Joseph Carlson — The Best Stock To Buy During The "AI Slowdown"
"If AI commoditizes, then Anthropic and OpenAI go to zero… if AI doesn't commoditize… at least he is not in his competitor's prison the same way that Apple's been. Heads, he wins, tails, he wins."
One-line take: The weekend's AI-safety blow-up — a departing Anthropic employee's "annihilation within a decade" warning, then Dario Amodei's essay We Must Pace the Frontier calling for a coordinated, industry-wide, regulated slowdown, with Sam Altman partly joining — knocked AI names like ASML and NVDA lower. Carlson runs two analyses on it. Who wins a coordinated slowdown? META, clearly: it spends the most on AI capex with the least direct AI revenue (Amazon has AWS, Google has Cloud, Microsoft has Azure — they rent capacity out), so a forced pause cuts Meta's spend without touching its core ad monetization — margins and free cash flow improve, the "Zuckerberg will spend endlessly" discount lifts, and a 21 PE on 27% TTM revenue growth rerates. Via a Colossus write-up he adds the heads-I-win-tails-I-win frame: if models commoditize, value flows to Meta's distribution/attention/commerce; if they don't, Meta is inside six months of the frontier with its own model, not in "his competitor's prison" like AAPL. Evidence of execution: the new agentic Muse app (books OpenTable reservations, shops for 10–15 minutes and checks out through Stripe; 4.9 stars / 3,800 ratings), the MSL rebuild in one quarter (the $14.3B Scale AI stake, Alexandr Wang), and a founder-controller's record of pivoting through every platform shift. Why does Dario want it? The unilateral-action test: if his models truly threaten civilization he could slow Anthropic tomorrow with no Congress needed — yet he asks only for a pace that "won't sacrifice commercial advantage." Carlson's read is incentive, not safety: with ~$500B of spending commitments, a promise of profitability and an IPO ahead, Anthropic (and OpenAI) need to cut training spend anyway, and a regulated pause freezes cash-rich followers (Meta, Gemini, xAI) behind them — "the regulatory capture will be complete." Also rebutted: the Manhattan Project analogy (a weapons programme; AI is more like the internet) and the Hugging Face agent incident (a containment failure, not rogue AI), plus the arithmetic that slowing a technology that "could cure most diseases" in 5–10 years has its own body count. Fail of the week: Hunter Biden's laptop memecoin (~$190 → under $2 in minutes, ~80% of buyers lost money) — no ticker, zero sympathy.
1. Stocks & names mentioned
One primary pick (META) argued as the beneficiary of a coordinated AI slowdown, the two frontier labs whose motives he dissects, and the hyperscaler / chip / private names used as comparisons. "View" is Carlson's stance in this conversation. Research: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis.
| Ticker | Name | Research | View | What he said | At |
|---|
| META | Meta Platforms | QT · SA · STK · FA | Positive — top beneficiary of a coordinated AI slowdown | "Meta will be one of, if not the top benefiting company from an overall coordinated AI slowdown": no company spends more on AI capex "with less direct revenue from AI," so a regulated pause cuts spend while ad monetization keeps compounding — "that would likely end up with the stock rerating… Meta currently trades at a 21 PE ratio" on 27% TTM revenue growth, flat all year. Capex cuts would lift margins and FCF and act as "another third party factor reigning him in." And either way Meta wins (Colossus): commoditized AI accrues to its distribution/attention/commerce; non-commoditized AI finds it within 6 months of the frontier with its own model. Execution evidence: the agentic Muse app (4.9 stars / 3,800 ratings), MSL built in a quarter, the $14.3B Scale AI stake. "Meta will be fine either way." | 6:16 |
| GOOGL | Alphabet | QT · SA · STK · FA | Neutral — monetizes directly via Cloud; Gemini "fallen way behind" | Grouped with Amazon and Microsoft as a capex spender that rents capacity out — "We have Google with Google Cloud" — so a slowdown helps it less than Meta. On the model race: "Anthropic and OpenAI lead the rest of the pack. Gemini has fallen way behind. Muse is catching up fast." Named among the companies a coordinated pacing rule would also slow down. | 26:28 |
| MSFT | Microsoft | QT · SA · STK · FA | Neutral — direct AI monetization via Azure | "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. | 5:01 |
| AMZN | Amazon | QT · SA · STK · FA | Neutral — heavy AI spend, but AWS monetizes it | "Amazon spends a lot of money on AI, but they have AWS that directly monetizes AI." The first of the three hyperscaler comparisons in the slowdown screen: high capex is not the tell — high capex without a direct revenue stream is. | 4:40 |
| NVDA | Nvidia | QT · SA · STK · FA | Neutral — the supplier a pause would hurt | Down on the day with ASML as investors priced "a serious slowdown in the frontier of AI." In the motive analysis Nvidia is the cost line the labs want to shrink: "They have to keep paying Nvidia over and over again at a time point where they already have market share dominance." No stance on the shares. | 27:23 |
| ASML | ASML Holding | QT · SA · STK · FA | Neutral — sold off on slowdown fears | "Many of the stocks that are AI stocks like ASML, Nvidia, these type of stocks have gone down today" — the market's first-order reaction to the pacing debate. Reported, not argued; no change to his ASML view. | 2:55 |
| BKNG | Booking Holdings (OpenTable) | QT · SA · STK · FA | Neutral — OpenTable as a Muse connector | Its OpenTable unit is the demo integration he tested inside Meta's Muse: "You can connect Open Table… I asked if it could really do dinner reservations and it says yes." The agent also books restaurants "not even on Open Table with no integration" by filling out the site's own forms — an implicit disintermediation point he does not draw out. No view on Booking. | 13:13 |
| Scale AI | Scale AI (private; Meta stake) | — | Neutral — the talent acquisition behind MSL | "He paid $14.3 billion for stake and scale AI and installed its founder, Alex Wing [Alexandr Wang], as Meta's head of AI," who is "every single day on X… trash talking competitors… He's not slowing down at all." Evidence of how fast a founder-controlled Meta can pivot, not a view on Scale itself. | 18:58 |
| xAI | xAI (private) | — | Neutral — a follower a pause would freeze | One of the challengers that explains why the leaders can't slow unilaterally: "they have Meta's Muse and Gemini nipping at their heels. They have XAI right behind them. So if they slow down for safety reasons and these competitors don't slow down, then the gap… will reduce." | 24:40 |
| Stripe | Stripe (private) | — | Neutral — checkout rail for the Muse shopping agent | Muse searches the web for 10–15 minutes for the best deal, "and then it integrates with Stripe or different checkout systems. So you don't even need to go to the checkout. It'll automatically fill out all your information… buy the item for you and then track it for you." Named as plumbing only. | 14:31 |
| AAPL | Apple | QT · SA · STK · FA | Negative — "in his competitor's prison" on AI | The counter-example in the heads-I-win-tails-I-win frame: if AI doesn't commoditize, Zuckerberg "is not in his competitor's prison the same way that Apple's been. Apple has no AI for themselves. They're just reliant on other people. They are dependent on AI eventually commoditizing." Also the historical toll-keeper Meta escaped — ATT "was a massive threat to Facebook," circumvented by rebuilding targeting with large AI models. | 7:26 |
| Anthropic | Anthropic (private) | — | Negative — the pacing call as regulatory capture ahead of an IPO | The leader asking everyone to slow down. His test: if its models could end civilization, "why don't you just slow down?… you could do this easily without an act of Congress" — but Dario wants a pace that "won't sacrifice commercial advantage." The real motive in his reading: ~half a trillion dollars of spending commitments, investors promised profitability (FT: profitable a second straight quarter), an IPO coming — "We need to show healthy margins, good cash flow" — so a coordinated pause lets it cut training spend while cash-rich Meta is forced to stop catching up. "Even if Dario's concerns about AI are sincere… it's completely irrelevant." Also criticised for doomerism and the Manhattan Project comparison. | 26:11 |
| OpenAI | OpenAI (private) | — | Negative — the other half of the proposed duopoly | Sam Altman partly joins the call ("we could slow down the frontier and be a little bit more cautious"), and OpenAI shares Anthropic's incentive: "Anthropic and OpenAI have a very big incentive to reduce costs. They don't want to keep paying hundreds of billions of dollars for training." If the pacing regime is granted "they will be the duopoly that controls the world and AI, the regulatory capture will be complete. Every other company trying to catch up will be legally prohibited from catching up." Commoditization is its existential risk — "Anthropic and OpenAI go to zero" is his Colossus-sourced (exaggerated) bear case. | 29:39 |
"View" is Joseph Carlson's stance in this video, not a price rating. People are not securities: the departing Anthropic researcher ("Jacob Coxin" in the captions), Dario Amodei, Sam Altman, Alexandr Wang, Shengjia Zhao, Nat Friedman and Daniel Gross appear inside the rows above. The Hunter Biden laptop memecoin (fail of the week) has no issuer or ticker and is covered in the talking points only, following the hub's practice of not giving fail-of-the-week items a row unless they are a company. Hugging Face (named only as the site of the agent incident), Snapchat/Instagram/TikTok/Reels/Threads (Meta history), Gmail/Google Docs (Muse connectors), Colossus (the publication he quotes) and Financial Times (the source for Anthropic's profitability) carry no view and are excluded. Nothing is framed as a short-term trade. Research: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis.
2. Talking points
0:00 An "all-out war against AI," led by the AI companies
- Framing: the anti-AI push is being led "by none other than the top companies themselves, primarily Anthropic" — a whistleblower's extinction warning last week, then Dario "calling for a big slowdown" and "greater government regulation."
- The agenda: the top company to benefit from a slowdown, Dario's real motives, and the Hunter Biden memecoin fail of the week.
1:39 How the hysteria built
- An exaggerated Hugging Face "hack" story, then the departing Anthropic researcher saying there's "a good risk that we'll all be annihilated by AI over the upcoming decade," amplified on Anderson Cooper and other "highly viral interviews."
- "That caused a lot of hysteria" among people who "weren't aware of these risks."
2:21 We Must Pace the Frontier — and the market reaction
- Days later Dario publishes the essay — "interesting timing." Concern "hit a boiling point" over the weekend.
- AI stocks "like ASML, Nvidia… have gone down today," and Sam Altman joins in "to some degree."
3:21 The key word is coordinated
- Dario doesn't want just the two frontier labs to slow — "Daario wants everyone to slow down, no matter really where they are in this race."
- His quoted line: a coordinated pacing strategy would give developers time "without sacrificing commercial advantage" — regulation that slows "Microsoft and Google and Meta and all the rest of them."
4:00 Who benefits: capex without direct AI revenue
- "Meta will be one of, if not the top benefiting company from an overall coordinated AI slowdown."
- The screen: "what company is spending more money on capex… with less direct revenue from AI than Meta? Right now, there really isn't one." AWS, Google Cloud and Azure all rent capacity out; Meta is "the only company that's really spending on it as an indirect way of creating revenue."
- "Meta does not need to be on the bleeding edge to monetize their core assets directly."
5:40 Growing 27%, flat all year — the capex discount
- Revenue +27% TTM, yet "Meta Stock really hasn't been rewarded… flat all year," because investors "can't link [capex] to a direct revenue stream."
- In a slowdown capex tapers "but yet, their monetization would likely not slow down at all" — Meta "has more than enough capacity."
6:16 The rerating case — 21 PE
- "Meta currently trades at a 21 PE ratio… Meta has probably the best setup": growth continues, margins and FCF improve, and the fear of "Zuckerberg's willingness to spend endlessly on AI" eases — "another third party factor reigning him in."
- He hedges the horizon: "appreciated by the market, at least in the short term."
6:57 Colossus: Meta wins whether AI commoditizes or not
- If AI commoditizes (free open-weight models), Anthropic and OpenAI "are worth dramatically less," and value accrues "to the complements of meta… distribution, attention, personalization and commerce."
- If it doesn't, Zuckerberg "is not in his competitor's prison the same way that Apple's been" — he only has to "remain within 6 months of the frontier." "Heads, he wins, tails, he wins."
- So the spending "has reduced risk in Meta," the opposite of the reckless-capex narrative.
8:22 Doomers vs a leader who sells the upside
- Dario and Sam warn of job losses (the claim hasn't shown up — "the unemployment rate has continually gone down") and edge cases like hacking and bioweapons; "there's little focus on the upside."
- Zuckerberg (quoting Colossus): AI "is not in fact God… simply useful and cool." Carlson: "Daario and Sam have a god complex."
10:03 The Manhattan Project is the wrong analogy
- The Manhattan Project "was explicitly a weapons producing project… The only thing that they wanted to create was something that could destroy."
- AI "could more readily be compared to the internet" — a general platform with far more good uses than bad.
11:45 Zuckerberg's plan: cheap, useful, and price-undercutting
- The goal is booking airline tickets, dinner reservations and photo edits — driving the cost of serving users "to zero" and lifting ad revenue.
- Owning data centers and chips that "Anthropic and OpenAI have to pay to rent" lets Meta "undercut them on price"; 3.6B daily users "don't care whether or not the AI model is within 6 months of the frontier."
12:42 Muse, the agentic app — hands-on
- A plushy cream-coloured assistant with connectors (Gmail, calendar, OpenTable); he tested reservations — it books via OpenTable or fills out a restaurant's own web forms, with a final confirmation step.
- Shopping is the big use case: one command, "it works for 10 to 15 minutes" hunting deals and coupon codes, then checks out via Stripe and tracks the order.
14:51 Early traction and the distribution lever
- "The guys at Anthropic and OpenAI are warning about the death of humanity. Meanwhile, Zuckerberg is creating the most relevant products."
- 4.9 stars on 3,800 ratings — "a ways to go" but bullish; Meta can push it to users as it did Threads ("hundreds of millions of daily active users").
15:48 A company that has survived every transition
- Mobile ads (2013), Instagram Stories vs Snapchat (2017), scandals without user loss (2020, the $5B FTC settlement), Reels vs TikTok, ATT circumvented with AI-built targeting, the FTC break-up attempt beaten, a near-$2T record cap — all while refusing to cut Reality Labs.
- "This revisionist history that Mark Zuckerberg is not a good capital allocator… is not only ridiculous, but it's wrong."
18:17 MSL — rebuilt in one quarter
- After Llama 4, a "regime change": FAIR folded into Meta Superintelligence Labs, nine-figure-plus pay packages, Zuckerberg personally cold-calling researchers.
- The $14.3B Scale AI stake with Alexandr Wang as head of AI; hires including ChatGPT co-creator Shengjia Zhao, Nat Friedman and Daniel Gross. "I've never seen a company this big be able to change route that quickly."
- Why it's possible: "He is the owner… He has the controlling shares."
20:32 "Meta will be fine either way" — and why Dario wants it slowed
- With or without a slowdown Meta is "making incredible progress," and "Daario knows companies like Meta are going to soon catch up. And that's why he's trying to slow them down."
- Dario's interview on the resignation: "I agree with Jacob much more than I disagree with him" — two paths, a safe regulated one or "the highly risky one which we're going down right now."
22:22 The disease-cure paradox
- The essay concedes AI "could cure most major diseases in the next 5 to 10 years" before ten pages of risk.
- "You're sacrificing all the people that will die of those diseases if you slow it down" — a pause has a cost measured in lives, too.
23:14 The Hugging Face incident is a containment story
- The agent swarm attacked targets it wasn't asked to — but it was given terminal access, "huge amounts of leeway, and essentially no monitoring."
- "This is a containment issue more than AI being sentient or out of control," yet it's the pretext for "more AI regulation."
25:00 The sincerity test: why not slow down alone?
- The leaders can't slow unilaterally because Muse, Gemini and xAI would close the gap — so Dario wants "the best of both worlds."
- "If you really thought that your models are so powerful… why don't you just slow down?… you could do this easily without an act of Congress."
26:11 Anthropic's real position: the lead, the bills, the IPO
- Anthropic leads, OpenAI is "neck and neck," Gemini "has fallen way behind," Muse is 6+ months back.
- ~Half a trillion dollars of spending commitments, profitability promised (FT: "profitable for the second straight quarter"), an IPO ahead; enterprise already "plugged into Claude Code."
- Phase change: "What they're focused on now is monetization… They want to reduce costs."
28:57 "The perfect solution" — pacing as regulatory capture
- A coordinated pause lets the leaders "bolster up their balance sheet, make their cash flows look better for their upcoming IPOs, while also slowing down all the competitors."
- "They will be the duopoly… the regulatory capture will be complete. Every other company trying to catch up will be legally prohibited from catching up."
- "Even if Daario's concerns about AI are sincere… it's completely irrelevant" — the ask disproportionately benefits Anthropic.
30:46 Cash-short leader vs cash-rich challenger
- "Meta is flush with cash… Zuckerberg can fund the growth of AI endlessly. And Daario knows that."
- So Anthropic, "which is short on cash, needs to slow down anyway," while asking Meta to slow too — "a blatant attempt at creating a duopoly."
31:36 Fail of the week — the Hunter Biden laptop memecoin
- "Are we still doing the crypto meme coins? Apparently, we are." Launched as community-building and charity for those who "lost money on the Trump grift" — "try to profit the exact same way… execute it a little bit worse."
- From ~$190 a token to under $2; ~80% of buyers lost money within minutes; 726 lost $1k–10k, 112 lost $10k–100k, two lost $100k–1M — while 10 wallets made $100k–1M.
- "Zero tears shed by me… you know 100% what you're getting into." Layers of fail: Trump starting the craze, Biden copying it worse, and the buyers.
3. In plain English
A jargon-free summary of the thesis behind each substantive name — what it is and why he holds that view. (Plain-language companion to the table above; renders on each ticker's consolidated page.)
META — Meta Platforms Positive
Meta (Facebook, Instagram, WhatsApp, Threads) is spending enormous sums on AI — chips, data centers, researchers. Investors dislike that because, unlike Amazon, Google or Microsoft, Meta doesn't rent that computing power out to paying customers. Its AI spending only pays off indirectly, through better ad targeting and new products, so the market can't see a clear return and has held the stock flat all year despite revenue growing 27%.
Carlson's point is that the AI-safety debate could flip this into an advantage. If regulators force every big AI company to slow down together, Meta is the one that loses least and gains most: its spending falls, but its ad business — which doesn't need the very best AI model to work — keeps growing. Lower spending means fatter profit margins and more free cash (the money left after paying for the business). At about 21 times earnings, he thinks the stock would be revalued upward.
He also likes that Meta wins either way on the bigger question. If AI models become cheap and interchangeable, companies that own billions of users (Meta) capture the value, while companies that only sell models suffer. If models stay valuable, Meta has built its own and isn't stuck depending on a rival, as Apple is. The new Muse app — an AI assistant that books restaurants and shops for you — and the speed with which Zuckerberg rebuilt Meta's AI lab are his evidence the company can execute.
AAPL — Apple Negative
Apple appears here only as the cautionary comparison. Carlson says Apple "has no AI for themselves" and relies on other companies' models. That puts it in what he calls "his competitor's prison": if AI models stay scarce and valuable, Apple has to pay rivals for a technology central to its products. Apple's best outcome is for AI to become a cheap commodity — a bet on something outside its control. It's a strategic criticism rather than a call on the share price.
Anthropic — Anthropic (private) Negative
Anthropic, maker of Claude, is not listed on the stock market, but its CEO Dario Amodei just called for all leading AI companies to slow development together under government rules. Carlson is negative on the motive, not the technology.
His test is simple: if Amodei truly believes his own models are that dangerous, he can slow Anthropic down tomorrow without asking anyone. Instead he wants a slowdown only if everyone else slows too, so Anthropic keeps its lead. Carlson thinks the real reason is money. Anthropic has reportedly committed around half a trillion dollars to computing, has promised investors profits, and is heading toward a stock-market listing (IPO), where it needs to show healthy margins. A rule forcing everyone to spend less would let Anthropic cut costs while stopping cash-rich rivals like Meta from catching up. Using rules to protect your position from competitors is called "regulatory capture," and that's his charge.
OpenAI — OpenAI (private) Negative
OpenAI, maker of ChatGPT, is Anthropic's closest rival at the frontier, and CEO Sam Altman partly endorsed the slowdown idea. Carlson lumps the two together: both face the same huge training bills and would both benefit from a pause that locks in their lead — "the duopoly that controls the world and AI." Their long-term risk is the opposite scenario: if AI models become cheap and interchangeable, a company whose whole business is selling the model is worth much less.
Summary & timestamps derived from the public YouTube video (transcript in transcript.txt) for personal study. Not investment advice. © Joseph Carlson for source material.