In short: Passing mention — Epperson (39:18) lists it with NVIDIA, Apple, Alphabet and Amazon as the five tech giants that make up ~30% of the S&P 500, hence of most 401(k)s.
In short: Held and kept as a likely beneficiary: "the agentic operating system." Microsoft sits inside work identity, email, documents, databases and security, and Foundry offers hosted agents, multi-agent workflows and persistent memory. Its edge is permission context — which employee may see which record, who approves spending, how activity is audited. "Far more insulated from agentic threat than many other companies," despite lagging as a software stock.
Inside a company, an agent needs to know who is allowed to see which files, who can approve spending, and how every action is logged. Microsoft already controls that layer — company logins, email, documents, meetings, cloud and security — so agents at work will most likely run through Microsoft. That makes it a likely winner rather than a victim, even though software stocks have been weak. He holds it.
32:32Its foundry strategy describes hosted agents, multi-agent workflows, persistent memory, Microsoft 365 integration. The architecture puts Microsoft in a position to coordinate tasks across applications rather than merely add a chat box to each one. This is something that Sachin Adella has done incredibly well.
In short: Neutral on the equity, positive on the credit: in his hyperscaler list whose cash flow covers the coupons — "I just don't see a bankruptcy with the hyperscalers." The Global Crossing analogy (overbuilt capacity priced at a tenth of plan) is his risk case for the equity.
Microsoft's bonds are "solid" in his view, because its cash flow easily covers the coupon payments. His warning for the equity is the Global Crossing story from 25 years ago: companies laid huge amounts of fibre-optic cable expecting one price, and capacity ended up selling for a tenth of that. Overbuilding hurts owners, not necessarily lenders to cash-rich firms.
16:17The other guys are borrowing the money to use Anthropic's products. So, can Anthropic go bust? Sure. But that's an equity problem, not a bond problem. And I think the hyperscalers, I think Meta, Google, Amazon, Microsoft, Oracle, they have plenty of cash flow coming on through to fund, to make the coupon payments and pay these things back over time.
In short: BUY. ER 15.90%; fwd PE 24.5 vs 30.1 (18.6% under); RDCF 18.0% vs 15.0% (−3.0pp). Fair value $1,173.8 vs $494.65.
In short: Passing mention: among the hyperscalers that signed "behind the meter" nuclear power deals with owners of existing or soon-online plants — which, for nuclear, just diverts power utilities could have sold.
In short: Link: the hyperscalers should be green because the ROI is showing up. On Fundstrat's Mark Newton calling an imminent Mag-7 breakout, and Alphabet/Meta/Microsoft green through the weekend's negativity: "They should be green because you're starting to see monetization… when you saw Microsoft and you saw Google Cloud and you saw Amazon AWS all see these numbers in their cloud businesses accelerate last quarter substantially. That was real confirmation that you are starting to see the ROI… as long as they're going to spend and then they start to see the ROIs, that does benefit the entire food chain." Saccocia: "the hyperscalers are going to outperform… these frontier AI names over the course of the next couple of months."
The worry about AI has been that the big tech companies are spending hundreds of billions on data centres without proof it pays off. Link's evidence that it does: the cloud businesses that rent out that computing power — Microsoft's Azure, Google Cloud and Amazon's AWS — all grew faster last quarter.
That is the return showing up in revenue. So when AI-safety headlines hit, she thinks these companies should hold up better than the chipmakers and the private AI labs, and Saccocia agrees for the next couple of months. She makes an exception for Meta, which spends heavily but doesn't rent out cloud capacity the same way.
In short: Liked in May 2025 but only up 12% since; asked where he comes down, he declines to rate it: "pretty much any stock you mention, I've probably been long and or short depending on the period of time… When the facts change, I change." Today "I'm just a lot more nervous and that just cuts across to most stocks." Also cited for the overinvestment mindset — "Microsoft has said the biggest risk is underinvestment, not overinvestment" — and as a neocloud customer that can simply stop renting that capacity.
Microsoft was one of his picks in May 2025, but has only gained 12% since. Asked where he stands now, he wouldn't commit either way: he has been both long (betting it rises) and short (betting it falls) at different times, quoting Keynes — "when the facts change, I change." His general mood is "a lot more nervous" across most stocks.
Microsoft also comes up as a symbol of the spending race (it has said the bigger risk is spending too little on AI) and as a big customer of the "neoclouds" that could drop that rented capacity if demand slows.
59:57Where do you come down? Were you long and short over the course of a year on Microsoft? — Pretty much any stock you mention, I've probably been long and or short depending on the period of time. And there's a famous quote by John Maynard Keynes, right? When the facts change, I change.
In short: The example of a hyperscaler that "has shown an appetite to cut" AI spend, and "the market has told them that their stock can go back up again." Corporates won't hand 10% of free cash flow to AI immediately — more like 3–4%.
Microsoft is one of the "hyperscalers" pouring money into AI data centers. Taylor points out that when it signaled willingness to cut that spending, its stock went back up. Shareholders are rewarding restraint rather than endless capex funded by negative free cash flow and more debt. He expects big companies to put only 3–4% of their free cash flow toward AI at first, not the 10% the build-out would need.
19:50Very hard because your number is going to go down unless your productivity goes through the roof. And so for them to jump on and take a 10% hit immediately to do this is improbable. They're going to start out at three, maybe four. Some guys that really see it, five or six. And the ones that have shown an appetite to cut it, like Microsoft,
In short: Lebenthal's holding and the day's "port in the storm" — the spender that benefits if the frontier slows. Santoli: "Jim, you own Amazon, Alphabet, Microsoft, so you know, that's the port in the storm today… somebody wants to own these rather than Broadcom." Lebenthal: "it's hard to get awfully negative on the markets, Mike, when those names… are actually up… Microsoft, let's not forget, has had a heck of a nice bid recently. So when something as big as Microsoft is performing as well as it is, again, really hard to get negative on this market overall." Weiss explains the bid: "we're seeing Meta and Microsoft and Alphabet move up because say, OK, great, they're going to slow down spending."
Microsoft is one of the "hyperscalers" — the giant cloud companies that buy the chips and build the data centres AI runs on. When the Anthropic CEO suggested the industry should slow down how fast it pushes the most powerful models, chipmakers fell because they sell the picks and shovels. Microsoft rose, because slower spending means less cash going out the door for a company that is already earning money from AI through its cloud business.
Lebenthal owns it and treats its strength as a reason not to turn bearish on the whole market: when the second-largest company in the world is holding up on a bad day, it is hard to argue the market is breaking.
In short: Hyperscalers are mostly FCF-negative with capex outgrowing earnings, so every capex dollar is new debt; if stocks don't fall, Microsoft and Amazon keep issuing and rates keep rising "until the stock market pops."
Microsoft is one of the "hyperscalers," the cloud giants building AI data centers. Woo says these companies now spend more than the cash their businesses bring in, so the extra spending is paid for with borrowed money. That flood of new corporate bonds is a big reason long-term interest rates are rising. As long as stocks keep climbing, the borrowing continues and rates keep going up, until something breaks. It's a macro point, not a verdict on Microsoft's business.
21:51for Trump to taco, for oil price to stop going up. So if the stock market doesn't go down the oil price is going to keep going up until the stock market basically pops. The same thing with interest rates, right? Right now interest rate is trading on this whole AI capex story, right? If the stock market doesn't go down, that means Microsoft, Amazon, it's going to continue to issue debt to finance their AI capex and rates are just going to keep going higher until the stock market pops. So at this point, that
In short: "We have Microsoft with Microsoft Azure. All of them are renting out their capacity" — capex with a billable revenue line attached, the contrast that makes Meta the unique indirect spender. Also named as a company Dario's coordinated pacing would slow alongside OpenAI and Anthropic.
5:01We have Microsoft with Microsoft Azure. All of them are renting out their capacity. The only company that's really spending on it as an indirect way of creating revenue is Meta. And the ways that Meta actually makes money from AI, their core business, that'll continue to monetize. They can continue making improvements on that without the leading AI.
In short: Same data-center/security point: already under constant state-actor attack and will "increase their security as a result" of AI.
14:46state actors. Does Jacob think that actors in Russia and hackers aren't already looking at critical infrastructure in the United States? They're not already looking at exploits. Microsoft, Apple, all these companies, Amazon, they're being hacked every single day. hacking attempts every
In short: The most concrete disclosure in the captured section, and it is a revenue-quality attack: "In fiscal year 2026, Microsoft reported $34.3 billion in AI revenue, and $24.1 billion of that came from OpenAI alone." Net out its own investee and "Microsoft's own AI compute and software business is comparatively modest, a single-digit-billion-dollar business" — i.e. roughly 70% of the headline AI line is money routed back from a company Microsoft funds.
Microsoft says it earned $34.3 billion from AI in its 2026 financial year. The argument here is about where that money came from. $24.1 billion of it — about seven dollars in every ten — was paid by OpenAI, a company Microsoft is itself a major investor in and funder of.
Why that matters: normally, revenue is evidence that outside customers want what you sell. If most of it comes from a company you are simultaneously writing cheques to, the revenue is partly your own money coming back around, recorded as a sale. Strip OpenAI out and what is left — Microsoft's AI business selling to everybody else — is described as "a single-digit-billion-dollar business," roughly $10 billion rather than $34 billion.
The practical test this sets up is simple: the part of the AI revenue line to watch is the part that does not come from a company the seller also funds, because that is the only part that proves independent demand exists.
In short: Lebenthal's sentiment-reversal name. "We've talked about Microsoft a lot. Very forgiving valuation. And most importantly… the sentiment is coming back to Microsoft. This was a stock under $400.00 just a few months ago, and people were looking at it, actually tossing it into the heap of the SaaS names and thinking that it might somehow become disenfranchised. I think that will be looked back on." His general point is that the Mag 7 recovery is idiosyncratic rather than a basket trade: "each of these names has their own idiosyncratic reasons for going higher, and the valuations give support."
Microsoft sells business software and rents out computing power through its Azure cloud. A few months ago the market lumped it in with the ordinary subscription-software companies that were being sold off — Lebenthal's phrase is that it was "tossed into the heap of the SaaS names" — and worried it might get left behind by AI rather than benefit from it.
His argument is that sentiment, not the business, has turned. The valuation is "very forgiving" for a company of this quality, the stock was under $400 not long ago, and the people who wrote it off will look wrong. Note what he is not claiming: no new product, no new number. This is a re-rating call — the same earnings, valued more generously as the fear recedes.
In short: The second historical exhibit, on the same footing as META. Its multi-year-ceiling breakout chart accompanied the February 2024 note ("charts showing breakouts above multi-year ceilings for META… and MSFT") and is republished here on a six-year view with prior overhead resistance marked, purely to demonstrate what happened after the range expansion. No fresh view on Microsoft's business, valuation or rating is given, and it is not on the published portfolio tables.
In short: Named with Meta for the capex cash burn that makes board members "very very very very uncomfortable" — and the reason the whole complex is now financed by banks against double-A-rated future free cash flow.
Microsoft is named alongside Meta for the same reason: capital spending on AI data centers has been burning cash so fast it has made boards "very very very very uncomfortable."
Banks were happy to lend against these companies' top-rated future cash flow. But the loan exists today and the cash flow does not yet — so if the turnaround date keeps slipping, the risk transfers straight onto the lenders, and the crowded, record-priced bank stocks are where he expects it to show up.
7:20What these guys have done behind the scenes is when you talk to people that are close with these companies, the board members in San Francisco, the cash burn at say Meta, Microsoft has been so vicious in terms of the capital expenditures. And when you burn a lot of cash, it makes the board members very very very very uncomfortable.
In short: Dawson stacks three separate disclosure complaints on it. Capex optics: "Microsoft lowered its capex guidance because it reclassified financial and operating leases that move certain parts of its capex down from operating cash flow into investing cash flow. And so the capex calculation that they use looked to be lower than it actually was." Segment opacity: "there was a great article in, I think it was the FT, effectively saying Microsoft has done a terrible job at disclosing its AI business and it hides it in with a bunch of other businesses… it makes it really murky and hard to see." And the 25% OpenAI stake carried by the equity method. Counterweight in the same episode: Niles names it as the one survivor of his failed-buy-and-hold list — "Microsoft has done great through three different decades, right? But that's one company."
Three separate complaints, all about what you can and cannot see, none about the business. First, capex: Microsoft moved some lease costs from one part of the cash flow statement to another, which made its reported capital-spending number smaller without making the actual spending smaller. Second, disclosure: an FT piece argued the AI business is buried inside other segments, so nobody outside can size it. Third, OpenAI: Microsoft owns 25% of it, which under the "equity method" means a slice of OpenAI's profit or loss flows into Microsoft's income statement — so a private company's economics land in a public company's numbers with very little detail attached.
None of these is fraud, and each has a defensible technical justification. Dawson's point is about the pattern: when a company that is doing well starts making its numbers harder to compare, the question worth asking is why comparison has become inconvenient.
Set against that, the same episode gives Microsoft the only compliment in Niles' entire buy-and-hold demolition — it is the one company he names that "has done great through three different decades." Both things are in the episode, and both belong on the page.
21:37So Microsoft lowered its capex guidance because it reclassified financial and operating leases that move certain parts of its capex down from operating cash flow into investing cash flow. And so the capex calculation that they use looked to be lower than it actually was. Then you add on top of that, there's a bunch of stuff that's not even hitting the balance sheet or cash flow statements because it's all happening in these SPVs.
In short: Same role, same segment. Nathan cites it as one of the three hyperscalers whose cloud business gives the AI spend somewhere to be recouped — the thing Meta does not have — and names Azure as one of the two platforms the commoditized models will sit on and be price-shopped across. No view on the business or the stock.
30:06One of the knocks on Meta is that they basically have been lagging in the performance of their models to OpenAI, Anthropic, even Gemini. And the monetization of all that spend. They don't have that big cloud business like Amazon, like Microsoft, and like Google.
In short: 7% / $106k (+$52.5k). 13% growth × 28× = 12.23%. Will now disclose Azure directly ($29.4B last quarter, $101.9B FY26); Maia 300 custom chips (300k+ in 2027) to reduce Nvidia dependence. $500 (5%).
In short: Wapner's candidate for the real engine of the comeback — "maybe this is the one; it doesn't get the headlines that the NVIDIA does." The move: "from earnings day until today… there it is, 27%. This stock got a reboot after a bit of squirrely price action." Rob Sechan makes the quality-plus-weight case: "certainly, because of its weight in the index… it's a significant contributor. It's been kind of a laggard over the last year, basically a flatliner over the last year… PEs come in quite a bit on Microsoft. I still think it's one of the highest quality tech, broad-based tech names in the world. Their strength was on full display in their latest earnings report." Lebenthal cites Azure growing 40% as part of the CapEx-is-profitable evidence.
Microsoft is up 27% from the day it reported, having gone essentially nowhere for the previous year. Sechan's case has two parts, and the second is the more useful one.
First, size: it is one of the heaviest weights in the index, so when it moves the whole market moves with it — which is part of why the Mag 7 chart turned. Second, and this is the part that matters for anyone holding it, the price rose while the valuation fell. A flat year of earnings growth with a flat stock means the multiple — what you pay for each dollar of profit — compressed. So the rally started from a cheaper base than the chart alone suggests.
That is the same test Simpson applies to the whole group: a rise that comes with rising earnings is a re-rating you can hold; a rise that comes only from people paying more for the same profits is a rally you have to time.
In short: Context only: Azure is the second of the three accelerating clouds (35% → 43%, margins +2 points) he cites to argue the AI build is currently working. No stock view.
21:42And if you look at the biggest cloud platforms in the world, so Amazon Web Services, Microsoft's Azure and Google Cloud Platform, if you look at the growth rates for the cloud businesses that they have, they were growing 35% in the March quarter and that accelerated to 43% year-over-year growth in the June quarter.
In short: Azure is the second of the three accelerating clouds (35% → 43% growth, margins +2pt). It is also the single company he offers as the exception that proves his anti-buy-and-hold rule — "you always have some company that makes it through, like Microsoft has done great through three different decades, right? But that's one company" — and the first name on his daily hyperscaler credit-default-swap chart.
Azure is Microsoft's cloud platform and the second of the three that accelerated with expanding margins. Same conclusion as AWS: selling more, and keeping more of each dollar.
Microsoft has a second, unusual role in this conversation. Niles spends the closing minutes arguing that "buy and hold" is survivorship bias dressed up as wisdom — for every Apple there is an AOL, a Yahoo, a Nokia, a Cisco, an IBM. Microsoft is the one company he grants as a genuine multi-decade compounder: "Microsoft has done great through three different decades, right? But that's one company." The compliment is real, and so is the warning attached to it — one survivor out of a long list is not a strategy.
It is also the first name on his morning credit-default-swap screen. A credit default swap is essentially an insurance premium on a company's debt: when it rises, professional lenders are charging more to be repaid. Microsoft appearing on that watchlist is not a negative view on Microsoft — it is where he looks first for signs that the AI build's financing is getting expensive.
56:40IBM was a buy and hold, of market share leaders that then went into trouble. But you always have some company that makes it through, like Microsoft has done great through three different decades, right? But that's one company. And so I think you have to have strong conviction but loosely held, and that's the way you want to think about it. And just because a stock's down a lot, like one of my big disasters this year, we'll see how the rest of the year plays out, is Nike. Right.
In short: Reiterated buy at Bank of America with the target lifted to 600 from 500, and Jason Snipe owns it. His numbers: "Azure is now a $100 billion business — just crossed over that in the last quarter. Azure revenue was up 43% year over year and copilot adoption is improving, 30 million paid seats. So I continue to like this stock. It's now positive on the year, up around 4%." Wapner's chart point is that the print was the reset — "this stock hadn't done anything and I feel like after they delivered the stock feels like it got that next leg" (+27% from earnings) — while Snipe notes it "hasn't moved really that much since." The credit he singles out: the Copilot seat numbers and the commitment, from CFO Amy Hood, that they "remain cash flow positive." The counterweight comes from the options tape — Oliver Renick reports "at least one mega bull" closing two bullish call spreads (a 595-705 worth $19m and a 610-725 worth about $17m, more than 30,000 contracts, ~$36 million of credit) that still had until December 2027 to work: "the owner chose to bail instead on a day the stock is going against them."
Snipe owns Microsoft and his case is two numbers. Azure, the cloud business, has crossed $100 billion of annual revenue and is still growing 43% a year — a very large business compounding at a very high rate. And Copilot, the AI assistant sold on top of Office, now has 30 million paying seats, which is the first hard evidence that customers will actually pay for AI features rather than just trial them. Bank of America raised its target to $600 from $500 today.
The counter-evidence comes from the options market, and it is worth understanding. A "call spread" is a bet that a stock rises to a particular level by a particular date. Someone had roughly $36 million of such bets on Microsoft, expiring as far out as December 2027 — and today closed them all, banking the money rather than waiting. Renick's point is that this holder had years of time left and still chose to walk away on a weak day. It is one investor's decision, not a forecast, but it is the kind of thing worth noting when a stock has stalled.
In short: His illustration that capex is an option, not concrete: "Microsoft talked about this in the latest call… you're buying powered shells, you're not filling it with the chips yet. You're essentially buying a couple years of optionality" — capacity you can ramp fast if you need it. Rational even at a low-but-positive ROI.
Microsoft supplies Keller's clearest example of how to think about AI capital spending: it isn't concrete, it's an option. On its latest call the company described buying "powered shells" — data-center buildings with electricity connected but no expensive chips installed yet. That is deliberately buying a couple of years of the ability to scale up fast, without committing the full cost.
Read that way, the arithmetic everyone argues about ("what return does this $1,000 in the ground earn?") is the wrong question. You spend because not spending could cost you the whole business, and because the upside if you win is open-ended. A modest positive return on that option is still a rational purchase — which is why he expects the spending to continue no matter what the share price does.
30:46There's other ways that you can generate new products. There's just sort of this open-ended upside that maybe if we become the leading provider of compute then whatever it may be there's open-ended upside. So it's really an optionality thing. And I think the way they've talked — I think Microsoft talked about this in the latest call, maybe it was one of the other hypers, but you can structure this where you're buying powered shells, you're not filling it with the chips yet. You're essentially
In short: Scaffolding for the credit argument rather than a view. One of the seven hyperscalers carrying the $3.1 trillion of off-balance-sheet arrangements, "not all of which are actually guaranteed. But if you look across Meta, Google, Microsoft, Amazon, Oracle, Nvidia, and Broadcom, you can see that there is a lot there that doesn't need to be on an individual's balance sheet." Also one of the ten most active options names on Wednesday, and — via Amkor's advanced-packaging programmes — the customer behind the Cobalt CPU alongside Nvidia's Vera and AMD's Venice.
Microsoft appears only as one of seven names in a single number, and that number is the point of the section. Between Meta, Google, Microsoft, Amazon, Oracle, Nvidia and Broadcom there are $3.1 trillion of "off-balance-sheet arrangements" — long-term obligations, leases and commitments that are disclosed in footnotes rather than counted as debt on the face of the accounts.
They are not all guarantees, and Singh says so. But the credit market has started pricing them as if some are: the cost of insuring these companies' debt against default hit all-time highs this week. His view on whether that recedes is blunt — spreads "could calm down if they stop issuing more and more debt, but I don't think that that is the goal."
This is scaffolding for an argument rather than a view on the shares. The argument is that the AI build-out has grown large enough that the way it is financed, not the demand for it, is now the thing that decides whether the forecasts get hit.
Full passage: premium transcript (PDF).
In short: Peak-date evidence, not a stance: paired with Meta as the leadership names that "peaked over a year ago" — the longest-dated of the Mag7 / hyperscaler / chip highs he charts, and the clearest sign the complex has not been topping together.
In short: Referenced as part of the hyperscaler class, not named individually in this interview. The sizing that applies to the group: "Nvidia sells chips to hyperscalers. 70% of hyperscaler AI revenue is from Anthropic and OpenAI. And that equates to about 25 to 35% of their cloud revenue."
4:45Which of those companies, we don't know who they are, but we could all pretty much guess a bunch of them. — Mhm. — The way I see the entire AI story is that Nvidia sells chips to hyperscalers. 70% of hyperscaler AI revenue is from Anthropic and OpenAI. And that equates to about 25 to 35% of their cloud revenue.
In short: Buy target $400 from $490 (holding eight, $104k / 7%, +$50k) — and the most mechanically plausible trigger on the board: "after the last earnings report, Microsoft shot up around 25%… before that, it was at $400 per share. So, if Microsoft simply gives up the gains of its last earnings report bump, then this will be the one that gets that $10,000." He keeps a high multiple in the model — 28, "given the moat and the structure of the company" — on 12.4% EPS growth, giving a 16.8% annual return at $400, "which I believe would soundly outperform the market."
Microsoft's target is the most plausible one on the board, and for a very mechanical reason. The stock jumped roughly 25% on its last earnings report, from about $400 to $490. His buy price is $400 — so "if Microsoft simply gives up the gains of its last earnings report bump, then this will be the one that gets that $10,000." No crisis required; just a normal round-trip of one good quarter's enthusiasm.
He models it with a modest 12.4% earnings growth rate but keeps a high multiple of 28, on the view that the business quality justifies a premium — "given the moat and the structure of the company." At $400 that produces a 16.8% annual return over five years, which he says "would soundly outperform the market."
15:46after that, we have Microsoft. Microsoft is number eight. It's a $104,000 position. $50,000 of that being gains, and it's a 7% weighted position. For Microsoft to get that $10,000, I will need it to drop down to $400 per share from the $490 that it currently trades at. After the last earnings report, Microsoft shot up around 25%.
In short: Present on three sides of the AI plumbing this week: one of the hyperscalers told its Nvidia server bill is going up more than 15%; one of the five big-tech issuers funding capex with debt rather than free cash flow; and the counterparty behind the $17.4 billion cloud deal whose Vineland, New Jersey data-centre approval cleared "a major hurdle." No stated position — it is scaffolding for the crowding-out argument.
Microsoft appears three times in this call without ever being a recommendation. It is one of the hyperscalers told that its Nvidia server bill is rising more than 15%. It is one of the five big technology companies now funding AI construction with borrowed money rather than cash flow. And it is the customer behind Nebius's $17.4 billion cloud contract, whose New Jersey data-centre approval cleared a major hurdle this week.
Taken together these are the load-bearing facts of the call's central argument — that AI spending is now large enough to move the government bond market — rather than a view on the shares.
Full passage: premium transcript (PDF).
In short: BUY on Compounding Quality's sheet — in the same issue that opens with Chris Hohn selling all of it. The house numbers: ER 15.90%; fwd PE 24.5 against a 30.1 average (18.6% under); but the reverse DCF requires 17.5% growth against 15.0% expected (−2.5pp). Hohn's reason for exiting what was TCI's third-largest position at end-2025 is quoted without rebuttal: "AI could disrupt Office and Azure faster than the market thinks."
Microsoft sits on the buy list this month, and the same issue opens by explaining why Chris Hohn — whose fund made nearly $20bn last year, and whose tollkeeper approach the archive openly admires — sold every share of it. His argument is that artificial intelligence could erode Office and Azure faster than the market believes.
The letter's own figures are mixed: the shares are about 19% below their five-year average multiple, but the reverse-DCF test says the price already assumes 17.5% annual growth against 15% expected. No attempt is made to answer Hohn's disruption case. For a reader, the honest summary is that this issue contains a well-argued bear case for Microsoft and a rating that ignores it.
In short: Referenced twice by Harrington, both times as a name whose outcome she is deliberately not forecasting. On the NVIDIA print: "I don't think anything NVIDIA says is going to matter to NVIDIA. It might matter more to CoreWeave or Microsoft or OpenAI." And inside the Cisco thesis: "it doesn't matter if Microsoft does well, Meta does well, or Apple does well" — the point of owning the infrastructure layer is that you don't have to pick the model or the hyperscaler.
In short: Owned "quite close to 15 years" and now being sold out entirely into the last month's run-up: "what we're seeing in the Microsoft look right now is not great… we've sold Microsoft really all the way up two or three or four times, just little bits at a time. But this time we basically decided to sell it all… the valuation doesn't hold the price." The only brake is tax: "some accounts have it so low that we can't sell all of it."
Oxbow has held Microsoft for close to fifteen years across many client accounts and is now selling it outright. That's a significant act for a firm whose default holding period is three to ten years, and Oakley is explicit that it isn't a judgment on the company: it's arithmetic. Run the valuation forward, ask whether the multiple can hold, and "for us it looks like the valuation doesn't hold the price. That's why we sell it."
Two details matter for anyone copying the method. First, the selling is staged, not a single exit — "we've sold Microsoft really all the way up two or three or four times, just little bits at a time," and only now "decided to sell it all." Second, tax is a real constraint, not an excuse: some accounts bought so low that the capital-gains bill prevents a complete exit, so those positions are worked down over time rather than dumped.
12:55Now, we've owned Microsoft for a number of accounts quite close to 15 years. Okay. What we're seeing in the Microsoft look right now is not great. And so, on this little runup we've had in the last month, we're actually selling Microsoft. Okay. Now, I will say some accounts have it so low that we can't sell all of it for taxes.
In short: The clearest de-rating in the round-up. "Microsoft, once a fixture on this list, appeared only once at the end of June after falling more than 20% YTD." A name that used to be automatic across top-five holdings has fallen out of nineteen of twenty — smart-money concentration rotating to AMZN, TSM and GOOGL. A disclosed author holding.
Microsoft is Azure, Office, Windows and the deepest commercial relationship with OpenAI — for years an automatic core holding for almost any large fund.
In Q2 it nearly vanished from this list: "Microsoft, once a fixture on this list, appeared only once at the end of June after falling more than 20% YTD." One appearance out of twenty funds' top-five holdings.
Two readings, and the article doesn't choose between them. Either institutions rotated the money into Amazon, TSMC and Alphabet on a view about relative AI positioning — or a 20% drawdown mechanically pushed Microsoft out of "top five by value" lists without anyone selling a share. Both are consistent with the data, which is exactly why a top-five-holdings table is a weak instrument for detecting a change of heart.
In short: Named only as a mega-cap earnings driver. "It's been the mega-cap stocks, like Microsoft (MSFT) and Alphabet/Google (GOOG), that have been reporting blockbuster earnings" — the mechanism by which a cap-weighted index with its top 10 at 38% of the total delivered a 50% total return while its forward multiple fell. Unlike GOOG, no quantified accounting critique is attached to Microsoft in this post; it is cited as one half of the illustration. No view is taken on the shares.
In short: "I still like Microsoft" — and the sorting variable is capex discipline: "Microsoft is the only company… up 13% after earnings, because they said capex would be roughly flat. But outside of Microsoft, all of the hyperscalers are now burning cash flow."
Microsoft is the exception that proves the rule. It rose 13% after earnings for one reason: it said capital spending would stay roughly flat instead of rising again. "Outside of Microsoft, all of the hyperscalers are now burning cash flow."
That makes spending discipline, not revenue growth, the variable that now separates the big tech names from one another — and Microsoft is the only one currently on the right side of it.
18:14So, I think Microsoft is the only company, so Microsoft was up 13% after earnings, because they said capex would be roughly flat, they weren't increasing it. But, outside of Microsoft, I think all of the hyperscalers are now burning cash flow. And so, even though earnings look strong, you have to be really careful with these companies, because they're not the capital-light free money generating businesses they were in the past.
In short: The episode's biggest disagreement. Hohn took a 17% position to zero on a moat call — Claude "could overtake many of the responsibilities and suppress the pricing power of Microsoft Office," with Azure falling behind it as a domino. Carlson concedes the pressure is real ("certainly transitioning some of the value to Claude's interface") but keeps the stock: "I'm not so concerned about Microsoft that I'm going to be selling my position… I believe Microsoft still remains very wide moat… In fact, I become more bullish on Microsoft over time." Ackman added another 10%, and Microsoft has since recovered into the green YTD — Hohn sold at the lower valuation.
This is the episode's real argument. Chris Hohn — a manager Carlson rates as among the best in the world — deleted a 17% Microsoft position entirely. His reasoning is a moat judgement, not a valuation one: AI assistants like Claude increasingly do the work people used to do inside Word, Excel and PowerPoint, which could gradually "disintermediate" Office — sit between Microsoft and its customer and take the pricing power. And because Office is what gets Microsoft into every Fortune 500 company, Hohn thinks Azure would fall as a domino behind it.
Carlson accepts half of it. He agrees value is genuinely shifting toward the AI interface. He just doesn't think the moat is narrow enough to sell: "I believe Microsoft still remains very wide moat, at least wide enough to have as a big position in my portfolio. In fact, I become more bullish on Microsoft over time." Bill Ackman took the opposite side of the same trade, adding another 10%.
The scoreboard so far favors the holders: Microsoft has recovered into the green for the year, meaning Hohn sold at the lower valuation. Carlson's own resolution is deliberately non-committal on the pair trade — he owns both Microsoft and Google and is bullish on both, so he never has to be right about which one wins.
14:02Now, when I look at Chris Hohn, I agree with some of the concerns to some extent. I think that there is some pressure on Microsoft with Claude. They are dealing with something that is interacting with their software and certainly transitioning some of the value to Claude's interface. But, I'm not so concerned about Microsoft that I'm going to be selling my position.
In short: The best of a deteriorating set on Eisman's numbers — "Microsoft's, which was the best of anybody, was about 19 billion [of quarterly cash flow] but it was down around 25% versus last year" — and the strongest tape on Verrone's: "The 5-day change in Microsoft is the best we've seen since the IPO in '87," evidence that the negativity of two-to-three weeks earlier had gone too far. Also the historical template for the antitrust segment ("Microsoft in the '90s, very similar story" to IBM).
Microsoft is the best-placed of the hyperscalers on both measures used in this conversation, which is what makes its numbers sobering. Its quarterly cash flow of roughly $19bn was "the best of anybody" — and still down about 25% from a year earlier. The whole group is converting a smaller and smaller share of its revenue into cash it can hand to shareholders.
On the tape it is the strongest of the lot: its best five-day advance since the company came public, which Verrone reads as evidence that the negativity three weeks earlier had overshot rather than as proof of anything fundamental.
It also plays the historical role in the antitrust segment. Microsoft spent the late 1990s in court, "won" in the same sense IBM did, and the cost was the attention of its management during a technology transition.
23:18And I think that's been mistaken for some deterioration in AI fundamentals when it's just been we have to think about them a little bit differently. You look at some of these moves. Google's basically back at the highs here. Amazon just made new highs. The 5-day change in Microsoft is the best we've seen since the IPO in '87.
In short: Named in the "value accrues to the infrastructure layer" conclusion — the big-three public cloud vendors' revenue growth accelerated from 23% y/y in 1Q25 to 35% in 1Q26 to 43% in 2Q26, and "arguably more important is public cloud operating margins expanded from 34% to 37% and to 39%," which is what makes the hyperscaler capex acceleration (+84% → +92% → ~+100% forecast in Q3) fundable. Also unveiling the Maia 300 AI chip in September, and Norway's sovereign fund's second-largest holding.
Full passage: premium transcript (PDF).
In short: The counterparty-transformation case: an AAA/AA-grade credit routing its spending through a far weaker data-centre developer or neocloud, which then shows investors a "giant backlog" that is really the hyperscaler's own off-balance-sheet commitment. His challenge, put directly: if Microsoft wants to borrow $10bn, the cheapest route is a plain-vanilla general-purposes corporate bond — so "flexibility is probably a convenient excuse" for optics that hide the shift "from being lean mean cash machines into being capex hungry utilities."
47:31You want freedom? Yeah, definitely. But in practice, it comes at a cost. — Broadly speaking, — if Microsoft wants to sell 10 billion, wants to borrow 10 billion to build a new data center, what is the cheapest way for a large mainstream public company to do so? Is it to sell to like 10 private credit firms or maybe a handful of private credit firms to do it without a rating, doing it quickly? We can do it opportunistically that way.
In short: Named by the host among the software names "slaughtered" on AI-disruption fear and since recovered. Newton's response is on the group, not the stock: "initially we saw the software deterioration that started actually late last year. That has since bottomed. We've seen software start to push off the lows" — the first of three tech legs to stabilise.
Microsoft was named by the host, not by Newton, as one of the big software names "slaughtered" on the fear that AI would erode their businesses. Newton's answer was about the group rather than the company.
His point is a sequencing one: software was the first part of technology to break down, starting late last year, and it is the first to have healed — "that has since bottomed. We've seen software start to push off the lows." Semiconductors and then memory chips broke down after it and have since stabilised too. So treat this as a neutral, group-level observation: the software damage is repaired, with no specific view on Microsoft's own chart.
13:27Initially we saw the software deterioration that started actually late last year. That has since bottomed. We've seen software start to push off the lows and then of course it shifted to more semiconductor semi-cap equipment weakness and then the final shoe to drop was really a lot of these memory names and one by one all of these areas have stabilized including the mag seven, the hyperscalers.
In short: The defended layer, and the explicit contrast case: "Google and Microsoft have multiple revenue streams from established businesses which are very unlikely to simply disappear. They also have hyperscaler businesses to balance their vulnerability." Caveat kept in his own words — "their LLM businesses are also questionable."
Microsoft sits on both sides of the AI business. It runs Azure — a hyperscaler, meaning it owns the data centres AI models live in — and it also has a large stake in the model layer through OpenAI. Eisman treats the first half as durable and the second half as the risky part.
His argument for the durable half is unusual: normally a business that has to spend colossal sums is considered weaker, but here the spending is the defence. "The amount of money it takes to be a hyperscaler is insane and that expenditure itself is a moat" — very few companies on earth can write those cheques, so competition is structurally limited. On top of that, Microsoft earns money from Windows, Office and enterprise software that would not vanish if AI disappointed. He still flags the caveat in his own words: "their LLM businesses are also questionable."
10:24But on the other hand, the companies he's less bullish on are the LLM providers, specifically the ones that exclusively offer LLMs. So, we're talking about Open AI and Anthropic. — Anthropic and Open AI are also problematic because they don't have the breadth of revenue streams of Google and Microsoft. Google and Microsoft have multiple revenue streams from established businesses which are very unlikely to simply disappear.
In short: Named with OpenAI and Meta as the customers CoreWeave rents compute to — i.e. one of the counterparties behind the 98% committed-contract revenue and the $104B backlog, which is also the customer-concentration risk the issue flags ("customer concentration matter[s] almost as much as growth"). Azure also appears on the other side of the trade, as one of the "broad software ecosystems" neoclouds deliberately do not try to replicate. Referenced; not a stance call. (Also a disclosed author holding.)
In short: "Microsoft, Google and Amazon seem to be the hyperscaler winners." The hyperscalers "at least on the surface are not the problem" and "have existing franchises" — but the cloud business is "an ever-increasing percentage of the revenue," so "in a sense the future of these companies is now very dependent on the future success of Anthropic and OpenAI."
Microsoft is one of the three "hyperscalers" — the companies that own the giant computing warehouses everyone else rents. Eisman puts it in the winners' column: the infrastructure layer has a real barrier to entry (the sheer cost), a real existing business, and revenue that keeps compounding.
The caveat is a dependency, not a flaw in the business. Because cloud is an ever-larger share of Microsoft's revenue and a large slice of that cloud revenue traces back to Anthropic and OpenAI, its future is "very dependent" on two private companies it doesn't control. If those two stumbled, he expects margin compression and a return to slower growth — a downgrade, not a collapse, because "the hyperscalers have existing franchises."
12:49This bare case of entropic and open AI could occur but it might occur a year from now and for markets that is an eternity. Winners and losers are beginning to emerge amongst the big companies. Microsoft, Google and Amazon seem to be the hyperscaler winners. Oracle is questionable with its debt barely rated above junk. Meta is having problems.
In short: Same list of his own prior positions — "I had Microsoft before he had Microsoft in his portfolio" — establishing that the overlap with Pershing Square runs the other way. A continuing holding, not re-argued here.
1:35I've already been in a lot of these companies. I bought Meta before it was revealed that Bill Ackman bought Meta. I had Google before he bought Google. I had Microsoft before he had Microsoft in his portfolio. I had Amazon before he bought Amazon. I had Netflix twice before he bought Netflix. And the same goes for now MasterCard and S&P Global.
In short: #3, and the pick that is also its own proof. Microsoft is one of the six companies present in both the 2005 and 2026 top-twenty tables — second at $278bn then, fourth at $2.778trn now. The case: "AI will become even more important. Microsoft is a leader in AI through Windows, Azure, Microsoft 365, and its partnership with OpenAI… Switching is expensive and difficult. Microsoft's ecosystem is so integrated that customers are unlikely to replace it." Also named among Chris Hohn's tollkeepers in the 30 June profile.
Microsoft sells the software companies run on — Windows, Office, Teams — and rents them computing power through Azure. Once an organisation's documents, email, identity system and applications all live in one vendor's ecosystem, moving is a multi-year project nobody wants to lead.
It is also the pick that proves the article's own point: Microsoft was the second-largest company in America in 2005 and the fourth-largest in 2026, one of only a handful to hold that position across two decades of technological upheaval. It did so by moving from selling packaged software to renting cloud capacity — exactly the "continuous adaptation" the piece names as the common factor.
In short: "Microsoft added almost half a trillion dollars of market value in a single session, a record for any stock, after Azure revenue grew 43%" — cited as the scale of what the data-center/AI buildout is still underpinning, the demand side of the same story whose financing side he is questioning.
Full passage: premium transcript (PDF).
In short: "For the first time in 15, probably 14, 15 years… we've been a seller of Microsoft," out of almost all portfolios (slower where the embedded gains are largest). "We don't see a lot for a stock, and that's been a great company for us. Don't get me wrong… Where am I going to be in 2 or 3 years here with this company?" People are "blinded by these big hyperscalers."
This is the headline move of the interview. Oxbow has sold Microsoft out of almost every portfolio — "for the first time in 15, probably 14, 15 years" — trimming more slowly only where clients have huge embedded gains and would face large tax bills.
He is careful not to call it a bad company: "that's been a great company for us. Don't get me wrong." The objection is forward return at this price. His test is a simple question — "Where am I going to be in 2 or 3 years here with this company? Is it really going to produce for me?" — and his read is that "people are blinded by these big hyperscalers right now. They're not thinking about the next two or three or four years."
The same instinct sits behind his refusal to invest in the data-centre build-out: he compares it to the fibre-optic construction boom of 1999–2000, where the capacity built for a "new paradigm" turned out to be far more than anyone could use, and the stocks suffered for years.
22:03It's a one-time throw. You say that won't happen again, but you're right on that. I go I look at it a different way, actually, but we don't go in and look at companies that are going to get the most money cuz we a lot of those come I'll give an example. If you look at our chief investment officer and I and myself, we for the first time in 15, probably 14, 15 years, we've been a seller of Microsoft.
In short: Firestone: "I'm a believer that Microsoft can continue to run" — up 39% in this recent run after "people were talking about Microsoft is dead, no one's going to touch Microsoft. It all seemed crazy… how do we live without Outlook?" Bernstein raises its target to 660 from 647 (Outperform). Renick's options color: combined Palantir + Microsoft premium is nearing $1 billion on the day, with calls outnumbering puts almost 3-to-1 in Microsoft.
Microsoft is up 39% in the current run, after a stretch when, as Karen Firestone puts it, "people were talking about Microsoft is dead, no one's going to touch Microsoft. It all seemed crazy." Her bull case is deliberately unglamorous: businesses cannot function without Outlook and the calendar, so the revenue is about as sticky as software gets. Bernstein nudged its target up to $660 and she is "a believer that Microsoft can continue to run."
The options market agrees loudly: buyers of call options (bets the stock rises) outnumbered put buyers almost three to one, and the combined money spent on Microsoft and Palantir options approached $1 billion in a single day.
In short: Held and endorsed, but explicitly not a buy-today call: at $507 "I'm not highlighting this one as great value or one that you have to jump and race to buy, cuz it's already gone up." He uses it as the worked example of the method — the time to highlight it was January 2025 at $220 (his "Big Tech is cheap" video), when something was visibly wrong; he bought at 244 and around there, and it has since outperformed the S&P by going up ~120%. "I have a huge holding in Microsoft, $82,000 with $40,500 in gains… In my other portfolio… $25,000 worth with $13,000 in gains."
Microsoft is a large holding — $82,000 in one account (up $40,500) and $25,000 in another (up $13,000) — and it is up around 120% since he highlighted it. But the point of raising it here is the opposite of a recommendation: at $507 the opportunity is gone. "I'm not highlighting this one as great value or one that you have to jump and race to buy, cuz it's already gone up."
The moment that mattered was January 2025, when the stock was $220 and something looked wrong with it. That is the uncomfortable rule the whole episode rests on: the only time a quality business is cheap is when there is a visible problem, so the buying has to happen while the news feels bad, not after it feels good.
1:44It just recently went up 16%, but even prior to this year, during the point where it was really the best time to be buying Microsoft, that point in time was right there. January of 2025, Microsoft sold for $220 per share, and that's the time to be highlighting Microsoft as a buy. And I made content highlighting Microsoft as a buy during that time period.
In short: Eisman's cold open: "Last night, Microsoft reported great numbers and the stock was up 10% plus… Today, Thursday, the market has had a big rally because of the Microsoft results." The single print carrying the tape on the day of taping (Jul 30).
Context only — the episode was taped the morning after Microsoft's quarter, and the print is what had the whole market rallying that day: "great numbers and the stock was up 10% plus." The detailed dissection (Azure re-accelerating to 43%, free cash flow down 23%) is on the Jul 31 wrap.
0:05Hey, it's Steve Eisman. Welcome to another episode of the Real Eisman Playbook. This episode is being recorded on Thursday, July 30th. Last night, Microsoft reported great numbers and the stock was up 10% plus. Meta reported hilarious numbers. They had negative earnings growth, 28% revenue growth, 55% expense growth.
In short: "There's only one hyperscaler left with positive free cash flow to equity right now and that's Microsoft… everyone else is kind of in cash burn mode." It also held capex flat while guiding above consensus, took Azure from 39% → 43% with 45% guided, and rallied 16% in a session (its fifth-largest one-day move ever, helped by the Situational Awareness forced sale) and added $450B of market cap on Aug 3 — the largest single-day gain on record ex-Nvidia's $441B in April 2025. Copilot is at 30M of 450M paid M365 seats (<10%), which Niles treats as the runway. Morgan Stanley has Microsoft winning the cloud in all three model scenarios. The offset: OpenAI is 40% of its cloud backlog and 70% of AI revenue.
The cleanest statistic in the call: Microsoft is the only one of the big cloud companies still generating spare cash for shareholders after paying for its AI build-out. Every other hyperscaler is now spending more than it earns and funding the gap with debt.
It also did the two things the market rewards at once — held its spending forecast flat while guiding revenue above expectations — and its cloud arm, Azure, accelerated from 39% to 43% growth with 45% guided next quarter. The stock rose 16% in a session and added $450 billion of market value in a day, the largest single-day gain on record.
Two things to watch. Copilot, its AI assistant, is used by only 30 million of 450 million paying Office subscribers — the upside case, but also proof adoption is early. And its biggest cloud customer is OpenAI, which represents 40% of its committed future cloud revenue, tying Microsoft into the same circular AI financing loop as everyone else.
Full passage: premium transcript (PDF).
In short: Up ~8.5% on the week after "the stock had done virtually nothing — it was such a disappointer." Simpson owns it and is harvesting the move with covered calls: the stock round-tripped from 550 in January to 350 within weeks while he held a 590 call, he added in the 370s and 380s, "and then all of a sudden this report blew it out of the water — exceeded any expectation that I could have dreamed of. So we wrote a 530 call expiring in a couple of weeks. It's 8% out of the money, but we're bringing in 10% annualized premium… no loss in conviction in the name whatsoever, but just trying to add a little alpha, a little premium on the edges." Talkington, who owns it, was "exhausted with the stock performance and actually with the company. The only reason I hadn't sold it is Amy Hood and Satya, which continue to be a masterclass" — and she credits them with "reading the room": committing to free cash flow and staying out of the debt market ("I believe Microsoft and J&J are the only AAA companies out there"), unlike Google. Cloud came in at 43 against a 38–40 run rate with Hood guiding upward — "so I do think it's a re-rating. People are looking at this in a new light." Her caveat: Microsoft and Google both have a bureaucracy problem, "and I think that's why you have to continue to watch the departures leaving to go do their own thing."
Microsoft rose about 8.5% this week after a long stretch of doing nothing. Bryn Talkington admits she was "exhausted with the stock performance and actually with the company," and the only reason she didn't sell was her faith in CEO Satya Nadella and CFO Amy Hood. What they got right, in her view, was "reading the room": committing to remaining free-cash-flow positive and staying out of the debt market, unlike Google — Microsoft and Johnson & Johnson being, she believes, the only two AAA-rated companies left. And the number that mattered: cloud growth came in at 43% against a 38–40% run rate, with guidance raised. "So I do think it's a re-rating."
Kevin Simpson's contribution is a lesson in managing a position rather than a view. He owns it, added in the $370s and $380s when it fell from $550 to $350, and has now sold a covered call at $530 — meaning he collects a cash premium today in exchange for agreeing to sell his shares if they rise above $530 in the next couple of weeks. That's a 10% annualised return on a strike 8% above the current price, taken deliberately after a big move because option premiums are rich. "No loss in conviction in the name whatsoever, but just trying to add a little alpha on the edges."
In short: Dean. A beneficiary of open weights — enterprises running cheap open models still have to rent someone's cloud, and "the economics translates to whoever controls the GPU." Capital allocation is "a lot more measured and a lot more rational" than Meta's because Microsoft is actually in the cloud business. But it sits in the ~10× EV/sales hyperscaler cohort where "a lot is going to have to go right."
54:00— We believe capital allocation with Meta is a bit different than Microsoft and Google who are in the cloud business. So that spending is a lot more measured and a lot more rational. And you could say for instance Meta is doing it as a defensive measure against Tik Tok and other incumbents that could build out GPUs and then they hit some type of app where it's magical and AI is being produced on there and it attracts people's attention which directly hurts Meta. So you could say that it's
In short: One of the four hyperscalers whose reaffirmed and raised spending plans Ethridge calls "reason to feel bullish" — "obviously trickle-down economics… it's going to go through the 2nd and 3rd order effects." Also one of the "largest purchasers" NVIDIA is diversifying away from by funding neoclouds (Brown). No individual call this episode.
In short: Jason Snipe's final trade: "Azure was strong, Copilot starting to show some life." Also cited by Terranova as proof the tape "wants to get behind the underdog and the underperformance" — Microsoft (and Palantir) rewarded for exactly that, which is why he finds Disney's muted reaction so telling.
Jason Snipe picked Microsoft as his final trade on two specifics: Azure — the cloud business — was strong, and Copilot, the AI assistant bundled into Office, is "starting to show some life," i.e. beginning to show up in revenue rather than just in demos. Joe Terranova separately cites Microsoft as evidence that this market rewards previously unloved names.
In short: Named among the customers lined up behind the Helios ramp (with Oracle and Anthropic), following the first deployments at Meta and OpenAI — evidence that AMD's rack-scale AI system has hyperscaler commitments beyond the initial two. (Referenced; not a stance call.)
In short: The same pattern as AWS: "you look at Azure, similar situation where the growth rate for that accelerated by about 4% to 43%, and the margins expanded by a percent." Azure is also named as one of the platforms that will do the routing — "if you're on Azure or Google Cloud platforms… those services will route whatever you're trying to do to the best model available for the task."
Azure is Microsoft's cloud platform, and it shows the same shape as AWS: growth accelerated about 4 percentage points to 43%, with operating margins up a point. Same conclusion — faster and more profitable at the same time.
Azure has a second role in his argument, and it is the more interesting one. He names it as one of the services that will "route whatever you're trying to do to the best model available for the task" — meaning the platform quietly decides, request by request, whether your question goes to an expensive frontier model or a cheap open-source one. The customer just gets an answer.
That routing layer is what turns the individual AI labs into replaceable suppliers. Whoever controls the routing keeps the customer relationship and the pricing power; the model behind it becomes a component you swap out the moment a cheaper one is good enough. Microsoft owns one of the few such control points.
3:22AND WHEN YOU LOOK AT THE RESULTS, AWS SAW REVENUE GROWTH ACCELERATE 9% FROM THE MARCH QUARTER TO THE JUNE QUARTER, THE 37% GROWTH, BUT MORE IMPORTANTLY TO ME IS THE OPERATING MARGINS ALSO EXPANDED BY A PERCENT. YOU LOOK AT AZURE, SIMILAR SITUATION WHERE THE GROWTH RATE FOR THAT ACCELERATED BY ABOUT 4% TO 43%, AND THE MARGINS EXPANDED BY A PERCENT.
In short: "Microsoft also put up a great report" — up 16% in a single day at a $3T-plus cap, then another 5%, "completely erasing all losses for the entire year of 2026." He credits it as part of the sentiment reversal (though Amazon is the bigger driver); his Passive Income position is back to $80,000 with $37,000 of gains.
Microsoft is the other big cloud spender the market had turned on. Its report was strong enough to move a $3-trillion-plus company up 16% in a single day — then another 5% — wiping out every loss it had taken in 2026. Carlson holds it (an $80,000 position, up $37,000) and counts the report as part of what flipped sentiment.
His nuance is about causation: the rest of the group only moved after Amazon reported, not after Microsoft. So Microsoft helped, but Amazon is what changed investors' minds about heavy AI spending generally.
0:40Completely erasing all losses for the entire year of 2026. So Microsoft was struggling. People are concerned about it. All of that's changed in the past one week. Then we also have Google, which investors surprisingly sold off after they released an incredible earnings report, but the stock is starting to get a lift again from Amazon's report.
In short: The one capex spender the market rewarded — "it crushed earnings." Revenue $90B (+18%) / EPS $4.81 vs $4.24 expected; Microsoft Cloud $60B (+27%); Azure +43%, past a $100B run-rate; Copilot >30M paid seats; capex $41B with FCF still +$19.6B and 2027 capex guided flat at ~$117.5B (unlike Meta). Nadella promised positive FCF in 2027 and showed his ROIC math on $357B of cash capex; depreciation life extended from 15 to 25 years. Stock +9% after hours, +13% the next day.
The market has spent a year punishing big tech companies for pouring cash into AI data centers without proof it pays off. Microsoft is the exception the market rewarded, and the difference is discipline: it spent $41 billion in the quarter and still generated $19.6 billion of spare cash, and — unlike Meta — it did not raise next year's spending plan.
The business underneath is also accelerating: its cloud arm Azure grew 43% and passed $100 billion of annualised revenue, and its AI assistant Copilot now has over 30 million paying corporate seats — evidence that the AI spending is actually being sold to customers. One accounting note worth knowing: Microsoft extended how long it assumes its data centers last, from 15 years to 25, which spreads the cost over more years and flatters near-term profits. The stock rose 9% after hours and 13% the next day.
Full passage: premium transcript (PDF).
In short: The share benchmark: total cloud infrastructure spending grew 43% Y/Y to $143B in Q2 — an 11th consecutive accelerating quarter and the fastest pace in eight years, with GenAI-specific cloud services +165% (Synergy Research Group). AWS held its leading 28% share against Azure's 20% and Google Cloud's 15%; the big three now control 63% of cloud infrastructure and 67% of public IaaS/PaaS. Microsoft is still growing faster than AWS, but "all hyperscalers remain supply-constrained, so small quarter-to-quarter market-share movements should not be overanalyzed." Also cited in the cash-engine chart as one of the hyperscalers whose trailing operating cash flow grew 34% to a combined $660B. A disclosed author holding. (Referenced; not a stance call.)
In short: The capex thesis proving out in real time — down 17% YTD on capex fear, then +15.5% in a single day on a 13% EPS beat and a revenue beat, back above $450 and erasing the year's losses. "Microsoft just proved to investors that this capex spend is a good thing" — monetizing AI, gaining AI users in its bundle across the Fortune 500. He owns ~$90k combined ($73k passive income + $22k Story Fund, >$40k in gains).
Microsoft spent most of the year down 17% while the rest of the market rose, purely because investors were nervous about how much it was spending on AI data centers. Then one earnings report changed their minds: profit per share came in 13% above expectations, revenue also beat, and — the part Carlson cares about — the company showed the spending is actually working, with more customers using its AI features, easier bundle sales across big corporate clients, and AI helping its other products.
The stock jumped 15.5% in a single day and erased the whole year's decline. Carlson's point about scale is worth pausing on: Microsoft is a $3.4 trillion company "moving like it's a penny stock," which tells you how much fear was priced in over something he thinks investors "should have already known." He owns about $90,000 across his two portfolios, over $40,000 of it profit.
5:56Worried if these bets are going to pay off, worried if it's going to be profitable. Microsoft fell 17% year to date. The stock was going down while the rest of the market's going up. And then in a single day with this report, investors have changed their mind. We look at Microsoft's report and their earnings per share grew way above expectations.
In short: "Microsoft had a good quarter… caught in the crosshairs of the AI debate all year" — feared as a software victim and as a capital-hungry hyperscaler, so the stock was −19% YTD into the print. EPS 474 +30%; revenue +18%; "most importantly, Microsoft's cloud business Azure saw revenue growth accelerate to 43% versus 40% in the March quarter." The blemish: "free cash flow of 19.6 billion was down 23%. Still, I would characterize this as a very good quarter" — stock up strongly after hours.
Microsoft entered this print hated from both directions — as a software company AI might replace, and as a data-centre builder whose spending has exploded — which is why the shares were down 19% on the year before reporting.
The quarter answered the second worry. Profits rose 30%, revenue 18%, and the number that matters most, Azure cloud revenue growth, accelerated to 43% from 40% the prior quarter. Acceleration is the key word: growth speeding up means the AI spending is producing revenue, not just cost.
The blemish is the same one that has punished everyone else — free cash flow fell 23% because of the building. Eisman notes it and moves on: "still, I would characterize this as a very good quarter." Under his new framework Microsoft is the ideal shape — a hyperscaler with a genuine moat (the sheer cost of entry) plus established software revenue that doesn't depend on winning the model race.
13:07Wednesday night, some very important companies reported and I'd say the overall results were very mixed. On the positive side, Microsoft had a good quarter. Microsoft has been caught in the crosshairs of the AI debate all year. It's a software company, so some investors are worried that AI software will replace it. On the other hand, it's also a hyperscaler and therefore its capital needs increased dramatically.
In short: The main story — the "open-weight hedge": a multibillion-dollar closed-model alliance with OpenAI on one side, Azure as the indispensable open-weight deployment platform on the other. "Microsoft does not need to predict which model architecture wins. It wants to provide the compute, governance, and security layer underneath all of them." Q4 FY26: revenue +18% Y/Y to $90.0B (a $2.4B beat), gross margin 67% (−1pp), operating margin 45% (flat), non-GAAP EPS $4.74 ($0.50 beat). Segments — Productivity & Business Processes +14% to $37.8B (M365 Copilot, E5, early E7); Intelligent Cloud +32% to $39.3B on 43% Azure growth, above the 39–40% cc guide and >$1B above the segment high end, on better fleet efficiency and faster capacity landings; More Personal Computing −4% to $12.9B as weak Windows OEM and Xbox (content and services −10%, third straight decline, ~3% margin plus impairments) outweighed +10% Search advertising. Copilot passed 30M paid seats (from 20M in Q3, 15M in Q2 — doubled in six months) at under 8% penetration, with >50,000-seat customers up over 7x Y/Y and weekly engagement "on par with Outlook and Teams," now billed as "seat plus usage." Backlog: commercial RPO +84% to $678B, more than twice annual revenue — OpenAI is still the largest disclosed component, but all sequential growth came from non-frontier customers and RPO rose 25% excluding OpenAI. The cost: cash CapEx more than doubled to $35.8B, FY26 total $115.9B (+80%), operating cash flow +30% to $55.4B yet free cash flow −23% to $19.6B; Microsoft Cloud gross margin −3pp to 65% and Intelligent Cloud to ~57%, though segment operating margin held near 41% as opex rose only 10% against 32% revenue growth. Quality-of-earnings flag: a $480M OpenAI gain was excluded from non-GAAP but a $3.2B Anthropic gain was left in. Efficiency offsets: 40% better performance per watt on Maia 200, 89% lower GPU cost in Dynamics 365 (MAI-Voice-2-Flash), up to 84% lower in PowerPoint (MAI-Image-2.5). Q1 Azure guided to ~45% cc with H1 growth accelerating and demand still exceeding capacity. Accounting flag: from FY27, data-centre and office useful life extends from 15 to 25 years and more future data-centre leases shift from finance to operating leases — which sit outside reported CapEx. A disclosed author holding. (Analysis, not a stance call.)
There's a fight in Washington over whether AI companies should be allowed to publish their models' "weights" — the trained internals that let anyone download a model and run it on their own computers, instead of renting it from the owner over the internet. Microsoft has quietly taken both sides, and that's the point of this issue. It keeps a multibillion-dollar partnership with OpenAI, whose models stay locked away, and it is making Azure the best place to run the downloadable rivals. Microsoft doesn't need to guess which type wins; it wants to sell the electricity, plumbing and security underneath all of them.
The quarter was strong. Sales rose 18% to $90 billion and beat expectations, and Azure — the cloud-computing business, the one investors actually watch — grew 43%, better than the 39–40% management had guided. Copilot, the AI assistant bolted onto Office, added 10 million paying seats in three months to pass 30 million, though that's still under 8% of Microsoft's business customers, so there's a lot of runway. And the order book is extraordinary: $678 billion of contracted-but-not-yet-delivered revenue, more than twice what the company earns in a year.
Two cautions. First, the order book leans heavily on one customer, OpenAI — the useful check the article performs is that it still grew 25% with OpenAI excluded, and all the growth this quarter came from ordinary businesses, so the concentration isn't getting worse. Second, the cost: Microsoft spent $115.9 billion on data centres this year, up 80%, which is why free cash flow — the money left after building things — dropped 23% even though the business generated 30% more cash. Profits are also being flattered a little by investment gains: Microsoft removed a $480 million gain on its OpenAI stake from its adjusted profit, but kept a $3.2 billion gain on its Anthropic stake.
The one item worth writing in a diary: from next fiscal year, Microsoft will assume its data centres last 25 years instead of 15 (which spreads the cost over more years and lifts reported profit), and it will sign more leases in a form that legally sits outside the capital-spending line. The money still goes out the door; it just becomes harder to see. As the article puts it, "the spending remains, but the accounting optics improve." The author owns the stock; this is analysis, not a recommendation.
In short: Luria's #1 winner: it "gets the raw end of both" debates (over-spending hyperscaler and doomed software), yet "Microsoft has accelerating growth right now because they're actually executing very well where AI is a tailwind not only to the Azure business but to the office business and the infrastructure software business… and yet they're both trading at these very low multiples." Five years out "I'm still going to… get on Outlook and use Teams and PowerPoint and Excel and Word… guess who is going to stand in front of the model when that happens? Microsoft."
Microsoft is punished twice over by the current AI mood — once for spending heavily on data centres that may not earn a return, and once as a software company AI is supposed to make obsolete. Luria's view is that both fears are misapplied to this particular company, which is why it is his number-one pick.
On the software half: agents will be using Excel, Word, Outlook and Teams, but the human is still there too, and "guess who is going to stand in front of the model when that happens? Microsoft." On the data-centre half: growth is accelerating right now — Azure, Office and the infrastructure software business are all being helped by AI, not hurt. "And yet they're both trading at these very low multiples." Good business, bad narrative, low price.
35:38Salesforce is in the category of software that they're trying to cut. Really? Yes. Why? Because Salesforce has not been adding value to them in years and it keeps charging them more and more for that less value every year. That's a bad business that's been declining. regardless of AI. And then you have Microsoft that has accelerating growth right now because they're actually executing very well where AI is a tailwind not only to the Azure business but to the office business and the infrastructure software business.
In short: Grouped with Alphabet and Apple as former capital-light compounders now spending like industrials — "it wasn't that long ago they were trading at 10 11 times earnings but maybe someday they'll be back there." Also cited (per the host) as asking permission to use cheaper Chinese models over OpenAI/Anthropic.
45:45free cash flow business models capital light that's why they were with great network effects that's why They were such great stocks for so long and Google's been my favorite in the space for a long long time but Microsoft and Apple too and it wasn't that long ago they were trading at 10 11 times earnings but maybe someday they'll be back there given what they're doing.
In short: Wait for sell-the-news. Azure faces a high bar after GCP accelerated from 63% to 82%; the offset is Microsoft 365 Copilot, which benefits from AI cost control because it "operates natively within the ecosystem where enterprise work already happens." The risk is the 27% stake in OpenAI, which Niles remains negative on — "caught between Google in consumer AI and Anthropic in enterprise." Microsoft is also among those adopting open-source Chinese models (token minimization).
Same "wait for the print" stance. Azure has to clear a high bar after Google's cloud accelerated to 82% growth. The positive: as companies try to control AI spending, Microsoft's 365 Copilot is well placed because it lives inside the apps where office work already happens, so it doesn't require a new habit.
The risk is the 27% stake in OpenAI. Dan Niles is negative on OpenAI because it's squeezed from both sides — Google in consumer AI and Anthropic in enterprise AI — and Microsoft owns a quarter of it. Microsoft is also quietly adopting cheaper open-source Chinese models, which tells you where costs are heading.
Full passage: premium transcript (PDF).
In short: A profitable hyperscaler spender, but Azure "tripping over themselves" vs AWS/GCP; incremental ROIC compressing, and an "OpenAI–Microsoft thing coming" (litigation).
Microsoft is one of the profitable "hyperscalers" doing the AI spending — so it has real earnings underneath it, unlike the smaller players. But Chanos notes its Azure cloud has stumbled versus Amazon's AWS and Google's cloud, and that a legal fight with OpenAI is likely coming.
More importantly, it's part of the group whose "return on incremental invested capital" — the profit earned on each new dollar of AI spending — is falling fast (see the Oracle note). Profitable today, but the trend is the worry.
31:52— Yeah. — So this is going back to I think it was like last fall, Satya Nadella, they've been tripping all over themselves with their strategy, and Azure is not the cloud provider that a lot of folks, if you put it up against AWS and even the way GCP is rallying, but he said they were no longer compute constrained.
In short: Demand-side example: Three Mile Island had shut because it wasn't economic, but when Microsoft contracted its power for years "all of a sudden it became economic to do that again."
0:37But I think the difference now is, well, one, we're seeing tech get into it, right? So you see somebody like a Microsoft being connected with Three Mile Island. What the heck is that all about? Well, Microsoft didn't necessarily decide to turn on Three Mile Island again. Three Mile Island was shut down because it wasn't economic to run it.
In short: Named in the mega-cap ramp (up ~3% with Apple, Meta, Amazon, Alphabet) as money rotates back into the trillion-dollar names. No individual call this episode.
In short: Named as the accounting-aggressiveness exemplar, not a rated call. The "massive non-cash gains on investments in loss-generating entities like Open AI" are made "even more aggressive" because those start-ups "are often major customers of companies such as Microsoft" — i.e. a circular, profit-inflating loop (mark up the stake and book the revenue). A caution on the quality of hyperscaler earnings, not a view on the shares.
Microsoft isn't a buy or sell call in this post — it's used as the clearest example of how big-tech profits can be flattered. Microsoft owns a large stake in OpenAI. When OpenAI's paper value rises, Microsoft can book a "non-cash gain" — a profit on its income statement even though no actual cash came in. What makes Hay call this "aggressive" is that OpenAI is also a major customer of Microsoft (it runs on Microsoft's cloud), so money effectively circulates: Microsoft's investment helps fund a customer that then pays Microsoft, and the rising value of that money-losing customer gets booked as profit. None of that is illegal, but it means the reported earnings overstate the real, cash-generating quality of the business. The takeaway isn't "avoid Microsoft" — it's "don't take hyperscaler earnings at face value; check the cash."
In short: The survivor: "taking the right steps," offering low-cost model options (even considering DeepSeek) to protect its enterprise platform. But it will still "get smashed" in a 2000-style downturn (Microsoft fell 60% even as best performer) — stock already −20% YTD; then it becomes a buy.
Microsoft is the name he most respects here: it's "taking the right steps" by offering customers cheaper AI options (even open-source Chinese models like DeepSeek if needed) so it keeps control of its enterprise platform rather than getting undercut. He's confident it survives the bust. But "survives" isn't "spared": in the 2000 crash the NASDAQ fell 83% and Microsoft — the best performer — still dropped 60%. The stock is already down 20% this year. His plan is to buy it after the washout, not now.
24:14But at the end of all of this, I do think that Microsoft, see, Microsoft's taking steps. They're taking the right steps. They're offering their customer base the low-cost options now. They're going to survive here. If they have to, they'll offer DeepSeek. It's the way to survive.
In short: Lebenthal's other "value territory" mega cap (with Meta) — the price-aware, stay-true-to-your-style bargain example. Green on the day. No fresh buy/sell.
In short: "Microsoft fits firmly in the same category as Amazon" — a commoditizer monetizing AI through cloud (Azure) and its in-house platform rather than charging for a flagship model. Part of the hyperscaler bloc Carlson expects to win as models become "cheaper and cheaper and less and less differentiated."
Microsoft "fits firmly in the same category as Amazon" in Carlson's framework — it wants AI models cheap and interchangeable and monetizes AI through its cloud (Azure) and in-house platform rather than by charging for a special model. It's part of the hyperscaler group he expects to win as models commoditize.
16:32And I believe that Microsoft fits firmly in the same category as Amazon. So you could say Amazon or Microsoft here. They're going to be monetizing AI through AWS infrastructure through offering everything in-house as the hub themselves as Amazon already does. So these three companies, Meta, Google, and Amazon, mostly want to monetize through either competitive interchangeable models, distribution, cloud hosting, or all of their massive platforms that they offer.
In short: Added during the quarter — used the "distinct Q2 underperformance of high-quality, strong cash-generating equities as an opportunity to build exposures like we added to Microsoft."
Nothing dramatic here — Singh simply used a window where high-quality, cash-rich companies lagged the market in the second quarter to add to Microsoft. The idea is that when the best "compounders" go on sale because money is chasing hotter AI-chip names, that's the time to build a position in them.
Full passage: premium transcript (PDF).
In short: The industry-math cautionary tale: Xbox is cutting 3,200 jobs (20% of its workforce) and divesting four studios. CEO Asha Sharma told staff the business is "not healthy," running at operating margins three-to-ten times below comparable platform and publishing peers as AAA budgets keep climbing. (Recap, not a stance call.)
Microsoft owns Xbox. The article uses it to show how tough the console-game business has become: Xbox is laying off 3,200 people — a fifth of its staff — and selling off four game studios. Its own new CEO bluntly called the business "not healthy," noting it makes far less profit per dollar (three to ten times less) than comparable rivals. The takeaway for the whole piece: if even Microsoft, with near-limitless resources, is struggling with rising game-development costs, the industry badly needs the extra economics that killing the used-game market provides. A recap, not a call.
In short: Giroux: hyperscaler winner — "the market will be shocked" by how much free cash flow Microsoft/Amazon/Alphabet generate in 2030–32 as custom silicon claws Nvidia's rents back and lower-cost internal LLMs kick in; in-sourcing the neoclouds' business. Black: 22x this year's earnings — reasonably priced for the growth.
In short: Cited for its CEO's WSJ remarks that Microsoft is considering hosting versions of DeepSeek on Copilot as an optional, cost-effective model — operated inside Azure so customer data stays in its cloud — giving customers cheaper choices alongside US proprietary models.
Full passage: premium transcript (PDF).
In short: Rumored (with Meta) to follow Google and Oracle with a capital raise — the hyperscalers have "transitioned from no need for capital to massive need," a group that once "generate[d] so much cash they have trouble figuring out what to do with it all."
Microsoft is rumored (with Meta) to be next in line to raise capital for its AI build-out. Eisman's shorthand: the hyperscalers have gone "from no need for capital to massive need." A company that used to have "trouble figuring out what to do with" all its cash now potentially tapping investors is, to him, the clearest evidence that AI has turned an asset-light software business into an asset-heavy one.
11:21Oracle raised capital and there are rumors that Microsoft and Meta will follow. Why is this important? Historically, these are companies that don't ever raise capital. They generate so much cash that they have trouble figuring out what to do with it all. Now, all of a sudden, the hyperscalers have transitioned from no need for capital to massive need.
In short: Bryn owns it but is the skeptic: "I don't know if Microsoft can execute" — Copilot is "more like Teams" and mediocre, the cloud business is fantastic "but the money they're spending, I just don't see the output." Also down ~32% off its 52-wk high, and the question stands of how much CapEx value is left for the spender. Her easier frame: who are the losers, not the winners.
Microsoft is the committee's example of an AI spender whose payoff is in doubt. Bryn Talkington owns it but is openly skeptical it can "execute": she thinks its Copilot AI assistant is mediocre — "more like Teams" — and that despite a fantastic cloud business, "the money they're spending, I just don't see the output." The stock is also down about 32% from its high.
Her broader framework is that in the murky enterprise-AI layer it's easier to name the likely losers than the winners, and a giant spender with an underwhelming product is a candidate. It's a caution on execution and on paying up for a heavy spender, not a claim the business is broken.
In short: BUY. FV $971.7 vs $390.4 = 59.8% under; ER 16.1%; fwd PE 24.5 against 30.1 (18.6% under); RDCF 8.1% vs 15.0% expected. YTD −17.5%, worse than at the June upgrade.
In short: "The last of the big tech that I believe is uniquely ideal" — below $400, under a 24 PE. Microsoft normally trades a 30–38 trailing PE (nearly 40 in 2024); being sold with the rest to fund semis, it has "plummeted to a 5-year low… at the 23" — the exact 2022-selloff bottom. Analysts still model high-teens EPS growth and Office isn't losing share.
Microsoft is the third of his big-tech July buys. The whole case is valuation: the stock is below $400 and trades at about 23× the last year's earnings. That matters because Microsoft normally trades much richer — usually 30 to 38× earnings, and nearly 40× as recently as 2024. The only other time in the last five years it was this cheap was the 2022 market bottom.
Why is a company this good suddenly this cheap? Because it isn't a semiconductor stock, so it's being sold along with "everything else" to fund the chip buying. Nothing about the business has broken — analysts still expect high-teens earnings growth and its Office software isn't losing ground — so Carlson sees a rare chance to buy a top-tier company at a bargain-bottom price.
16:14The last of the big tech that I believe is uniquely ideal right now is Microsoft. It continues to trade below $400 per share. That puts Microsoft at below a 24 PE ratio. If we look at the current valuation and compare it to its history, this can inform us of what type of situation Microsoft's in. Microsoft typically trades around this range around a 30 plus PE ratio on a trailing basis.
In short: Passing — an analyst flagged a Cerebras appearance at Microsoft Build as a sign its addressable market could open well beyond today's handful of customers. Also a disclosed App Economy Portfolio holding. Referenced/neutral.
In short: AI-disruption fear ("software gets ripped out") is "absurd" — switching costs across Office/Azure are huge; Azure +37% last quarter, now ~22-25x vs 35-40x before. Copilot is clunky but enterprises won't leave.
Microsoft fell to a new low on the fear that AI will let companies "rip out" Office, Word and the rest. Harris thinks that fear is absurd: a business focused on making its own product won't rebuild its entire software stack on some untested newcomer, and there's a huge amount of legal/compliance plumbing baked into incumbent software.
He concedes Microsoft's Copilot AI assistant is clunky today, but giant institutions are too interconnected and risk-averse to switch. Meanwhile its Azure cloud grew 37% in a quarter, and you can now buy the stock at ~22-25x earnings versus 35-40x not long ago — a great business at a much fairer price.
48:40a lot more comfortable with the spend in some ways. Microsoft is it the same? Um no, I I think Microsoft in terms of characteristics, well from a multiple point of view, it's trading. It's fallen a lot over the last little while. I think it made a new low yes a couple of days ago. new 52- week low. Um, I think that everybody thinks that Microsoft is people are going to rip like all all these software
In short: Hohn's #3 at 16.3%; reported, not endorsed. "Microsoft is the tollkeeper of the digital age," with four overlapping moats — network effects in Office 365, switching costs ("companies cannot easily migrate away from Microsoft ecosystems"), scale in Azure against AWS, and recurring subscription revenue. "Whether businesses use cloud services, productivity software, or operating systems, Microsoft is often the unavoidable choice." No Compounding Quality stance is offered, and the big-tech corner is one Slegers has said elsewhere he deliberately avoids.
Microsoft appears here as Hohn's third-largest position, described as "the tollkeeper of the digital age": office software that becomes more useful the more colleagues use it, corporate systems that are painful to migrate away from, a cloud business with the scale to compete with Amazon, and subscription revenue that arrives every month whether or not anyone buys anything new.
The stance is Neutral because this is a report of another investor's reasoning, not a Compounding Quality recommendation. Slegers has been explicit elsewhere that he deliberately avoids the crowded large-cap technology corner where he believes he has no analytical edge, and no valuation work is done on the shares in this piece. Read it as an illustration of the tollkeeper test applied to software, not as an addition to the buy list.
In short: On the Mag-7 pullback the multiples are attractive: Microsoft at 22x with Azure growing 38–40% over three quarters. The caveat she stacks on the group: it's not just margin contraction, it's free-cash-flow burn and no operating leverage during the investment cycle.
After the Mag-7 pullback, Link thinks the prices finally make sense relative to how fast these businesses grow. Microsoft trades at about 22 times earnings while its cloud arm, Azure, has been growing 38–40% — fast growth for a fairly modest multiple. The "multiple" (P/E) is just how many dollars you pay for each dollar of annual profit; paying 22 for that growth rate is the bargain she's pointing at.
Her caution applies to the whole group: the bigger risk isn't shrinking profit margins, it's "free-cash-flow burn" — these firms are spending so heavily building AI data centers that the cash left over after spending can actually go negative this year, and you won't see the usual benefit of revenue growing faster than costs ("operating leverage") until the spending cycle eases.
In short: "Similar to Meta if not weaker" — already almost retesting the spring (April) lows, a red flag, and ugly against a market making new highs.
Microsoft is in "a similar position as Meta, if not weaker." The damning detail: it is already almost back down to its spring (April) lows, even as the broad market makes new highs. Retesting prior lows while everything else is at new highs is a clear sign of relative weakness — the stock is being left behind.
For a technician, an "ugly" chart that can't keep up with a rising market is a red flag, which is why Microsoft lands on the negative side of his ledger despite being one of the largest companies in the index.
10:12MSFT (Microsoft) — similar position as Meta, if not weaker. Already almost retesting the spring (April) lows, which is a red flag. It's ugly compared to a market making a new high.
10:34GOOGL (Google/Alphabet) — different story. Got very extended, super overbought — take some money off the table and revisit; it will likely be viable. Compared on one screen, Google has the best chart of the big names — most similar to GEV (power in the AI of Google).
In short: Down ~35% from highs / worst June since 2008 — Singh would add into the Mag7 de-rate (gradually; Burry covered Palantir and bought Dec-2028 $700 MSFT leaps). Capex-light vs peers, but OpenAI (45% of its cloud backlog) delaying its IPO and Xbox memory-cost hikes are overhangs.
Microsoft has fallen about 35% from its high — its worst June since 2008 — caught in the broad selloff of big tech that's pouring cash into AI. It actually spends less on data centers than its peers, and famed bear Michael Burry just bought long-dated call options on it. Singh would add gradually into this weakness, while watching two overhangs: its biggest cloud customer (OpenAI) is delaying its IPO, and memory costs are forcing Xbox price hikes.
Full passage: premium transcript (PDF).
In short: Weiss owns it but questions it "every day": huge CapEx with unproven returns and an unclear AI strategy (could it be disrupted?). Talkington: not executing — Copilot keeps pivoting (OpenAI→Anthropic→DeepSeek) and "still isn't great." But +5% on the day (worst month since Dec 2000), and Santoli notes 18× is cheap relative to its history.
Microsoft is the episode's poster child for the "is the AI spending worth it?" debate. Weiss owns it but questions it daily: the company is pouring enormous sums into AI infrastructure (CapEx = capital expenditure, spending on data centers and chips) with no clear proof of the payoff, and even its AI strategy keeps shifting — its Copilot assistant has flip-flopped between AI suppliers (OpenAI, then Anthropic, now reportedly DeepSeek), which Talkington reads as a company that "isn't executing." The stock just had its worst month since December 2000 and is down about 25% on the year.
The other side: it bounced 5% on the day, and Santoli notes it's now only 18× earnings — cheap versus its own history. His added insight: the lower the stock falls, the more pressure builds on management to "modulate" (dial back) spending or otherwise reassure investors that the money is being invested profitably. So it sits in the middle — too cheap to write off, but unproven on returns.
In short: Down 3.7% in the selloff but "already a relatively cheap stock based on its growth and its quality." The consumer-inflation proof point: it's raising Xbox console prices $100–150 on memory/storage costs (up >2.5× and "another doubling by fall of 2027").
Microsoft dropped about 3.7% in the broad sell-off, but Carlson notes it's "already a relatively cheap stock based on its growth and its quality." He brings it up mainly as the second proof point of his thesis: Microsoft is raising Xbox console prices by $100–150 because storage and memory costs have more than doubled (and it expects them to roughly double again by late 2027).
That's the same dynamic as Apple — AI-driven hardware inflation reaching consumers, this time through game consoles. It's why he says "the floodgates are opened" and expects prices on many everyday products to start climbing. So the stance here is neutral on Microsoft itself (cheap, high-quality) but the Xbox hike is a flag for the wider consumer-inflation problem he's warning about.
23:43I don't usually do this, but I want to throw this in because I think it's very applicable to my opening monologue and major point in this episode, that the confined AI costs that investors were primarily funding have now broken full well into the consumer economy. We have not only Apple raising prices on all of their Apple devices, we have Microsoft now increasing prices of Xbox consoles due to soaring component costs.
In short: ~17% revenue / ~15% EPS growth. Moat strengthening as Azure/Office/GitHub/security become more embedded with AI tools — the Ackman thesis (the Office value-surplus is so large businesses won't cancel). Rates it slightly below Meta on conviction because Hohn's Claude/Azure-disruption argument has "more potential disruption."
Microsoft keeps getting more deeply woven into how companies operate — Azure cloud, the Office suite, GitHub, security — and adding AI tools makes it even harder to rip out. That's the bet Bill Ackman made: the value businesses get from Office so far exceeds what they pay that they simply won't cancel. Carlson agrees the moat is strengthening (revenue ~17%, earnings ~15%) but rates his conviction just a notch below Meta, because there's a credible argument (from Chris Hohn) that AI assistants could pressure Azure and Microsoft's customer relationships.
In short: The buyer of the West Texas data-center power (the GEV/Chevron PPA) — and also one of the hyperscalers (with Amazon, Meta, Oracle, Google) that were "all down on Thursday" even as chips/power rallied, evidence of the capital-intensity/no-moats rollover.
5:15on Monday. Chevron announced that it has signed a 20-year power purchase agreement with Microsoft to develop a power facility in West Texas that will supply electricity to a Microsoft data center. A majority of the facility's power generation will come from GEV's turbines and electrical infrastructure. This deal is another example of GEV's bundled offering of turbines and electrification equipment.
In short: Net S&P detractor — worst month since Dec 2000; the hyperscaler "footing the bill," an investment cycle that yields no operating leverage. (Lebenthal still owns it as an Oracle-style AI play.)
In short: Passing reference — GitHub Copilot named among the brutally competitive AI-coding field SpaceX/Cursor must fight (alongside Anthropic's Claude Code, OpenAI's Codex, Google).
In short: The offtaker — its West Texas data centers (co-located at the plant site) take the 2.67 GW over 20 years. Stock down 2.1% on the day as tech slipped; mentioned as the contract counterparty, not an investment view.
In short: Comstock's "new neighbor" via its Black Hawk data-center project at Silver Springs — visible adjacent transactions help re-rate Comstock's land on a per-megawatt basis.
In short: The "watershed": moving Copilot to usage-based token pricing and adding a heavily-modified, Azure-self-hosted DeepSeek v4 to cut AI spend — "we do not want a world where enterprises cede all value to a small number of models." Down >20% from highs; sued by a Michigan pension fund alleging it overstated AI cloud-profitability.
Full passage: premium transcript (PDF).
In short: "Microsoft's unchanged for two years" — one of the big-cap "dogs with fleas." Its forward earnings he respects (millions of counterparties), but the stock is dead money while its cash flow is decimated to fund the smaller chip names.
6:24The big stocks are actually, Nvidia's unchanged since October, Microsoft's unchanged for two years. They're real dogs. Like Meta is unchanged for almost two years. Tesla's unchanged since 2021. So there's a lot of dogs, as Gordon Gekko says in Wall Street, dog with fleas.
In short: UPGRADED HOLD → BUY, four days after the superinvestor issue reported Ackman buying it at 21x forward earnings and Hohn cutting it >80%. FV $1,037 vs $416.7 = 59.8% under; ER 16.1%; fwd PE 24.5 against 30.1 (18.6% under); RDCF 10.7% vs 15.0% expected. YTD −12.4%.
In short: Rumored to be doing a similar capital raise; also moved its GitHub Copilot to a token-usage pricing methodology on June 1 — the token-pricing reckoning Eisman is tracking.
10:16Microsoft moved to a similar pricing for its GitHub Copilot on June 1. There is already pushback. Uber went through its entire AI budget for the year in 4 months. Reddit boards are filled with complaints about the new pricing systems. It's possible that customers will start using AI less as they become more cost-conscious.
In short: Cited (with Amazon & Meta) as a 20-year compounder that has cycled through "different cultures, struggles and challenges" — selling on each bad-culture patch would mean selling the lows every time.
16:31Over the past 20 years, all these companies like Microsoft, Amazon, and Meta have gone through different periods of different cultures, different struggles, and different challenges. If you were to sell the stock every time it went into some challenging environment, you would sell it at precisely the lows every single time.
In short: Evidence, not a stance: "MSFT walking away from a $3bn ORCL cloud-leasing deal last night" is cited as a sign datacenter capex is starting to crack — a risk to a big pillar of the copper-bull / datacenter-rollout case.
Microsoft is the software-and-cloud giant behind Azure and a backer of OpenAI. Paulo isn't rating the stock — he flags one data point: Microsoft reportedly walked away from a $3 billion deal to rent datacenter capacity from Oracle. He reads that as an early sign that the enormous spending on AI datacenters ("capex") may be starting to crack. That matters beyond tech: a chunk of the bullish case for copper rests on all that datacenter construction, so if the building boom cools, one of copper's demand pillars weakens.
In short: Passing reference — named as one of the ~50 vetted partners in Anthropic's "Project Glasswing" defensive program for the more powerful Mythos 5 model.
In short: One of the hyperscalers carrying the hidden leverage — "over $500–600 billion of off-balance-sheet debt that the hyperscalers, the Facebooks of the world, the Microsofts… have off the balance sheet" — and one of the names SpaceX's projected revenue would have to come out of.
Microsoft is one of the hyperscalers carrying what he calls the hidden leverage of this cycle — "over $500–600 billion of off-balance-sheet debt." Off-balance-sheet means the borrowing sits in separate financing vehicles rather than on the company's own accounts, so an investor reading the balance sheet doesn't see it. That's the specific echo of Lehman he's pointing at.
On top of that, its cash flow is already being consumed by data-center spending — and if SpaceX's projected revenue is real, another trillion dollars of it gets handed over.
22:37It's over 5 600 billion dollars of off-balance sheet debt that the hyperscalers, the Facebooks of the world, the Microsofts, they have this debt that's off the balance sheet. And to your point, you've got this risk to the data center construction. If you're going to build a thousand data centers in five years, Elon Musk is promising a rail system to space, right? Where he would be the railroad and SpaceX would take and create data centers in space.
In short: Evidence in the AI-economics bifurcation: per Citadel's tokenomics work, Microsoft has begun canceling Claude rollouts as inference costs bite — frontier models concentrate in the biggest balance sheets while everyone else moves to cheaper models.
Full passage: premium transcript (PDF).
In short: Half of the AI disagreement — bought and sold by different superinvestors in the same quarter. Bought: "Bill Ackman's Pershing Square made Microsoft a core holding," in "at 21 times forward earnings," calling it "really cheap" after a post-earnings dip. Sold: "Chris Hohn's TCI Fund Management did the exact opposite. He cut his Microsoft position by more than 80%"; and the Gates trust "sold all his Microsoft stock in the first quarter of 2026." No Compounding Quality view is given here — but the stock was upgraded Hold → Buy on the 18 June list.
Microsoft is the clearest example in the article of professionals disagreeing with each other rather than with the market. Bill Ackman made it a core position at twenty-one times expected profits, calling it "really cheap" after a fall following a results announcement. In the same quarter Chris Hohn cut his holding by more than 80% and put the money into Alphabet, and the trust holding Bill Gates's own shares sold out entirely.
No view of Compounding Quality's own is offered here — but it is worth reading alongside the Buy-Hold-Sell list published four days later, where Microsoft was upgraded from Hold to Buy and Alphabet was downgraded from Buy to Hold: the archive quietly sided with Ackman.
In short: "Perhaps even Microsoft" may have to raise capital — same rotation-away logic as Meta: investors will shift toward AI beneficiaries that fund from cash flow, not capital-raisers.
Same story as Meta: Eisman says investors are rethinking names that "might have to raise capital — like Meta and perhaps even Microsoft." Microsoft is one of the biggest AI spenders, so it's exposed to the same shift.
His call is to rotate out of these capital-raisers and into companies that benefit from AI without needing to keep tapping the market for cash.
8:44At least not yet. Anyway, Oracle's numbers were powerful, especially the backlog. And the news from Google and Oracle is that participating in the LLM agentic AI game is now incredibly capital intensive. These hyperscalers are not going to stop. So I think investors are rethinking their commitments to companies that might have to raise capital like Meta and perhaps even Microsoft.
In short: One of the hyperscalers whose capex-cut announcement could trigger the unwind — "at some point next year you could have Microsoft, Meta or one of the neocloud guys say… we're going to need to cut back."
42:13So at some point next year you could have Microsoft Meta or one of the Neocloud guys say hey these data centers are taking a long time or we can't expand in markets where we want to expand and we're going to need to cut back because right now what Nvidia is doing is because you need like you in the beginning they thought you needed 7 million worth of memory for every Vera Rubin GPU rack you know that they were basically overordering and locking up inventory.
In short: "In the old days, Microsoft came public at less than a billion dollars" — vs companies now coming public at $1.8T, "a colossal failure of common sense."
37:52So passive investors are the bag holders up against really billionaire investors that have been in these IPOs for a decade now and maturing. In the old days, Microsoft came public at less than a billion dollars, right? Facebook came public at a hundred billion, right? And now companies are coming public at 1.8 trillion.
In short: Named as an ERock customer — a hyperscaler using ERock's on-site generators for data-center power.
In short: Named among the largest constituents ("Google and Meta and Microsoft and Amazon and Nvidia") that passive indexes must sell to source the capital for the IPO wave — "that's where the forced selling will be the biggest."
0:19You have ample opportunity to evaluate businesses. So, what one of the things we've been thinking a lot about is Newton's third law uh uh physics, which is for every action there's an equal and an opposite reaction. So, yes, SpaceX is going public, but where is the capital going to come from? And your point on the passive indexes, they're going to have to sell the largest constituents, which make up uh you know, are made up by names like Google and Meta and Microsoft and Amazon and Nvidia.
In short: One of the four hyperscalers; down 27% (bottom of its P/E range; FCF yield only mid-range because of heavy capex). He argues that capex is more monetizable than feared. Buy.
Microsoft is the Windows/Office and Azure-cloud giant, one of the four "hyperscalers" pouring huge sums into AI data centres. The stock has dropped 27% — a big dip — which puts its price-to-earnings at the very bottom of its range.
Its cash-flow yield only looks middling, but that's purely because so much cash is going straight into building AI infrastructure (capex) rather than being kept. Carlson's core argument is that this spending is more likely to pay off than the market fears, so on an earnings basis Microsoft is cheap — a buy.
8:52That's quite a ways from from the highs. A lot of times these companies trade 15 to 20%, but when you're 27% off your highs, that's a pretty big dip. So, it's towards the bottom end of its 52- week range. And on the 5-year historical valuation, Microsoft is very bottom of its PE ratio. But on the free cash flow yield, it's towards the middle because they're spending so much on capex.
In short: Cited as still funding the majority of its AI capex from existing operations — i.e. not yet forced toward debt or an equity raise.
Microsoft makes Windows, Office, and runs the Azure cloud, and is a heavy AI spender too. The article's point about it is reassuring: Microsoft is still paying for most of its AI buildout straight out of the cash its business throws off — no need to borrow or sell new shares yet. That puts it at the calm, self-funded end of the cash-vs-debt-vs-equity choice the piece is built around.
In short: One of his four hyperscalers spending hundreds of billions in AI capex; also rumored to need to raise capital. He's "fully invested and fully exposed" to this group.
Microsoft is the Windows/Office and Azure-cloud giant and the third of his four hyperscalers spending enormous sums on AI infrastructure. He describes himself as "fully invested and fully exposed" to this group, so Microsoft is a core holding he's comfortable owning straight through the current AI jitters.
7:14When I look at my portfolio, I'm heavily invested in artificial intelligence through four hyperscalers. We have Meta, Google, and Microsoft, all spending hundreds of billions of dollars in capex. And in my other portfolio, the Story Fund, we hold Amazon, which is another one spending hundreds of billions of dollars in AI capex.
In short: Rumored to be raising capital; its GitHub has started charging the real cost of tokens (sparking Reddit pushback) — a tell that AI economics are about to be tested as users finally pay full freight.
Microsoft is Eisman's evidence that the "AI is cheap" illusion is ending. "Tokens" are the metered units companies pay for when using AI, and providers have been charging less than the tokens actually cost — effectively subsidizing usage.
Now Microsoft's GitHub has started charging the real cost, and users are pushing back loudly online. The open question he poses: once people actually pay full price for AI, how much will they really use it — and can these businesses thrive? (Microsoft is also rumored to need a capital raise.)
3:58AND THAT'S ACTUALLY STARTING TO CHANGE. MICROSOFT'S GITHUB IS NOW CHARGING ACTUALLY FOR THE TOKENS. AND IF YOU GO ON THE REDDIT BOARDS, YOU'RE SEEING THAT THERE'S A LOT OF PUSHBACK. SO IT'S GOING TO BE A. — QUESTION WHETHER THE BUSINESS CAN SURVIVE. — I DON'T KNOW IF SURVIVE BUT THRIVE. — MAYBE I SHOULD SAY. — IT'S GOING TO BE VERY INTERESTING TO SEE HOW HEAVILY PEOPLE ARE GOING TO BE USING AI ONCE THEY GET ACTUALLY CHARGED FOR THE ACTUAL COST OF THE TOKENS.
In short: Hyperscaler whose combined CapEx (with peers) fell vs Q4; reportedly canceled most of its Claude Code licenses as Copilot/Cursor/Codex close in.
Microsoft is one of the five "hyperscalers" — the cloud giants spending enormous sums building AI data centers. Woo's key warning sign: their combined spending on this gear actually fell in Q1 versus the previous quarter, hinting the AI buildout is cooling.
He also flags a Microsoft-specific clue: it reportedly canceled most of its Claude Code licenses as rival coding tools (its own Copilot, plus Cursor and OpenAI's Codex) caught up — a sign the AI leaders' fast growth may not last.
9:06I see this as being the biggest problem facing Anthropic and ChatGPT ahead of their IPO later this year. For Anthropic, there's another problem. That is competition is closing in. Cursor, GitHub Copilot, OpenAI Codex are all quickly catching up to Claude Code. This is one reason why Microsoft reportedly has canceled most of its Claude Code licenses.
In short: Cited among history's great IPOs ("the Microsofts") that were all a small fraction of GDP — unlike today's mega-IPOs.
9:37This is this is a disaster Danny because right now SpaceX at the two billion uh two sorry two trillion valuation would be almost 6% above 6% of US GDP and if you look at all the great IPOs of all time the Facebooks of the world the Microsofts the Nvidia they're just a fraction you know less than a half a quarter% of GDP and so u you you've just got too much of the market that's going be either tech or u you know traditional tech or AI and it's probably at the end of the day going to be 50% and whenever we've seen this in our career Danny you
In short: Named two ways — Snowflake's Cortex Code / Snowflake Intelligence reach data inside Microsoft apps (widening Snowflake's moat), and Microsoft Fabric is a competitor in the agentic data layer. A disclosed App Economy Portfolio holding.
In short: CEG PPA counterparty — a 20-year power-purchase agreement for the entire 835 MW of the Crane Clean Energy Center (Three Mile Island Unit 1), the restart targeted for 2027 (DOE up to $1B loan). The contract is what de-risks the nuclear-restart revenue.
In short: A phase-3 buyer with "10 minutes" worth of AI monetization paths (Azure, Copilot, GitHub, Office, Security). $70k in the passive-income portfolio plus $21k in the Story Fund.
Microsoft is a classic "phase 3" buyer in Carlson's roadmap — a hyperscaler with, in his words, ten minutes' worth of ways to monetize AI: Azure cloud, Copilot, GitHub, Office, security and enterprise software. He owns $70,000 in his passive-income portfolio and $21,000 more in the Story Fund.
The logic is the same as for Google, Amazon and Meta: it's spending big on AI today (which holds the stock back relative to the hot sellers), but it owns deep customer relationships and will convert that spend into long-lived profit once the scarcity phase ends.
13:06Microsoft will have Azure, Copilot, GitHub, Office, Security, Enterprise Software, and on and on and on. That could go on for like 10 minutes listing the different ways Microsoft can monetize AI. With Google, we've already seen it. We've seen Search, YouTube, Gemini, Cloud, TPUs, Ads, Enterprise AI, and on and on. Even things like their new products, the Fitbit Air is using AI to monetize an existing product.
In short: Named (via Azure) as a hyperscale buyer powering NVIDIA's Hyperscale segment. Also a disclosed App Economy Portfolio holding.
In short: One of the five hyperscalers (GOOG, AMZN, MSFT, ORCL, META) "torching revenue growth and free cashflow on capex for chips and datacenters," with lease/purchase commitments over 100% of forward operating cash flow. Group reference, no individual stance.
In short: Cited as enterprise-software cash now being reinvested into data centers (and the Azure / Three Mile Island restart) — capital flowing into infrastructure/power.
Microsoft is cited the same way: its software business throws off huge cash that's now being reinvested into data centers and power (including the Azure deal to restart the Three Mile Island nuclear plant). It's an illustration of capital flooding into infrastructure and energy, not a stock call.
14:08So, Google search engine, Meta's social media platform, Microsoft's enterprise software business, um whatever cash that legacy business generates, we're going to go and reinvest it and build infrastructure, right? Build data centers. — Mhm. And so now you're seeing trillions of dollars of money that would used to go into these companies and and governments too, by the way, when you add in governments, now those dollars are flowing in to the treasuries of the companies that are focused on infrastructure, upstream commodities.
In short: A customer, and half of the concentration risk: "35% of the revenue comes from two companies, Microsoft and Meta, making Arista vulnerable to changes in their spending." Also cited among the hyperscalers "investing billions into AI infrastructure".
In short: Cited for its April return — +14.7%, "deeper integration of AI across Azure and its product ecosystem" — and named as one of the three cloud platforms ("AWS, Azure, Google Cloud") that models must run on, which is the pick-and-shovel layer. No stance.
In short: Good quarter — revenue 54.5B vs 53.8B est, EPS 4.27 vs 4.05 est (+23% y/y), Azure +39% vs 38% est, strong guidance — yet stock −4% because the Azure beat was "only 1% better than expected" and some wanted more. "On such data, stocks move."
Microsoft actually reported a good quarter — revenue and earnings both beat (earnings +23%), its Azure cloud grew 39%, and guidance was strong. Yet the stock fell 4%, because that 39% was only one percentage point above what analysts expected and some wanted more. Eisman's point is less about Microsoft's business than about how a hot market works: when expectations are sky-high, even a fine quarter can disappoint. "On such data, stocks move."
22:44So, why was the stock down? Well, the 39% Azure revenue growth was only 1% better than expected, and I think some investors were looking for more. On such data, stocks move. Amazon posted a nice beat on earnings per share of 278 versus $1.63 expected. Revenue of 181.5 billion exceeded the estimate of 177.3 billion.
In short: "The Mag 7 is a joke" — Microsoft ~unchanged for nearly 2 years; cash cows turning capital-intensive with ROIC diluted by China's theft.
Microsoft is the software giant. His blunt take: "the Mag 7 is a joke." Microsoft's stock has gone roughly nowhere for nearly two years. These former cash cows are now spending heavily on AI, and he argues the payoff (return on that capital) is being diluted by China copying the technology — so the profits may not justify the spend.
44:29Mag 7 is not at record highs. You know, that's because Microsoft's unchanged I think for almost 2 years. Meta is unchanged for a year and a half. These companies have over Look at Look at Look at the Look at the drawdown at Oracle. These companies have overinvested and now China's picking them apart. — Well, China's Foreign Ministry is denying this.
In short: Down ~28–30% off the highs; with Nvidia it was ~14–15% of the S&P — the top-2 concentration now unwinding (yet S&P only −6% = broadening).
Microsoft is the other half of that top-two S&P concentration (with Nvidia, ~14–15% of the index). It's down roughly 28–30% from its highs.
Strikingly, the S&P overall is only down ~6% even though its two biggest stocks fell that much — proof, he says, of a "broadening": money leaving the giant tech names and spreading into hard-asset stocks instead.
17:51Nvidia's down 19 since our last conversation. So, the thesis is playing out where I'm shocked the S&P's only down 6%. — Why do you think that is? — Because you're seeing a massive rotation and broadening out. So, 14-15% of the index was Microsoft and Nvidia. Microsoft's down 28-30% off the highs. Nvidia's down 19.
In short: Full 15-step deep dive; Total Quality Score 7.9/10; no verdict given. "Microsoft is a global leader with a wide moat. Their competitive advantage is based on high switching costs… Users and businesses find it costly and difficult to move away from its integrated software and services." Segments: Productivity 42.6%, Intelligent Cloud 39.4%, More Personal Computing 18.0%; Windows at a 75% share of personal computing, Azure at 20% of a cloud market growing 16.0% a year to 2033. Economics pass comfortably — 68.6% gross margin, 22.0% ROIC, 34.4% ROE, 39.0% net margin, net cash, 27.2% ten-year EPS growth. Three failures: CAPEX/Sales 27.2% and CAPEX/OCF 51.8% (both pass only after stripping growth capex, giving 2.8% and 5.3%); SBC at 10.3% of net income; and a reverse DCF requiring 12.8% annual FCF growth — "possible but ambitious if you ask me." Valuation: 22.8x forward against a 27.8x ten-year average, with an earnings-growth model return of 11.7% on 12% EPS growth and a multiple assumed to fall to 20.0x.
Microsoft sells three things: the Office software and subscriptions most companies run on (43% of revenue), the Azure cloud computing that other businesses rent instead of owning servers (39%), and Windows licences plus Surface, Xbox and Bing (18%). It also owns 27% of OpenAI, so buying the shares is partly buying ChatGPT.
On almost every quality measure it passes easily — 69% gross margins, 22% returns on the capital it employs, net cash rather than debt, and earnings that grew 27% a year over the past decade. The shares are down 25% from their high and trade at about 23 times next year's earnings against a ten-year average of 28, so on that measure they are cheaper than usual.
Three things do not pass. Capital spending has exploded to 27% of sales because of AI data centres — defensible only if you treat that spending as investment for growth rather than the cost of staying in business, which the analysis does. Stock awards to staff cost shareholders about 10% of profit each year. And working backwards from today's price, the shares already require Microsoft's cash flow to grow 12.8% a year for a decade to deliver a 10% return — described as "possible but ambitious."
The article scores it 7.9 out of 10 and then simply stops. No decision is stated, no price is named at which it would become interesting, and it does not appear on the same day's list of 45 buys. In the archive's own terms that is a pass by omission rather than by argument.
In short: With Nvidia, ~14–15% of the S&P; ~24× sales — owning the index puts ~$150k of every $1M into two crazy-expensive AI names.
Microsoft is one of the two biggest stocks in the S&P 500 (with Nvidia, together ~14–15% of the whole index). It trades at a very rich price relative to its sales.
His worry is concentration: if you just own the index, roughly $150,000 of every $1 million goes into two extremely expensive AI-linked stocks. That's far riskier than people realize, so he'd reduce exposure to these mega-caps.
25:30Well, you're really setting yourself up for failure because the S&P is nothing like it used to be. 10-15 years ago, the top two or three stocks in the S&P were maybe 7% of the index. In recent years, the top two stocks got up to 14-15% of the index, and they're both AI stocks — Microsoft and Nvidia.
In short: The issue's cleanest contradiction: #1 on both aggregate Top Buys tables and simultaneously on the quality-fund trim list. The stated reason for the trims: "Companies like Microsoft, Alphabet, and Interactive Brokers did really well recently. The valuation might become a bit too expensive according to some investors." Also cited as a serial copier — "Microsoft copied Netscape, Lotus, and many more to become dominant in software."
In short: "Microsoft stock is getting destroyed, but that has not stopped the retail army" — a mega-cap breaking down while retail buys the dip in the absence of fear.
In short: Cole/Bill — same as Meta. Grew its capital base a lot on AI investment; because revenue growth wasn't as much as hoped, the market "tore them apart" that day. Bad returns on incremental invested capital — the hyperscaler-capex problem.
24:22Now what was the difference? They were both bad returns on incremental invested capital. — But Meta was growing faster on — they perceive Yeah, they so so — but that doesn't matter in the long run. Revenue growth does not matter. So let's go back to to the.com bubble. Okay. So what happened was and in fact I was sitting at I was a member of Seattle Rotary and uh Microsoft reported earnings on December the 27th of 1999 and the stock soared that day up 6% uh and and became a $500 billion market cap which was the biggest at the
In short: 14× sales (2–3 std devs above its 30-yr mean) — half of every passive boomer's "two AI stocks"; the core of the dark-side-of-passive concentration.
Microsoft is the software giant. Like Nvidia, it's expensive (about 14 times sales, well above its long-run norm) and it's the other half of the "two AI stocks" that a typical index-fund saver unknowingly owns a big chunk of. He sees this passive over-concentration as the core risk.
Full passage: premium transcript (PDF).
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.