In short: Lebenthal bought today and doubled his position (18:32–19:05). He had trimmed half at ~$230 in May's parabolic run; now "what I see here is momentum. It's up over the last month about 19%" with a growing data-center business — "Qualcomm is not just a smartphone chip manufacturer, automotive, Internet of Things and data center. That's what the market's waking up to."
Qualcomm is best known for smartphone chips, but it also sells chips for cars, connected devices and now data centers. Lebenthal sold half near $230 in May and has now bought it back, doubling his position. His reasons: the stock is up 19% in a month, and investors are starting to value it as more than a phone-chip company.
In short: Collaboration with Amazon on custom AI chips and optical networking "designed to create competition with Nvidia" — plus a warrant for 25M Qualcomm shares at $161.26, in the money (stock >$175) and worth "over $4 billion." "Frankly, I'm a bit bewildered by this deal. Not by the partnership, by the warrants" — not circular financing, since no cash changes hands and Qualcomm "does not need any cash anyway." Subscriber answers he quotes: a "strategic customer acquisition cost" securing a ~$60B pipeline and AWS as anchor customer; or payment for Amazon IP, since Qualcomm's XPU expertise "is not established."
Qualcomm makes the chips that connect phones to mobile networks. It is teaming up with Amazon to build custom AI chips for data centers — a way for Amazon to depend less on NVIDIA.
What puzzles Eisman is the sweetener. Qualcomm gave Amazon "warrants": the right to buy 25 million Qualcomm shares at $161.26. With the stock above $175 those rights are already worth over $4 billion. Usually a company hands out something like that in exchange for money, but no cash changes hands now and Qualcomm doesn't need cash. So why give it away?
Two readers offered answers: it is effectively the price of winning Amazon as a flagship customer for a large order book in a market where Qualcomm has no track record, or it pays Amazon for technology Qualcomm lacks. Either way, Qualcomm is giving up a slice of its future share-price gains to buy its way into AI.
16:24Now, since Qualcomm stock is over 175, the warrants are in the money and are worth over $4 billion. Frankly, I'm a bit bewildered by this deal. Not by the partnership, by the warrants. This does not appear to be another version of circular financing. When warrants are issued, no cash changes hands until later and Qualcomm does not need any cash anyway.
In short: The day's mover — up ~5% on an AWS data-centre deal — and a rare three-member agreement. Belski: "this is a structural change for Qualcomm. This is a very good signal that they're getting in business with AWS… given what happened historically with their Apple position, this sets Qualcomm in a very good stance in a space they've been kind of behind on." Lebenthal wants it but not this month: "more progress in the Internet of Things… which de-emphasizes the smartphone business… I'm tempted to buy today, but because September is a heavy month, I'm not going to do it today… over the last year there have been several data center announcements… and it hasn't really gained traction in the share price. So I'm going to wait… that is one heck of a base forming… also a very cheap stock… I will add to it, just not in September." He trimmed half "during the parabola season… around $230," and needs "a price below 196" to offset taxable gains — "so at 176, yes, I'm tempted." Terranova agrees on the direction but shrugs at the move: "they need to diversify away from the smartphones and that's exactly what they're doing… this is not a unique deal… a lot of the hyperscalers like Amazon going out trying to get relationships to secure these customers."
Qualcomm designs the chips inside smartphones. That has been a problem, because its largest customer, Apple, is steadily replacing Qualcomm parts with its own — a shrinking business the company does not control. The news today is a deal with Amazon's cloud arm, AWS, which puts Qualcomm silicon into data centres instead of handsets.
Belski calls that a structural change rather than a good quarter: the company is finally getting paid in a market it had been losing. Terranova is less impressed by the size of the move, pointing out that every cloud provider is signing deals like this one to tie up suppliers.
Lebenthal's position is the most instructive because he separates the analysis from the trade. He likes the chart, the price and the direction, and he is still not buying today — partly because he expects a difficult September, and partly for tax: he sold half his holding at around $230, so he needs to buy back below $196 to leave his taxable clients whole. The stock is at $176, which is why he says he is "tempted."
In short: In the second tier of the broadening: "Intel, Astera Labs, Qualcomm, Nebius, and CoreWeave also made appearances" among the quarter's top buys — accumulation without the multi-fund clustering that TSM, CBRS, MU, AMAT, AMD and STX drew.
In short: Sold. Belsky: "we sold Qualcomm because from a dividend perspective — it's in our dividend growth portfolio — and we like financials a little bit more. So we did a one-for-one switch out of Qualcomm and into Truist." Asked directly whether he likes financials better than chips: "yes, we do. And we like the dividend growth in Truist in particular, and the yield relative to Qualcomm." A dividend-mandate sale rather than a fundamental call on the chip business.
Brian Belsky sold Qualcomm out of his dividend-growth portfolio and put the money one-for-one into Truist. The decision is about the mandate, not the chip business: he wants a higher yield and, more importantly, faster dividend growth — and he prefers financials to semiconductors for that purpose right now. Asked plainly whether he likes financials better than chips: "yes, we do."
In short: Paired with Seagate as a dividend payer with AI exposure that dividend ETFs briefly held. Context only.
21:45So they missed out on some of that. Some of the dividend ETFs were able to get their hands on Seagate or Qualcomm for a little bit. We own Broadcom, which we've owned for years, but it — when we bought that in 2019, it was a 4% yielding tech stock, and the words AI never came out of their mouth just yet.
In short: Q&A: "slightly cheap" and technically oversold (a 15× forward PE at the lows), "but it's actually not very, very cheap relative to historical valuations. If you look at the earnings growth, it's really not as high" — though it does benefit from the AI boom.
Qualcomm looks oversold on the charts and "slightly cheap," but Singh separates a technical bounce from a real bargain. Compared with its own history the valuation isn't unusually low, and the growth rate behind it is modest — nothing like the memory makers or AI-chip suppliers. It benefits from the AI boom at the margin, but it isn't the way he wants to own that theme.
Full passage: premium transcript (PDF).
In short: Diversification on trial. Q3 FY26 revenue (June quarter) −4% Y/Y to $9.9B ($280M beat) and adjusted EPS $2.21 ($0.01 miss); shares sold off on an accelerating Apple step-down and a Q4 EPS guide well under consensus. CFO Akash Palkhiwala reiterated that China OEM handset revenue bottomed in Q3 and returns to double-digit sequential growth in Q4, but handset revenue still fell 20% to $5.1B on the memory crunch — and Qualcomm now expects its modem share on the new iPhone to be "materially lower" than the prior 20% estimate, with the step-down accelerating from Q4 on supply constraints. The other side: automotive surged 61% to $1.6B (FY26 exit run-rate ~$7B) and the two data-center custom-silicon wins start generating revenue in December, feeding a raised FY29 non-handset target of $40B. Q4 guide $9.7–10.5B (midpoint above the ~$10.0B consensus) but EPS $2.05–2.25 under the ~$2.35 consensus — the gap is QCT margins compressed by memory costs, with double-digit price increases across end markets starting September 1 to claw it back. The diversification is real and ahead of plan, but racing a modem business disappearing faster than expected. (Recap, not a stance call.)
Qualcomm has one enormous known problem: Apple is replacing Qualcomm's modem chips with its own, and this quarter Qualcomm admitted the loss is coming faster than it had said — its share of the new iPhone will be "materially lower" than the 20% previously guided. The company's whole answer is diversification, and that part is genuinely working: car chips grew 61%, and two custom data-centre silicon wins start generating revenue in December, letting it raise its 2029 non-phone target to $40 billion. The near-term pinch is the same memory shortage hurting everyone — it compresses Qualcomm's chip margins, which is why next quarter's profit guidance came in well below expectations, and why Qualcomm is pushing through double-digit price increases from September 1. The diversification is ahead of plan; it just has to outrun a modem business disappearing faster than expected. A recap, not a call.
In short: Brown's final-trade pick.
In short: ~70–75% smartphones (plus the pure-profit cellular-IP licensing), now chasing data center: a CPU and 200MW AI-rack deal with Saudi Arabia's Humain (nothing delivered yet) and a just-announced mystery hyperscale ASIC — no part, customer, size or date — that still sent the stock up 70%: "tells you how nuts things are." Analyst day end of June.
Qualcomm makes the chips inside smartphones (processors, modems, Wi-Fi) — about three-quarters of its chip business — plus a lucrative licensing arm that collects a royalty on essentially every 3G/4G/5G phone sold, whether or not it contains Qualcomm chips. The problem: phones are a mature market, so the story is diversification into data centers.
So far that's mostly promises: a CPU and a 200-megawatt AI-rack deal with Humain (Saudi Arabia's AI consortium) with nothing delivered yet, and a just-announced custom AI chip ("ASIC") win where nobody knows the product, the customer, the size, or the timing — and the stock still jumped 70% on the announcement. Rasgon's read: that move "tells you how nuts things are," not that the business has been proven. Analyst day at the end of June should fill in details.
38:14ASIC stands for application specific integrated circuit custom chip. — Okay, — they got some type of of AI ASIC. We don't know what the part is. We don't know who it's selling to. We don't know how big it is. We don't know when it's coming. We don't know anything except they have a a win. But that — but that single announcement actually sent the stock up 70%.
In short: Cole — owned (US tech). One of only two "tech" names they own: good capital structure, produces good returns, cheap. Not caught in the SaaS destruction or the hyperscaler capex.
16:34Do you not own any tech stocks? And what do you see as the ultimate conclusion of the trillions that are being spent to build out AI infrastructure? — Yeah. So we yeah we own Qualcomm in the US portfolio and then uh former lead tech now considered discretionary is eBay. Okay. And we we we've always liked eBay because it's like the the redheaded stepchild of tech.
In short: ROS short candidate — "#6 on the list." A "catchup laggard in the semi/chips/hardware space" that is "sloppy": "if it can't find its footing when semis and chip names are screaming higher like it's 1999… then when?"
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.