Despite all the big factor whiplash this year, today managed to set a new record in Semis vs. Software with the spread at -10%:

A little context in the SOX vs. IGV performance:

Most sectors/names still sitting above “Leopold Lows” although it’s been a frustrating & choppy month and a half for semi investors:

Read more

Software +1.5% the main beneficiary of the talk so far on rotation and as a “potentially” slower frontier takes some urgency out of the SaaS-disruption bear case. Cyber +4% strongest sector. Internet +50bps also up led by META/GOOGL +2%

If you need to catch on what happened over the weekend, we thought Gavin Baker had a good rundown of the facts

Sam out with a post last night: When we talk about “pacing”, we do not mean “stopping”. Progress has been rapid and will continue to be. But it should be slower than it otherwise could be; interventions like safety cases and monitoring have significant costs.

MSFT CEO Nadella also weighed in:

Trump:

We're leading China in AI. And frankly, I want to keep it that way.

And we can put guardrails, we can do this and that, but I think you have a lot of negative forces… bringing up things that won't happen.

But whoever wins with AI wins. It's an expression that I came up with. It's true.

In Macro land, Oil +3.5% on more Middle East turbulence: ongoing Hormuz turmoil; Houthi territorial conquests in Yeme; and shuttering of the Saudi East-West Pipeline. Yields up 1-2bps across the curve ahead of the Fed meeting on Wed, where market is pricing in 92% chance of a 25bps increase.

Lots to get to, so lets get to it…

HPE: Evercore ISI downgrades to In Line after +192% post-Juniper rerating, says next leg needs margin proof

Evercore ISI says HPE has executed well but valuation now fairly reflects the Juniper upside, with shares at ~13x FY27 FCF and fewer near-term catalysts to drive further multiple expansion; PT stays $65. ISI sees FY27 revenue growth of +13–17% and EPS growth of +16–20%, but says the debate now shifts to Juniper integration, Networking margins and mix quality, with Networking margins still below historical Juniper levels. ISI views AMD Helios as an additional Cloud & AI tailwind, but likely not material until late FY27/FY28.

IREN: JPM upgrades to OW/$65, sees neocloud execution and pricing supporting further ARR upside

Read more

First, Gavin Baker lays out the facts saying the only tangible change is embedded third-party evaluators at OpenAI and Anthropic, not an agreed industrywide pause.

Read more

In Tech, Semis +1% lead the way higher while Internet/Software +50bps. ORCL +6% on a better print while ADBE -5% continues to frustrate any remaining bulls left, with the print not doing anything to lift the AI overhang.

We’ll hit ORCL and ADBE first then move onto the usual.

Let’s get to it…

ORCL +5%: OCI cleared the buyside bar and the ATM is complete. Better print but ROI/’27 funding concerns still weigh

A much better print overall with better IaaS growth and ATM being done the big positives, which is what buyside was focused on. OCI's fourth straight quarter of acceleration: +55%, +66-68%, +84% and +93% across FY26, now +121% — the largest step-up so far and mgmt expects IaaS acceleration through the year. Given better OpenAI sentiment recently and low expectations, positioning had a positive skew going into the print but this was still better than what bulls had hoped for - feedback was generally positive. This continues to be a tough stock, but a better print + continued momentum from OpenAI should help keep the bears away for now. Oct 28th investor day next catalyst

Read more

Day 3 at Citi and GS was busy. We’ll cover snippets below, among a few other items.

Post-close:

ORCL +2.5% solid print with IaaS accelerating and the $20B ATM overhang removed

IaaS growth of 121% Y/Y, or 120% cc, cleared the 116–118% buyside bar , with 850MW of capacity delivered showing actual backlog conversion.. Sw was soft, but stronger operating margins and substantially less cash burn made this a better-quality quarter than the headline revenue upside alone suggests. Biggest relief for investors: ATM is finished, the $90B revenue outlook is intact, and more than $30B of new AI contracts require no incremental change to capital-raising plans. Stock was up +7% but fading a bit on the call…

ORCL Key KPIs vs. Street: Revenue $19.35B vs. $19.14B; IaaS +121% Y/Y (+120% cc) vs. ~117% Street and 116–118% buyside bogey; Cloud rev $11.6B, +62% Y/Y vs. ~$11.5B; op margin 42.1% vs. 40.7%; EPS $1.92 vs. $1.74; FCF ~-$5B vs. ~-$10B. RPO reached $664B, +$26B Q/Q vs. Street modeling only ~$2B of net adds.

ORCL Guide vs. Street: FY27 rev at least $90B vs. $89.6B, with EPS $8.10 vs. $8.07. Q2 total rev growth 30–34% cc vs. 31.9%, Cloud growth 64–70% cc vs. 65.2%, and EPS $1.85–1.93 vs. $1.90.

ADBE -2.5% small headline upside, but still no convincing growth inflection. Unlikely to change bull vs bear debate.

Revenue and EPS came in modestly ahead, but ARR growth slowed to 11.2% from 12.5% last quarter, while 44% operating margins left little to get excited about on earnings leverage. AI-first ARR growth above 150% and Firefly app/credit-pack ARR up 40% Q/Q solid, but the 10.2% full-year ARR growth target still didn’t signal a broader reacceleration. FY rev guide actually moved slightly higher, rather than being cut, although the Q4 outlook offers little evidence of an imminent inflection.

ADBE Key KPIs vs. Street: Revenue $6.76B vs. $6.70B; EPS $6.13 vs. $6.09; ending ARR $27.5B, +11.2% Y/Y; op margin 44.0%, modestly softer than expected. AI-first ARR grew >150% Y/Y, while Firefly app/credit-pack ARR grew ~40% Q/Q.

ADBE Guide: Q4 rev $6.80–6.85B with EPS $6.30–6.35 and ~44% op margin. FY26 rev nudged to $26.576–26.626B and EPS to $24.45–24.50

Let’s get to the good stuff…

AI/SEMIS

Headline after the close: MICROSOFT PLANS TO TRIPLE DATA CENTER CAPACITY TO 38 GIGAWATTS BY 2032, UP FROM 12 GW NOW.

• SWKS +10%: Hey a non-AI semi leads the way on positive read-through from AAPL Duo launch…

  

• NVDA -2.4% as CEO Jensen Huang at Goldman opened with ‘I was right’ on the $3-4tn call, pitched NVIDIA as ‘the world’s first and only growth value stock,’ and reaffirmed 70% growth next year. New: per-system content of ~$18K in Hopper, $25K in Blackwell and ~$40K in Vera Rubin; GB NVLink72 rack shipments up 27% month over month; the Australia 2GW for 2027 sized at $80bn; per-gigawatt math of ~$60bn of cost against ~$50bn a year of rental revenue; and compute turned into an asset-backed, loan-securing asset.

• LRCX -5.6% as CFO Doug Bettinger at Goldman, a day after Citi, again promised no new statements and called the industry ‘fundamentally sold out,’ with leadership ‘more concerned about missing the upside’ than a downturn and 2027 WFE still DRAM-led with foundry closing the gap. New: Lam is ‘deeply engaged already’ with SpaceX’s Terafab, sized as a ‘very high’ opportunity; a $3bn US lab and engineering investment including a new Oregon lab; China WFE ‘slightly up’ this year but shrinking as a share; a company-wide no-M&A stance beyond tuck-ins.

• DELL -5.3% as CFO David Kennedy at Citi called the FY27 guide ‘realistic’ where 90 days ago it was ‘prudent,’ on ~$61bn of Q2 AI server orders and a $95bn non-cancelable backlog. New: customers now negotiate supply access for years two through four (’guarantee supply does not guarantee price’), next year looks tighter than this one with DRAM and NAND the big constraint, OpEx runs 8% of revenue versus 20% six years ago, storage adds $2.5bn this year, and the buyback stepped up from $1.6bn to $3.9bn.

• SPCX +40bps as CFO Bret Johnsen at Goldman said SpaceX has ‘even more conviction’ in the $100bn ARR target for year-end after closing another hosting deal this month worth about $1.11bn a month from Dec 1, with 5-10GW deployed next year versus a little over 2GW at year-end. Hosting deals are 90-day commits with 90-day outs because the company expects its own Grok products to monetize above the $30-50 per-watt range; the first orbital compute satellites fly next year; direct-to-device 5G service turns on in 1H28.

• ANET -2.0% as CFO Chantelle Breithaupt at Citi kept the $12.6bn, 40% guide and named the four signals to judge the model on (P&L growth, the deferred trend, purchase commitments, RPO), with commitments tripled to $9.6bn and already reaching into chips for 2H27. Scale-up carries ‘zero in our guidance and zero revenue’ for 2026 with revenue in 2028; XPO early deployments come 2H27 with CPO about 12 months behind; a third and possibly fourth 10% customer is expected, and the $105bn TAM gets revisited later this year.

• STX -2.7% as CFO Gianluca Romano at Goldman replaced the old 50% incremental-margin frame: incremental gross margin has run ‘70+%’ and ‘for the next few quarters, that is a good way to look at the business.’ New: 80-90% of data-center volume on HAMR within a couple of years; edge pricing changed as high NAND prices let HDD-SSD overlap segments take price; storage is low-to-mid single digits of customer capex, leaving room to raise price ‘for a very, very long period’; a dividend increase is likely after an Oct-Nov review.

Read more

In Tech, Semis -1.7% weaker to start as more AI doom/safety restriction news gets around and Deepseek V4.1 out looking strong & cheap. Software/Internet flat. TMT Mo -2%.

Asia mixed overnight: TPX +0.2%, NKY +0.2%, Hang Seng -1.27%, HSCEI -1.13%, SHCOMP -0.43%, Shenzhen -0.95%, Taiwan TAIEX -0.51%, Korea KOSPI -0.25%

Citi and GS Day 3 today with Jensen at 8:50am PT at GS the main event (fuller schedule below)

We also get ADBE/ORCL EPS tonight. RBLX developer conference starts today.

Let’s get to it…

Citi and GS Conference Lineup today:

Citi Global TMT Conference — Eastern Time

• AI/Semis: DOCN 8:10am (CEO/CFO) · AEIS 8:50am (CEO/CFO) · Crusoe 9:30am (CEO) · DELL 10:10am (CFO) · Infineon 10:50am (CFO) · Tokyo Electron 10:50am (America President/GM) · CSCO 12:35pm (Head of IR) · ANET 1:15pm (CFO/SVP Cloud & AI Networking)

• Software: ROP 8:50am (CEO/CFO) · Databricks 8:50am (SVP Product) · MSFT 9:30am (CMO, AI at Work) · VEEV 10:50am (CFO) · CRWD 11:30am (CFO) · SNOW 12:35pm (CFO)

• Other: T 9:30am (CFO/AT&T Business GM) · AMT 10:10am (CFO) · CHTR 10:50am (CFO) · COIN 1:15pm (CFO)

GS Communacopia + Technology — Pacific Time

Read more

In Tech, semis outperformed as the MS OpenClaw Muse basket back in vogue after a 6 month hiatus with many names below outperforming.

Not sure why AAPL is in there, but the bull case is that more Muse adoption creates demand for the infrastructure around inference, not just the GPUs: a single instruction can trigger browser sessions, API calls, code execution and retries, creating incremental CPU, server DRAM and storage/NAND demand as workflows become longer and more concurrent. NET through Workers, sandboxes and browser infrastructure & AKAM through cloud compute. The compute bulls will say consumption and infra grows multiples of MAUs as agents do more work on each user’s behalf. More thoughts on Muse and impacts on the landscape below in Internet section.

We also got the Duo flip phone launch from AAPL, which is priced at $1,999, with bigger increase in prices for higher storage tiers, helping offset some of the memory prices people are worried about. Siri AI is also launching as well as the same old a new Apple watch.

The Verge got its hands on it and said it feels impressively polished: smooth hinge, near-invisible crease, IP68 durability, and a matte 7.6-inch inner display that handles glare well. The wider, passport-like form factor is better suited to video, reading and multitasking than many rival foldables, while Apple adds fluid screen transitions and useful laptop/tent modes.

Another busy with lots of companies at GS and Citi. Jensen is key guy on the docket tomorrow at GS. We had full recaps in TMTB Slack and included some snippets below for many.

Lots to get to, so let’s get to it…

AI/SEMIS

• MRVL +4.3% as CEO Matt Murphy at Citi put the FY27-28 outlook at $30bn versus the $20bn framed a year ago, with data center at $15-16bn next year and supply flexed up 50%. New: Teralynx switching on track for $1bn-plus, next year’s NPO/CPO revenue framed well above the prior zero-to-$300m range, connectivity growing 70%-plus this year and next, and the Google warrant cast as XPU-attach with other undisclosed customers behind it. Gross margin stays flat-ish next year despite custom doubling.

• LRCX -1.4% as CFO Doug Bettinger at Citi promised nothing new and held CY26 WFE in the low $150bn range, calling clean-room space the industry’s binding constraint through 2027, with 2027 a growth year and 2028 ‘probably a pretty decent year.’ Incremental: served market above 36% of WFE this year, ahead of plan; the ~$8bn-of-WFE-per-$100bn-of-AI-capex rule raised by $1-2bn; NAND wafer starts ending 2026 about 20% below peak; DRAM de-speccing ‘a red herring.’ Gross margin of 52% is a 20-year high.

• STX -2.0% as CFO Gianluca Romano at Citi said the supply-demand gap is widening, with POs covering the next 4-5 quarters and 2-3-year LTAs in which customers keep asking for more volume. New: an exabyte CAGR of at least 25% for 2-3 years on flat units via the HAMR ladder (50TB quals start in about a year), KV cache framed as warm storage whose lower tiers can use HDD with no flash partnership needed, and one high-interest note left before buybacks step up further.

• WDC +1.0% as CFO Kris Sennesael at Citi said revenue is gated by supply, not demand: exabyte demand compounds well above 25% while supply grows only 22-23% between product transitions. Price per TB rose 18-19% y/y last quarter with more room, most of 2027 is under LTA and customers want 2030-31 terms WDC has not signed, and he sees no gross-margin ceiling from the mid-50s. HAMR is in qual with four hyperscalers for 1H27; High-Bandwidth drives sample at five customers, aimed at flash’s KV-cache tier.

• GLW +1.5% as CFO Ed Schlesinger at Citi said the $20bn Springboard run rate lands a quarter early in Q3, Q3 is at the high end or better with Q4 sequentially bigger, and high-teens growth runs into next year. The Verizon long-haul/DCI deal plus Lumen and Zayo bring the $1bn/yr DCI opportunity ‘much sooner’ than end of decade and bigger; photonics is a $10bn CPO/NPO build from zero with first revenue hoped for next year but the inflection 4-5 years out.

• ALAB +4.0% as CEO Jitendra Mohan and CFO Desmond Lynch at Citi said Scorpio X engagements grew from 10 to the high teens with several design wins expected by year-end for 2027 ramps, and Scorpio becomes the largest product family this quarter with X overtaking P. Content per XPU is near $1,000 (from under $100 at IPO). The optical roadmap got dates: aiXscale component sales as first 2027 revenue, NPO initial revenue in 2H27 and a full year in 2028, then 400G CPO in-house.

Read more

META +4.5% on the news. Our working assumption is that Muse keeps gaining traction and becomes a hit, and this is where META’s distribution advantage starts to matter. Once a product works, it can accelerate adoption enormously. We think this moves both ends of the valuation range higher. Our old ~$550 risk case was roughly 15x buyside’s $36 CY27 EPS number; we think the emerging consumer agent opportunity is worth at least a couple of multiple turns, moving that toward $600 (~17x), while opening up a bigger upside case. Zuck outlined to Alex Heath that this is an agent that can help run small businesses and connect directly to META’s ad systems. More importantly, we think this could move META closer to the transaction, rather than just monetizing the discovery that precedes it. In other words, META could double dip here on both ad and transaction fees: for example, a retailer pays META to show you an Instagram ad for shoes. Under the proposed agent model, you could ask Muse to find and buy shoes, and the retailer could pay META a small fee for bringing it the sale. While a product like this was expected this month, we’re impressed with the rollout so far. Still early, but we like what we’re seeing and think today’s +4% should at least hold (unless market weakness throws a wrench in things).

We’re also hearing plenty of frustration from GOOGL -2% investors asking why Google wasn’t out in front of this, with rotation into META potentially adding to the pressure today. This release should only intensify the agentic fears that Instinct has already started stirring up around OTAs, DASH, CART, etc. The risk isn’t necessarily that agents replace the underlying services; it’s that they take over the customer relationship and steer the transaction, forcing platforms to compete harder for the order and potentially give up some of the economics. Check out this conversation on Muse from skinnybundle on TMTB Slack that demonstrates this:

Read more

Astra also picked up plenty of hype over the weekend, although the early consensus is that it’s not yet a broad step-function improvement over Fable 5.1 outside 3D modeling, spatial reasoning and a handful of other use cases. More importantly, investors seem more comfortable with the September path given the lineup and cadence of model releases leading into OAI’s DevDay on September 29 (Grok 4.7, Fable 5.1, META’s Watermelon, etc.). Alongside that, we’re seeing more concrete applications of these models through Grok Bot for workplace tasks and Instinct on the consumer side, with META’s newly launched Muse adding another entrant today. The bullish semi read-through is that you don’t necessarily need a step function in intelligence if better harnesses make that intelligence useful for a much wider range of tasks, which might provide the elusive next great use case after coding that semi bulls have been hoping to see.

The flip side is that the same setup has sw and internet investors more nervous than they’ve been in several months. We’re already seeing Instinct and broader consumer-agent fears spill into internet today, with SHOP -7%, DASH -6%, CART -5%, EXPE/BKNG -8% among others. The concern isn’t that people stop shopping, ordering food or booking trips. It’s that an agent increasingly owns the starting point, compares the options and routes the transaction, potentially weakening incumbent customer relationship & pricing power. Still many questions rather than answers here (eg, the flip side is agents could also bring incumbents incremental volume and lower acquisition costs) but as we saw earlier this year with Claude Code and similar agentic fears, sometimes it’s shoot first ask later in Tech land. Lots of good discussion around agentic read-throughs in this thread here in TMTB Slack.

Busy day today with a full lineup at GS and Citi TMT conferences. We’ve included some quick snippets below on a lot of names, and we posted fuller recaps from the key presentations in TMTB Slack in the new channel #conference-key-quotes. We sent out OpenAI CFO Sarah Friar recap earlier here. Not surprising to anyone, most AI companies sounded good.

Let’s get to it…

Let’s get to the good stuff…

AI/SEMIS

• INTC +9% as Northland upgraded to Buy and Digitimes had an article calling out another 10% CPU price increase.

• AMD +6% getting a nice bid from Computer Use read-through to CPUs (you need more of them) and CFO and CEO were at Citi sounding confident on the MI450/Helios ramp despite the execution chatter, saying 2027 demand and volume are now running above AMD’s original plan, all three anchor customers (Meta, OpenAI, Anthropic) are forecasting more than AMD expected, and supply, not demand, is the constraint. The new items were margin commentary (gross margin in Q4 and through 2027 slightly below Q3’s 56% guide as MI450 ramps, offset by server CPU and embedded mix), a stance that wafer-cost increases get passed through but AMD will not raise prices just to expand margin, and optics detail for the 2H27 accelerator (scale-up domains above 72, copper and near-package optics side by side, not a binary shift). Data center doubling next year, server CPU up 80%-plus in 2H and 70%-plus in 2027, and the $220bn 2030 CPU TAM were all repeats from the Aug 4 earnings call and the Aug 11 KeyBanc forum.

• Optical (LITE +11% / COHR +7%) — Getting the bid off AMD at Citi TMT: the company plans to offer both copper-based and near-package optical connectivity for larger scale-up domains on the 2H27 accelerator, with copper and optical likely coexisting for multiple generations rather than transitioning through a single industry-wide shift. This is the first dated roadmap from a merchant GPU vendor putting optics inside the scale-up domain (where bandwidth per GPU is a multiple of scale-out) rather than just at the switch. It's the customer-side confirmation of what LITE's CEO said last week, that the next 12–18 months are pivotal for optics adoption inside server trays, and it comes on top of AWS bankrolling 1.6T optical connectivity with QCOM.

• AVGO +3.0% as CEO Hock Tan defended the $115bn/$230bn AI revenue path at Goldman and said the binding constraint is power-ready sites rather than silicon, which Broadcom can lock in for 2027. His value framework was the news: one gigawatt generates about $30bn of ARR for a top frontier model against roughly $10bn of running cost, and his token math has roughly $200bn a year of global inference cost against $150bn of model revenue, with frontier models earning at least 75% of the revenue on half the tokens while open-weight players spend about $100bn to earn $30bn, so value goes to where intelligence keeps improving and second best gets nothing. He pushed back on the circularity critique of the Apollo/Blackstone platform (using financing to create demand, not circular), sketched a die roadmap from two dies to four to eight, and previewed a December capital-return decision with record cash, mid-40s free-cash-flow margins next year, and a dividend increase plus buyback as the realistic options.

• BE +9% after getting added to SP500

• AMAT +4.0% as CFO Brice Hill at Citi said every indicator still points up: the rolling eight-quarter forecast from its largest customers has risen every quarter this year, the fab tracker now covers more than 100 factories, leading-edge logic and DRAM sit near 100% utilization, and the ramp is gated by clean-room construction rather than demand. The incremental piece was the DRAM greenfield math, 300-400K wafer starts a month added in each of the next several years at roughly $10bn of equipment per 100K starts versus about a quarter of that for an upgrade, which is why he calls this a much larger investment cycle. He also called HBM de-speccing neutral for equipment demand, since customers are shipping more systems per memory chip, not fewer.

• QCOM +3% after inking a multi-generation deal with AWS to co-develop custom AI inference silicon and 1.6T optical connectivity (leaning on its SerDes/optical DSP IP), and issued Amazon warrants for 25M shares at $161.26, ~$4B all-in. Warrants expire Sept 3, 2036 and vest in tranches tied to commercial milestones, including up to $60B of Qualcomm server chip/tech purchases so it’s a performance-gated stake, not a $4B cash injection. Second hyperscaler anchor after Meta…CFO Akash Palkhiwala was at GS walking investors through the morning's Amazon deal, calling it the execution proof point investors wanted, said it makes the roughly $5bn FY27 data-center target very high confidence, and that the undisclosed second hyperscaler is similarly proceeding, with a path to $15bn in FY29. Elsewhere the CPU TAM has tripled to over $200bn with half moving to Arm, low-end handsets are down low-teens on memory costs while premium is unaffected, and the $10bn auto target was pulled forward again.

• ALAB -7% — Read-through was negative from QCOM/AWS deal: Amazon their biggest customer and QCOM deal targets internal networking bottlenecks with co-developed optical connectivity up to 1.6T, using Qualcomm’s SerDes and optical DSP tech, which is squarely in ALAB’s signal-conditioning/interconnect lane (Aries, Taurus, Scorpio). Amazon already holds ALAB warrants contingent on product purchases reaching $6.5B, but QCOM inking a a parallel $60B warrant had some investors worrying it will lead to some share loss.

• SNDK -0.1% as CFO Luis Visoso told Citi the new business models now cover 50% of bits this fiscal year and two-thirds in fiscal 2028 with roughly 80% gross margin at the floor, and two contracts have already been reopened by customers wanting more volume and longer duration. Supply growth is 100% node-driven with no wafer additions, industry bit growth stays mid-to-high teens, and his line was that there cannot be more demand than supply by definition. The incremental bit was that the KV-cache demand model behind High-Bandwidth Flash has gone up in every recent revision, with products in customers’ hands in 2027; $4.5bn of last quarter’s $5bn free cash flow went to buybacks.

• ANET +0.6% as CTO Ken Duda and CFO Chantelle Breithaupt explained the anatomy of the guide raise at Goldman: supply-chain work plus two quarters of PO visibility by August, and the near-tripled $9.7bn of purchase commitments is a demand signal, not inflation. Scale-across is about 30% of this year’s $3.5bn AI target because customers cannot get enough power in one place, scale-up is white space where ANET has almost no share with ramp in late 2027 and volume in early 2028, and the price increase was curated to memory-heavy bills of material on future POs only, meant to be margin-neutral. On neoclouds, the CFO’s rule is prepayment or walk away, since not all of them make it.

• VRT +3.7% as CEO Gio Albertazzi told Goldman the market is strong and if anything strengthening, with pipelines building, sales cycles accelerating, and demand broadening beyond the US as frontier labs become hyperscalers, while the supply congestion flagged at Q2 is specific to the OneCore and SmartRun converged products. On 800-volt he gave timing: sidecar acceleration from 2H27, native powertrains commercial late 2027 with volumes in 2028, and Vertiv will field the whole range including a solid-state transformer because a one-trick pony would starve. The Utility Innovation Group acquisition puts Vertiv in the behind-the-meter power conversation, and hyperscalers doing their own power and cooling is decade-old behavior where the door never closes.

INTERNET

• RBLX +3.5% the standout after better weekend data and positive read-through from Astra’s capabilities to make cheaper and better games (ie, RBLX wins from its distribution)

• Instinct losers: DASH -6%; SHOP -7.5%; CART -6%

• GOOGL flat. Some good back and forth in Slack on where there roll is in the agentic world:

Cloud CEO Thomas Kurian was also at GS saying cloud is winning new customers about twice as fast as a year ago with $100m-plus deals up 2x, 17 product lines each above $1bn, and customers typically spending 50% more than they commit. The silicon claims were the incremental part: the TPU business is more than twice the next largest hyperscaler’s, AI-server payback is under two years and half that on Google’s own chips, the majority of infrastructure contract value sits in five-year contracts, and TPUs now sell three ways, including outright purchase for customers’ own data centers and a neocloud set up with Blackstone. Gemini Enterprise is in over 90% of the Fortune 100 and AI attach across Cloud customers is 80%.

Read more

🎤 OpenAI — GS Communacopia

Goldman Sachs Communacopia + Technology Conference 2026 · Sarah Friar, CFO

OpenAI CFO Sarah Friar used the session to put numbers on a business she said is now roughly half enterprise. The company entered 2026 at about 60/40 consumer to enterprise, set a goal of 50/50 by year-end, and reached it around mid-year because “the enterprise business has been on a tear”: in July, run-rate revenue grew about 20% month over month and enterprise grew 32% month over month “off a pretty big base.” The usage data behind that: a frontier firm, meaning a top-10% customer, now uses about 8x the tokens per user per week of an average customer, up from about 3x, and OpenAI’s own internal usage runs about 33x, which she offered as a preview of where existing customers are headed. Codex went from roughly 100,000 users at the start of the year to a last disclosed figure of about 25 million, and she said automated coding “is the thing,” with no developer left using AI merely to augment. Customer examples were concrete: Canva’s code is now “100% OpenAI” and its users create billions of images a week on OpenAI image models; Travelers took an AI claims-review assistant from eight states to nationwide in two months. Where customers ask next, in her walk-ins: cyber “very top of mind,” knowledge-work agents, and vertical transformation, with OpenAI investing in chip design, life sciences and financial services and hinting at a financial-services announcement in New York later this week. On the consumer side, with more than a billion weekly users, messages per user rise about 60% in the first six months and use cases double; Plus and Pro tiers are the year-to-date outperformers as users move “from asking to doing”; the ChatGPT and ChatGPT Work tabs merge by year-end; and the ads business hit a $1 billion run rate seven months after launch, “the fastest product to a billion-dollar run rate,” live in 40 countries with WPP and Dentsu on board and no AI-native ad format launched yet. Her framing of that opportunity: “if Google and Meta had a baby, you get ChatGPT,” high-intent search plus memory.

Read more

OpenAI related names up early: Softbank +5% overnight and now up 25%+ in a few days. ORCL +5%. Neoclouds also +2-3%. Semis +1.4% as a group while Software/Internet -50bps early

Busy week: AAPL’s iPhone event is tomorrow. Citi and GS conferences start today with a great line up at both. We get OpenAI CFO Sarah Friar at GS today at 12:30am ET & Jensen at GS on Thursday among a lot of other big ones. Here are times & schedule for the key presenting companies:

Subscribe now

Read more

Let’s get right to the question everyone is talking about over the weekend: Is OpenAI’s Astra enough to get the AI semis vibes (and trade) going again?

We always think it’s better to take a step back and look at the full mosaic instead of just one datapoint, so let’s start elsewhere before coming back around to Astra.

First, we know from HF Prime data that nets have come down significantly. We don’t have Prime brokerage data, but we do have something arguably more scientific: a TMTB Slack poll.

econd, semi charts have begun looking significantly better. NVDA is a stone’s throw from new ATHs, while DELL has shown strong follow-through after its huge beat and raise and is now hitting new highs. Both are good tea leaves and mark a notable change in character from the Q2 earnings season, when semi stocks consistently struggled to follow through to the upside.

At the same time, we saw the opposi…

Read more

Fairly slow day today. Next week we get the Citi and GS TMT Conferences so should be a good and busy week, with the first set of datapoints/mgmt commentary since the Q2 prints.

We’ll cover IOT +13% and ZS -3% earnings first then onto the usual.

Let’s get to it…

IOT +13%: Beat and raise as NNARR re-accelerates & FY27 raise exceeds the Q2 upside with record large customer adds

Read more

The other big item of the day was OpenAI announced release of Astra. We won’t post all the great benchmarks here (they look good), just go on X and your feed will be crowded with them. Here’s some real world feedback and one benchmark:

Yup, momentum has had a huge drawdown:

Let’s get to the good stuff…

AI/SEMIS

• CLS +11.5% as bulls will say AVGO’s $230B FY27/FY28 AI guide de-risked the size and duration of the custom-XPU build cycle, including OpenAI ramping from 1.3GW in FY27 to more than 5GW in FY28. The read-through: CLS is OpenAI’s named Jalapeño partner for board, rack and system integration and scalable production, so a larger OpenAI deployment means materially more potential CLS content. Investors also extrapolated upside from AVGO’s broader multi-GW XPU ecosystem and AI networking growing as fast as XPUs, although AVGO didn’t disclose a new CLS order on the call.

• NVDA +1.8% acting better and getting close to testing ATHs from back in May. If it does clear the May highs, I’m submitting this chart to the Guggenheim.

• AVGO -3% as debates on the stock weren’t settled. Bulls focused on 2 year outlook being de-risked and stock trading at 10x buyside ‘28 numbers while bears said they came away feeling the story has become uncomfortably more complex, saying growth is increasingly dependent on OpenAI and Anthropic rather than Google, where Hock’s commentary was weaker.

Read more

In Tech, AVGO -4% down early as Hock’s F28 guide failed to meaningfully change the debate & SNOW +24% on a big guide that has bulls thinking 40% a quarter earlier than expected. NTAP -9% on a weaker guide and HPE -5% on peak margin commentary. Semis -1% vs. Software +1% early. META +1% as their new Muse Spark has investors saying “frontier” earlier than expected ahead of Watermelon’s release.

Asia mixed overnight: : TPX +0.5%, NKY -0.17%, Hang Seng -0.39%, HSCEI -0.77%, SHCOMP +0.02%, Shenzhen +0.05%, Taiwan TAIEX -0.67%, Korea KOSPI +0.26%. Memory names weaker with Hynix -3%; Samsung -2.5% but Kioxia +1%. Softbank -6%

We’ll hit earnings first, then the usual.

Let’s get to it…

AVGO -4%: F3Q was modestly above Street and F4Q roughly in line, while the first explicit FY28 AI outlook of $230B and >$30 EPS materially reset AVGO’s out-year earnings power, but debates remain alive

F3Q revenue was $29.591B, +85.5% y/y (last q +47.9% y/y), vs. Street $29.443B, +84.6%, with non-GAAP EPS of $3.32 vs. $3.23 and operating margin of 67.9% vs. 66.9%. F4Q revenue guidance of $34.8B, +93% y/y, was essentially in line with Street, as AI semiconductor revenue of $21.7B came in ahead while software and non-AI semis were below.

The key #s on the call was FY27/FY28 AI revenue guidance of $115B/$230B, with FY27 below the roughly $130–150B buyside bogey but FY28 well above Street and supporting >$30 of EPS.

Bulls say AVGO’s two-year AI outlook is now materially de-risked, with the $115B/$230B FY27/FY28 revenue framework supporting $30+ of FY28 EPS and leaving the stock at only roughly 12–13x. They see FY27 as supply-constrained, view FY28 as achievable or beatable as OpenAI and Anthropic scale while arguing AVGO remains inexpensive versus NVDA and MRVL given its growth, margins and software diversification.

Bears came away from the print feeling the story has become uncomfortably more complex. The $115B FY27 outlook fell below $130B+ buyside expectations, the $230B FY28 number may already be embedded, and the growth is increasingly dependent on OpenAI and Anthropic rather than Google, where Hock’s commentary was notably weaker. The remaining debate is whether those customers can finance and absorb the required capacity, whether AVGO can sustain $20–30B of revenue per GW (Although Hock talked those #s down in the callback) and premium margins as deployment scales, and whether rising customer concentration and execution risk justify giving the stock a higher multiple in a still-poor AI tape.

Read more

Busy day after the close with earnings so let’s get right to it…

SNOW +20%: Another 5% product-revenue beat plus an accelerating guide clears high bar for the most crowded name in sw

Product revenue grew 37%, ahead of both Street and buyside expectations, while the Q3 guide calls for further acceleration and FY27 growth was raised five points to 36%. With this q3 guide, you now very likely get 40%+ growth in q3 a quarter earlier than even bulls expected. This is an unusually clean print: core demand remains strong, AI revenue stepped up meaningfully, and management passed substantially more than the quarterly upside into the annual outlook. Bulls in full control here. This explaisn the CoCO + core magnitude of the beat well (from our weekly on Sunday):

Key KPIs vs. Street

• F2Q product revenue: $1.49B, +37% y/y vs. $1.42B, +31% and bogeys at 36%

• F3Q product-revenue guide: $1.591B, +37.5% vs. $1.50B, +29% and bogeys of 32-33%

• FY27 product-revenue growth: 36% vs. 31% prior

HPE -1%: A large forward revenue guide, but DELL remains the cleaner enterprise-hardware play

HPE beat the July quarter, guided next-quarter revenue roughly 11% above Street and raised both its FY26 and FY27 outlooks, alongside an expanded Oracle networking partnership involving warrants. The standalone numbers are strong, especially revs and cash flow, but after HPE’s run and DELL’s stronger print, the bar had been set a bit higher, but otherwise solid q

Key KPIs vs. Street

• July revenue: $12.2B vs. $11.95B

• Gross margin: 40.4% vs. 36.0%

• EPS: $1.11 vs. $0.93

• Free cash flow: $1.60B vs. $0.93B

• F4Q revenue guide: $14.35B vs. $12.88B

• F4Q EPS guide: $1.17 vs. $1.07

• FY26 EPS: $3.75–$3.85 vs. $3.35–$3.45 prior

• FY27 revenue growth: 15% vs. 10% prior

• FY27 EPS growth: 18% vs. 14% prior

AVGO -4%: AI revenue beat, but the next-quarter outlook is too close to consensus for the current semiconductor bar

AVGO delivered modest upside in F3Q, including better AI semiconductor revenue, but Q4 revenue and margins were roughly in line to slightly below Street and the $21.7B AI guide fell short of the >$22B buyside bar. The quarter is fine in absolute terms but underwhelming relative to NVDA and elevated AI expectations; the call now needs a materially stronger FY27 AI outlook, with investors looking for roughly $130B or more, and Hock will do his best to focus investors on FY28 as best he can. Doesn’t look thesis changing in any way so far and the 27 vs 28 debate will continue

Key KPIs vs. Street

• F3Q revenue: $29.6B vs. Street around $29.3B

• EPS: $3.32 vs. $3.22

• Operating margin: 67.9% vs. 66.7%

• AI semiconductor revenue: $16.7B vs. $15.9B

• F4Q revenue guide: $34.8B vs. Street around $34.7B–$35.0B

• F4Q operating-margin guide: 66.0% vs. roughly 66.5%–67.0%

• F4Q AI semiconductor revenue: $21.7B vs. $21.3B Street and >$22B buyside expectations

AI/SEMIS

DELL +16% the standout showing the big beats can still be rewarded in AI semi land. CRDO -20% wasn’t so lucky on a roughly inline beat. NVDA +3% the other one where got lot of questions on, but no great answers. Otherwise fairly slow day here.

Since we didn’t get around to CRDO this morning, here is the bull vs. bear case:

Read more

In Tech, DELL +9% and GTLB +23% delivered the goods while MDB -13% & CRDO -11% missed elevated expectations. PANW -2% the most debated overnight among investors. We’ll dive into DELL GTLB MDB and PANW then move onto the usual. We’ll put some CRDO Bull vs. Bear Takes in the Slack as soon as we’re done here. Semis -70bps and Software 40bps both down early.

Asia weaker overnight: TPX -2.4%, NKY -2.85%, Hang Seng -0.07%, HSCEI -0.15%, SHCOMP -0.97%, Shenzhen -1.45%, Taiwan TAIEX -1.67%, Korea KOSPI -3.99%. Softbank -6%; Hynix/Samsung -3%

Let’s get straight to it…

DELL +8%: Another big rev and EPS guidance raise; AI pulls the whole stack as $61B of orders, storage-led margins and a $25B FY27 raise

AI-server revenue $74B, above bogeys of $70B and street in mid 60s. FQ2 revenue/EPS of $46.97B/$7.04 came in above Street at $44.93B/$4.91, while FQ3 and FY27 midpoint guides of $49.0B/$6.50 and $192B/$25.50 were far above Street at $41.36B/$4.47 and $174.13B/$18.95 and bogeys of low 20s (buyside was already sitting at $25 for the year)

AI server revenue of $16.4B was only modestly above Street, but traditional servers grew 122% and storage grew 26%, driving ISG revenue up 89% and ISG OM to 15%. AI orders jumped to $60.9B and backlog to $95B, while the five-quarter pipeline still grew sequentially and remains multiples of backlog after $131.7B of orders over the past year

Overall, not much to nitpick at the results as bulls come away happy with the huge beat, but bears will question how much of that broader growth and margin structure survives once component inflation, supply scarcity and initial enterprise refresh normalize.

Key Takeaways:

• Callback. The most important callback clarification was that 15% ISG OM should not be annualized. Mgmt expects approximately 13.5% in FQ3 and modest sequential improvement in FQ4, while AI server OM should remain in the mid-single digits through upcoming platform transitions. CSG OM is expected to normalize toward roughly 6% in FQ3 and 5.5% in FQ4, supporting a structurally better margin framework, but tempering the most aggressive extrapolations from FQ2.

• Enterprise. DELL’s AI customer count surpassed 6,500, up from roughly 5,000 last quarter. Of those, 3,300 were added during the past three quarters, versus eight quarters to acquire the first 3,200. Enterprise customer count, repeat buyers, revenue and pipeline all increased sequentially and y/y. However, mgmt explicitly said the overall mix did not necessarily change because sovereign and large-neocloud wins also remained strong, so concentration risk is improving but not disappearing. Enterprise customers also tend to attach more storage and networking.

• Servers. Traditional server and networking revenue reached $10.53B, +122% y/y versus Street around $9.3B, +~96%, and FY27 growth is now expected to exceed 100% versus the prior outlook of just over 60%. Most demand is coming from DELL’s historical enterprise base, supported by data-center modernization, higher core and memory content, security and resiliency requirements, and incremental CPU demand for agentic workloads. DELL gained more than 10 points of traditional-server share over the past two quarters.

• Refresh. Approximately 1.2M installed assets remain on 14G or older platforms. Mgmt sees 6–8:1 consolidation ratios on 17G and 12–14:1 on the upcoming 18G platform, with post-quantum cryptography and heightened security becoming additional forcing functions. This provides a concrete multi-year refresh argument beyond near-term memory pricing.

• Storage. Storage revenue was $4.85B, +25.8% y/y versus Street around $4.3B, +~11.5%, and FY27 growth is now expected in the mid-teens. Dell-IP demand has grown ahead of the market for six consecutive quarters, with mix shifting away from lower-margin partner IP toward higher-margin proprietary products. AI is adding demand from unstructured data, agents, logs, traces and KV cache, but Lightning remains early and is still in beta at several customers, meaning the current strength is primarily coming from the established Dell-IP portfolio rather than a fully mature AI-storage product cycle.

• Margins. Gross margin was 21.1% versus Street 17.4%, total OM was 12.6% versus Street 9.3%, and ISG OM was 15.0% versus Street approximately 11.0%. Scale and operating leverage were the largest contributors, followed by higher Dell-IP storage mix, pricing discipline and favorable product, customer and geographic mix. Mgmt cautioned that every FQ2 benefit will not persist at the same level, but also argued that the multi-year modernization program has structurally lowered the expense base.

• Pricing. Mgmt explicitly acknowledged that inflation and input-cost increases are contributing to revenue growth, but did not quantify the price-volume split. One detailed decomposition estimates that traditional-server growth included approximately 25% more CPU cores, 40% like-for-like pricing primarily from memory inflation, and 27% more non-CPU content per core. The magnitude of the pricing contribution remains one of the most important questions around revenue quality and the eventual growth comparison when memory prices stabilize.

• Supply. Demand continues to exceed available supply across DRAM, NAND, CPUs, disk drives, optical components, substrates and power equipment. DELL attributed the $25B FY27 revenue increase partly to better allocation and configuration of available components, including redirecting supply from a softer second-half CSG environment toward ISG. Large customers are ordering further in advance and sharing longer-term requirements to secure supply. This creates additional shipment upside if supply improves, but also raises the eventual pull-forward and double-ordering debate.

• CSG. CSG revenue was $15.03B, +20.2% y/y versus Street $14.99B, +~19.9%; commercial grew 22% and consumer grew 7%. FY27 CSG growth is now expected in the mid-teens, but the setup is less clean than ISG: large-enterprise refresh activity remains healthy, while more cost-sensitive customers are extending replacement cycles and FQ3 OM is expected to fall toward 6% as DELL balances units, pricing and share.

Bull vs. Bear Debate

Bulls believe DELL is evolving from a cyclical hardware vendor into the scaled merchant platform for enterprise and neocloud AI infrastructure. The $95B AI backlog alone exceeds one-half of the new FY27 revenue guide, while the more important five-quarter pipeline continued to expand after $131.7B of orders were booked over the past year. The customer base is also broadening: enterprise customer additions have accelerated sharply, repeat purchases are increasing, and enterprise buyers attach more storage and networking than large standalone neocloud deployments. That expands both the addressable market and the margin pool.

The broader portfolio is central to the bull case. AI is stimulating not only GPU-server demand, but also traditional CPU compute for agentic orchestration, storage for logs, context and KV cache, networking, services and financing. Traditional servers growing 122%, storage growing 26% and ISG OM reaching 15% materially weaken the old bear argument that DELL’s AI exposure is merely high-revenue, mid-single-digit-margin server assembly. Dell-IP storage mix, more complex rack-scale deployments and greater enterprise attachment create paths to structurally higher gross profit dollars even if standalone AI server margins remain modest.

Bulls also see durable execution advantages. Some AI engagements require more than 50 unique designs across workload performance, power, cooling and the data-center environment. DELL combines engineering, supply-chain scale, global deployment and support, financing, and the ability to move customers from initial design to first token. Those capabilities matter more as systems transition from individual GPU trays toward full rack-scale solutions. This quarter’s share gains, ability to redirect constrained components and position as an early Vera Rubin rack supplier reinforce that differentiation.

Bears argue that headline growth overstates underlying unit demand. Mgmt acknowledged that systems cost more than they did in prior quarters and that inflation is embedded in growth. Roughly 40 percentage points of traditional-server growth is bc of like-for-like pricing, in addition to richer memory and storage configurations. If memory prices flatten or fall, DELL could face slower revenue growth even if unit demand remains healthy. The extraordinary growth comparisons may therefore reflect a mix of true modernization, richer configurations, price increases and customer pull-forward rather than a clean doubling of underlying compute demand, and eventually leading to much tougher comps in ‘28.

The order and backlog figures also require judgment. Supply-constrained customers are ordering earlier and sharing longer-duration forecasts to secure components. That provides visibility, but it can create double ordering, project timing risk and eventual digestion. Large dollar deployments remain concentrated among a relatively small group of neocloud and sovereign customers, whose financing capacity and project schedules may be less durable than those of hyperscalers. DFS helps DELL close and finance deployments, but rising financing receivables add working-capital and credit exposure. The customer-count diversification is encouraging, yet mgmt explicitly said the overall mix did not necessarily change.

Margins are the second major bear argument. FQ2’s 15% ISG OM benefited from unusually favorable scale, storage mix, pricing, geography, customer and product mix, and mgmt does not expect all of those benefits to recur. The callback’s approximately 13.5% FQ3 ISG margin and more normalized CSG margins show that FQ2 is not the new quarterly run rate. Meanwhile, AI servers remain a mid-single-digit-margin product, and the mix of lower-margin AI revenue continues to expand. Reported FCF of approximately $1B was also far below adjusted FCF because of DFS dynamics, while inventory rose 41% q/q. Bears will want evidence that earnings growth converts consistently into reported cash rather than financing receivables and inventory.

Longer term, the risk is that FY27 represents a peak growth and pricing year. Traditional-server demand could normalize after the refresh wave, storage attachment may develop more slowly than expected, PC units may remain weak, and public-cloud or ODM alternatives could pressure pricing. DELL’s multiple has already expanded well above the roughly 10x historical forward P/E it used to trade, mean any slowdown in growth or margin disappointment re-rates it back.

MDB: 30% total growth was not enough; 29% Atlas, weaker Q3 guide, and an implied F4Q slowdown drove the selloff

Sentiment & positioning had sneakily increased into the print and F2Q27 was strong across revenue, EPS and margins, & FY27 was raised by more than the quarterly upside, but main KPI 29% Atlas growth fell short of buyside bogeys closer to 30%+ and guide implies Q4 Atlas growth of only 24% in Q4 as AI remains too small to drive an immediate acceleration.

Bears will say no accel and slower growth than DDOG/SNOW at similar multiple and expectations were high heading in. Bulls will say the 26% Q3 Atlas guide is conservative and likely comes in near 29-30%, which means no signs of decel and keeps accel hopes alive while at the same time giving some breathing room for AI revs to finally pick up. Overall, q unlikely to have bulls jumping ship but stock likely needs some digestion after recent run.

The #s:

Revenue: $771.8M, +30.5% y/y (last q +25.2%) vs. Street $733–735M, +~24%.

Atlas / EA: Atlas revenue was $565.9M, +28.9% y/y (last q +29.4%) vs. Street $553.4M, +26.1%, but below the 30%+ buyside bogey; EA & Other subscription revenue was $181.2M, +35.9% vs. Street $158.2M, +18.6%.

Non-GAAP EPS was $1.90 vs. Street $1.61–1.62, while non-GAAP operating margin reached 24.1% vs. Street 21.1%.

Guide: F3Q27 revenue midpoint is $758.5M, +20.7% y/y vs. Street $747M, +18.9%; FY27 midpoint is $3.01B, +22.2% vs. Street $2.967B, +20.4%. FY27 Atlas growth was raised to ~27% vs. Street 24.5%, although that implies F4Q Atlas growth of only ~24%.

Key Takeways

Read more

In Tech, a lot of red with Semis/Software both down 2% and Internet -1%. Broad low liquidity de-grossing day: everything red except Telecom (+0.3%), with the damage concentrated in the high multiple growth: Security SW -5.3% (CRWD/S/NET/ZS/FTNT all -5% to -7%), Expensive Software -4.3%, Optical -3.9% (CRDO -8.6%, CIEN -6%, LITE -5%, ALAB -6% all underperformed), Neoclouds/Miners -3.6%, plus the EDA pair (CDNS -7.6%, SNPS -5.6%). Bitcoin Related was the worst factor at -6% (COIN, CRCL) and took fintech down with it (AFRM, XYZ, UPST). BTC -2%.

Post-close Earnings:

MDB -13%: Sentiment & positioning had sneakily increased into the print and the the ~13% selloff is about Atlas, not the headline revenue beat: 29% growth fell short of the 30%+ buyside hurdle. Q3 Atlas guidance, potentially coming on the call, will be key focus where buyside wants to see at least 27%

Q2 revenue: $772M vs. $735M Street, 5% ahead; Atlas growth +29%.
Q3 revenue guide: $758.5M vs. $745M Street, 1.8% above.
FY revenue guide: $3.01B from $2.94B, a $70M increase versus $37M of Q2 upside.

DELL +10% looks really good. Another massive guidance raise, with FY $25.50 vs. $17.90 taken up past where buyside was at $25 and comfortably above the low-$20s buyside bar. Revenue $192B, up 15% from prior guidance; AI-server revenue $74B, above bogeys of $70B and street in mid 60s.

PANW +4% very solid despite high expectations as near-term security demand continues unabated and enough for bulls. NGS cleared bogeys. NGS ARR +63% vs. +59–60% guided, adding roughly $970M of net new ARR vs bogeys of low 900s. Guided NGS ARR of $9.54–9.56B (+63%) vs. $9.2B (+57%) Street; total revenue growth of 33–34%, versus roughly 14% a year earlier.

Let’s get to the good stuff…

AI/SEMIS

Fable 5.1 out today and benchmarks look really good:

Good pod from Alex Heath with Sam Altman. Key Quote:

“We delayed a frontier RL training run. We definitely have not slowed down, paused or delayed all training. This is specifically about frontier RL runs, where we think the biggest risk surface currently is. The model that caused the Hugging Face incident is, in AI-adjusted time, a relatively old and much weaker model.”

More…

1. Compute: demand keeps absorbing efficiency gains, but neocloud overbuilding is becoming speculative

Sam remains confident OpenAI can fund and profitably use its committed capacity, but sees new neoclouds announcing enormous buildouts without customers or supporting revenue. He warns that investing regardless of cost could produce poor returns if better chips and infrastructure sharply reduce compute costs. OpenAI is not immune to a broader downturn, but he distinguishes that risk from concerns about its own commitments.

Sam says efficiency savings repeatedly get consumed by rising token demand rather than relieve capacity constraints. Bringing broader adoption anywhere near today’s heaviest users, alongside larger and more capable models, requires substantially more abundant and cheaper compute.

Read more

Oil +2% as two oil supertankers struck in SOH as Iran’s president said country was seill seeking a negotiated settlement to the war. US long end yields +3-4bps as global yields continue to rise as well.

In Tech, Semis/Software both down 1% and TMT Momentum roughly flat. Memory and Optical -2% weakest.

Overnight, Asia was mixed: TPX +0.62%, NKY -0.15%, Hang Seng -0.93%, HSCEI -0.59%, SHCOMP -0.16%, Shenzhen -0.67%, Taiwan TAIEX +1.78%, Korea KOSPI +0.23%.

Slow day so far but we get DELL, PANW, GTLB, CRDO, and MDB tonight on the earnings front and CIEN tomorrow morning,

Let’s get to it…

Anthropic/NVDA: Anthropic Signs $35 Billion Cloud Deal Backed by Nvidia - WSJ

WSJ says Anthropic signed a $35B cloud deal with Nvidia-backed Lambda, with Nvidia itself leasing the underlying Texas data-center capacity from Hut 8. The structure highlights Nvidia’s expanding role as a financier/credit backstop for AI infrastructure, helping non-investment-grade customers like Anthropic and neoclouds like Lambda secure enormous amounts of compute.

OpenAI: announced its new ad model had reached $1B in ARR after 200 days of launch (out yesterday)

AKAM: Piper Sandler upgrades to OW, PT to $125 on Compute inflection and improving unit economics

Piper Sandler says Akamai’s Compute business is approaching a material growth inflection, with ~$2.8B of commitments already signed and capacity potentially rising from ~47MW today to ~242MW over time. Piper estimates Compute cash gross margins near ~50% today expanding toward ~62% by 2028, supporting attractive long-term economics despite higher near-term capex. Piper also sees AppSec reaccelerating, pricing helping security/web delivery, and estimates conservative enough to support ~10%+ growth as Akamai shifts toward longer-duration Compute contracts.

HOOD: Morgan Stanley upgrades to OW, PT to $150 from $124 on broader platform and more durable growth

Morgan Stanley says HOOD’s expanding product set is improving customer economics, with assets per customer +23% y/y and Gold users holding ~4.2x the AUC of average users, reducing “graduation” risk as customers get wealthier. MS now sees revenue compounding ~23% through 2028 to ~$8B, ~6% above consensus, as prediction markets, retirement, banking, advisory and Rothera broaden monetization beyond trading. MS also sees operating leverage supporting ~28% EPS CAGR through 2028 and raises PT to $150, implying ~43% upside.

DUOL +6%: Evercore ISI upgrades to Outperform, PT to $210 from $105 on reaccelerating engagement and major estimate upside

Evercore ISI says proprietary survey work, product improvements and 3Q tracking all point to stronger engagement, with DAU growth accelerating toward ~24% y/y in 3Q and user satisfaction/daily frequency rising. ISI says its 2028 revenue/EPS estimates sit ~17%/~22% above Street, with the biggest upside coming from higher operating margins as DUOL approaches its 100M DAU target. ISI also argues ChatGPT is more complementary than disruptive, with >50% of ChatGPT language learners also using Duolingo.

DUOL: D.A. Davidson raises PT to $175 from $160 as August DAU growth finishes strong

D.A. Davidson says August DAUs increased ~1.3% m/m, with recent weekly trends accelerating and 3Q DAU growth now tracking around ~27% y/y. D.A. Davidson notes comparisons ease after mid-August and sees further upside if September momentum holds, with its base case implying ~25.6–27.6% y/y growth. D.A. Davidson reiterates Buy and raises PT to $175, citing stronger engagement trends and improving growth visibility.

APP: Evercore ISI cuts PT to $510 from $630, sees solid Q3 but resets Q4/FY27 lower

Read more

Let’s get right to the good stuff…

AI/SEMIS

• SNDK choppy trading: +4% → -5% → finishing +6% after a 9% rally in the last 30 minutes. A perfectly normal day when the order book is three interns and a pair algos:

NAND Spot continues to hit new highs:

• SKHY +2% / MU +3%: UBS Asia was out raising DRAM estimates, saying they see DRAM supply-demand gap widening further in 2027, with blended ASPs +44% y/y and the sufficiency ratio falling to -6% even assuming inventory digestion, versus -12% without it. UBS says unconstrained DRAM bit demand could exceed +100% while supply grows only ~40%, with server DDR particularly strong and HBM demand still seen +91% y/y despite de-spec concerns.

• GEV -1.5% on news Elon was going to begin producing turbines

• MRVL -2% as NVDA’s $3.5B convert investment in MediaTek and expanded NVLink Fusion partnership likely added to custom-chip competition concerns.
  The bearish read-through: a better-funded MediaTek with more NVDA-provided connectivity and memory technology could compete more aggressively for hyperscaler ASIC wins, taking some shine off MRVL’s own NVDA partnership.

Otherwise fairly slow day here…

INTERNET

• OpenAI announced its new ad model had reached $1B in ARR after 200 days of launch. Hypers weak all day — didn’t hear any great reason other than the OpenAI news: AMZN -2.5%; GOOGL -2%; META -1%

• AMZN -2.5% saw from WSJ: FTC Files Lawsuit Alleging Amazon Deceived Advertisers. WSJ reporting the FTC and more than 20 states sued AMZN, alleging it secretly manipulated ad auctions to raise minimum prices and overcharged advertisers by more than $20B over seven years. Amazon denies wrongdoing, saying its auction practices are standard, transparent, and improved ad relevance/value.

Read more