The one thing
Nvidia’s July quarter landed after Wednesday’s close and beat on every line — revenue $96.2bn, up 106% year on year, data centre $89.0bn, up 117%, October-quarter guidance $108bn. None of that is the story. The story is that in her prepared remarks CFO Colette Kress gave a fiscal-2028 revenue growth number — approximately 70%, against sell-side consensus near 44–45% — the first time Nvidia has ever guided a full year ahead, and framed it in one sentence as “a supply-constrained outlook.” Huang’s gloss on the call: demand is much higher than 70%; supply is what lets the company commit to 70%.
Underneath that headline sits a second number that cuts the other way. Nvidia reset its forward gross margin to trough at 71–72% in the January quarter and settle at 72–73% through fiscal 2028, down from the 75.0% just reported, and named the cause as “extreme pricing conditions” in memory, worse than it had expected. That is Nvidia telling the market it is now a price-taker on HBM. For a coverage list that owns Micron and the memory chain on one side and the hyperscaler capex line on the other, that sentence is worth more than the beat.
Kress also sized the pool: cloud-industry backlog above $2tn, and top-five hyperscaler capex reaching nearly $800bn in 2026 and $1.3tn in 2027. Everything else overnight was second-order — Synopsys beat and fell, Salesforce and CrowdStrike both printed well, Meta settled with 51 attorneys general, and Apple set a date. One pattern worth carrying across the whole morning: the two largest prints of the night, Nvidia and Salesforce, both reported earnings materially inflated by mark-to-market gains on strategic investments — $7.8bn and $2.6bn respectively. Nasdaq futures were up about 1% pre-open, Nvidia itself indicated +5.9%.
Sector pulse — covered names with material news
| Semis & equipment | 3 / 16 |
|
| Cloud & AI infrastructure | 0 / 8 | |
| SaaS & enterprise software | 2 / 11 |
|
| Ads & adtech | 1 / 5 |
|
| Consumer platforms & commerce | 0 / 5 | |
| Hardware & devices | 1 / 3 |
|
Overnight & last session
Nvidia FQ2 2027: $96.2bn revenue, $108bn guide, and a first-ever year-ahead forecast of ~70% growth
Revenue of $96,221m was up 106% year on year and 18% sequentially; data centre $89.0bn, up 117% and also 18% sequentially, now 92% of the business. Gross margin 75.0% on both bases. October-quarter guidance is $108.0bn ±2% at 74.0% gross margin, and it explicitly assumes zero data-centre compute revenue from China. Nvidia returned $26.0bn to shareholders in the quarter with $99.0bn of authorisation remaining.
One trap in the EPS line. GAAP diluted EPS of $2.46 came in above non-GAAP of $2.22 — the reverse of the usual ordering. The gap is almost entirely $7,771m of pre-tax gains on equity securities that non-GAAP strips out. GAAP other income was $7,773m against non-GAAP other income of $300m. Nvidia’s marketable equity securities rose to $42.8bn from $12.9bn and non-marketable to $51.2bn from $22.3bn. This is not a one-off: the April quarter carried $15.9bn of the same.
The read: Two things to separate. First, a company that guides a year ahead for the first time is telling you its order book has become long enough to be an accounting object rather than a forecast — and that is exactly the claim that deserves the most scrutiny. Second, and more useful for a model: a growing share of Nvidia’s reported GAAP earnings is now mark-to-market gains on stakes in its own customers and ecosystem, nearly $24bn across two quarters. That is a circularity that shows up in the income statement rather than the revenue line, and almost nobody is adjusting for it. It was not the only such print last night either — Salesforce booked $2.6bn of strategic-investment gains that account for most of its own EPS beat. The clean version of the question, one the NVIDIA workbook can already answer, is what Nvidia’s GAAP earnings look like with equity gains stripped and how fast that line is growing relative to the operating business.
Memory is now visibly repricing the AI build-out — Nvidia concedes margin, Amazon concedes capex
Nvidia attributed its gross-margin step-down to “extreme pricing conditions” in memory, running worse than the company had assumed. Separately, and independently, Amazon’s move to roughly $220bn of 2026 cash capex from about $200bn was attributed to memory prices rather than a larger physical footprint — a $20bn increase that buys no additional capacity. Micron also announced leadership changes on 26 August, appointing Manish Bhatia President and COO and Scott DeBoer President and Chief Technology and Products Officer, both effective immediately.
The read: This is the cleanest evidence yet that the AI margin pool is migrating from logic to memory. Nvidia has been the price-setter in this chain for three years; conceding 200–300bp of gross margin to an input cost is the first time it has behaved like a price-taker, and Amazon spending $20bn more for the same footprint is the second data point in the same direction on the same day. The modelling consequence is sharp: if a rising share of each incremental capex dollar buys memory rather than compute, then delivered FLOPs per dollar of capex is falling, and the “capex equals capability” assumption underwriting most sell-side AI models breaks. The Micron and SK hynix initiations already built now have their corroboration from the customer side — which is the strongest form of evidence for a pricing-power claim.
Synopsys beats above the high end of guidance and raises the year — and the stock falls
Fiscal third-quarter revenue of $2.477bn against $1.740bn a year ago, with non-GAAP EPS of $3.91; both came in above the high end of company guidance. Full-year fiscal 2026 was raised to roughly $9.715bn of revenue and $15.07 of EPS at the midpoint. Shares fell despite the raise.
The read: A beat-and-raise that sells off is usually the market repricing the multiple rather than the numbers, and EDA is where that first becomes visible in the semis chain — Synopsys and Cadence get paid on design starts, which lead silicon revenue by a year or more. If the EDA names stop being rewarded for beats while Nvidia is guiding to 70% growth, the market is telling you it has already discounted the design pipeline and is now only paying for delivered volume. Worth pairing with Cadence into its next print: two EDA names, one identical demand signal, and a divergent multiple response is a publishable observation about where the cycle actually sits.
Vera Rubin ramp partners named — CoreWeave listed alongside the three hyperscale clouds
Nvidia’s release put Vera Rubin in full production and named CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud and Nebius as ramp partners. Background that matters: CoreWeave closed a $2.6bn delayed-draw term loan on 10 August, oversubscribed, priced at SOFR + 5.50%, with an approximately five-year maturity against customer contracts averaging roughly three years — the company’s own comparison, in its own release. That took it past $30bn of debt and equity raised year to date, against roughly $35bn of debt at the end of Q2 ($31.4bn recourse plus $3.7bn non-recourse; the $35bn is a sum, not a disclosed line).
The read: Being named beside Azure and Google Cloud is a commercial win and a funding-cost argument. The structural point is unchanged and under-covered: CoreWeave is financing assets with a maturity longer than the contracts that service them, so every renewal is a refinancing event in disguise. FactSet’s July work shows the same shape one layer up — incremental debt has gone from 9% of hyperscaler capex in FY24 to 32% on a trailing-twelve-month basis by mid-2026, aggregate debt near $700bn, and FY26 free cash flow expected near zero or negative everywhere except Alphabet and Microsoft. In fairness to that note, it also concludes gross leverage across the group remains low — the debt is growing fast from a small base, which is a different claim from distress, and the brief should not conflate them. The honest framing is that the levered periphery and the investment-grade centre are converging on the same financing model, not that either is in trouble today.
Salesforce raises FY2027 guidance — but the raise is mostly acquired and the EPS beat is mostly investment gains
Second-quarter revenue of $11,345m, up 10.8% (11% as the company states it, and in constant currency). Full-year FY2027 guidance was lifted to $46.1–46.4bn from $45.9–46.2bn — a $200m raise, $300m in constant currency, of which only about $100m is organic; the rest is the pending Contentful and Fin acquisitions. Reported revenue also includes a $456m contribution from Informatica, so organic growth is materially below the headline. Current remaining performance obligations reached $33.5bn, up 14% both reported and in constant currency, a point ahead of guidance and the fastest net-new bookings in four years. Income from operations was $2,331m — effectively flat against $2,332m a year ago.
The same trap as Nvidia, on a bigger scale. GAAP EPS of $4.29 and non-GAAP of $5.90 both include $2,613m of net gains on strategic investments — worth $2.43 of the GAAP figure and $2.53 of the non-GAAP one. Against consensus near $3.27, the operating beat largely disappears once those marks are removed. Note also that Agentforce ARR “exceeding $1.5bn, up over 240%” is not like-for-like: the definition was widened this quarter to fold in other AI offerings, Slackbot and Headless 360. Shares rose 12–13% after hours.
The read: There is a real story here and it is not the one the headlines are telling. Three separate things flatter this print — Informatica in revenue, Contentful and Fin in guidance, and $2.6bn of investment marks in EPS — while operating income is flat year on year. That is the note: Salesforce’s growth is being purchased, in three different currencies, and the organic engine is running at roughly zero operating leverage. The genuinely good number is cRPO at +14%, which is a clean bookings signal and should anchor the bull case. Then note that Salesforce and Nvidia, the two largest prints of the night, both reported earnings materially inflated by marks on strategic investments — that is a cross-coverage observation nobody is making, and with the Salesforce segment-label work already in the workbook it is a study CoreValue can write this week.
CrowdStrike posts record net new ARR of $332.8m, up 51%, and raises the ARR guide
Revenue $1,470.9m, up 26%, with subscription revenue $1.40bn, up 27%. Free cash flow $377.4m against $283.6m, a Q2 record, on operating cash flow of $530.3m. Net new ARR of $332.8m was up 51% year on year and a record from new logos, taking ending ARR to $5.84bn, up 25%. FY2027 revenue guidance rose to $5,991–6,011m and ending-ARR guidance to $6,603–6,612m, framed as a 630bp lift to about 34% net new ARR growth at the midpoint. The company also reported GAAP net income of $5.3m — its first GAAP profit since the July 2024 incident — and agreed to acquire the technology assets of XM Cyber. Shares rose 10.9% after hours, from $189.18 to $209.77.
Two things to handle carefully. The net new ARR “record” beats the prior high — Q4 FY2026’s $330.7m — by about $2m, so it is a record on the tape rather than a step change; the 51% year-on-year growth is the more meaningful number. And FY27 non-GAAP EPS guidance of $1.25–1.26 on 1,044m shares looks like a collapse against the prior $4.88–4.96 on 259m shares — it is not. A 4-for-1 stock split took effect on 2 July 2026; split-adjusted, prior guidance was about $1.22–1.24, so this is a raise.
The read: The CrowdStrike study spine is built and the chapters are unwritten — this print is the argument for writing them now, and the framing has just clarified. The thesis-defining datapoint is not the ARR record, which is marginal, but the combination of net new ARR up 51% with a record contribution from new logos and the first GAAP profit since the 2024 outage. Together those say the customer-acquisition damage is behind the business, which is a far stronger claim than a revenue beat and is exactly what the study’s central question was. It also dates the cybersecurity sector initiation split out of the same work: if CrowdStrike is winning new logos at this rate, that piece’s competitive-position section needs rewriting before publication, not after. Flag the split for the data spine — a 4-for-1 in July will silently break any per-share series that spans it, which is the same failure mode already logged in the ServiceNow and Broadcom workbooks.
Meta settles with 51 attorneys general for $16.7bn cash over ten years
Announced mid-trial in the Northern District of California, with a bipartisan coalition of 51 attorneys general led by California’s Rob Bonta: $16.7bn in cash over ten years, subject to court approval, with California receiving $1.5–2.1bn. The allegations were that Instagram and Facebook features were designed to drive compulsive use by minors and that internal evidence of harm was concealed. The injunctive terms are the substantive part — a default two-hour daily limit for under-18s, a midnight-to-6am block, school-hours notification blocking, bans on like counts and cosmetic-surgery filters for minors, and independent auditor oversight. Note the figure varies by source: the California AG release headlines “$17 billion” and some outlets ran “up to $18 billion” including the ascribed value of the injunctive terms. $16.7bn is the cash number.
The read: $1.67bn a year against a company guiding to $130–145bn of capex is a rounding error, and the market will treat it as overhang removal. The forecast risk is the injunctive terms, not the cash: a default two-hour cap and an overnight block on under-18s is a direct cut to engagement minutes in the cohort with the longest monetisable life, and there is no line in any published Meta model for it. That is the angle — everyone will write the settlement number; almost nobody will attempt to size the impression loss. Meta is also the one hyperscaler funding its build-out purely from advertising cash flow with no customer contracts behind it, which makes any structural hit to engagement matter far more to its capex sustainability than to Microsoft’s or Oracle’s.
Apple confirms a 9 September event — “Surprise and shine.”
Invitations went out on 26 August for an event on 9 September at 10:00 PT (14:00 BRT) at the Steve Jobs Theater, two weeks ahead, consistent with recent practice. Reporting expects the iPhone 18 Pro line and a foldable, with the affordable models held to spring 2027 and new silicon on a 2nm process.
The read: The 2nm node is the coverage-relevant detail, not the handsets. Apple is the anchor customer that funds the leading-edge ramp before AI accelerators migrate onto it, and TSMC has lifted 2026 capex to $60–64bn substantially to pay for that transition. The unwritten section in the Apple initiation is not the product cycle — it is Apple’s role as the involuntary underwriter of the AI build-out’s process node, absorbing first-year 2nm cost that Nvidia and Broadcom inherit at maturity. That framing also gives the existing Apple work something to say about AI without pretending Apple has an AI revenue line.
Full-year capex guidance — hyperscalers and largest semis
Where each name stands
| Name | Period | Reported | Key figures from that print |
|---|---|---|---|
| Semis & equipment | |||
| NVDA | FQ2 FY27 | 26 Aug 26 | Revenue $96.2bn +106%; Data Center $89.0bn +117%; GM 75.0%; GAAP dil. EPS $2.46 vs non-GAAP $2.22 (incl. $7,771m equity-securities gains); FQ3 guide $108.0bn ±2% at 74.0% GM, ex-China; FY28 growth guided ~70%, FY28 GM 72–73% |
| AVGO | FQ2 FY26 | 3 Jun 26 | Revenue $22.19bn +48%; AI semis $10.8bn +143%; Semi Solutions $15.0bn / Infra Software $7.2bn; FQ3 guide $29.4bn with AI semis ~$16.0bn; FY26 AI semis ~$56bn; FY27 AI semis “>$100bn” |
| AMD | Q2 CY26 | 4 Aug 26 | Revenue $11.54bn +50%; Data Center $6.72bn +107%; Client $3.06bn +23%; non-GAAP GM 56% |
| QCOM | FQ3 FY26 | 29 Jul 26 | Revenue $9.95bn −4%; QCT $8.50bn −5% (handsets $5.09bn −20%, auto $1.59bn +61%); non-GAAP EPS $2.21 |
| MU | FQ3 FY26 | 24 Jun 26 | Revenue $41.46bn +346%; DRAM $31.3bn (76%); non-GAAP GM 84.9% (record); data centre $11.5bn, up more than sevenfold; FQ4 guide $50.0bn ±$1.0bn. 26 Aug: Bhatia to President & COO, DeBoer to President & CTPO, both immediate |
| INTC | Q2 CY26 | 23 Jul 26 | Revenue $16.1bn +25%; Foundry $5.8bn +31% on a $(2.09)bn operating loss; GAAP EPS $(2.16), non-GAAP $0.42. Risk-factor disclosure on the US government’s 9.9% stake |
| TXN | Q2 CY26 | 22 Jul 26 | Revenue $5.46bn +23%; Analog $4.37bn +26%; operating margin 42.3% vs 35.1%; dil. EPS $2.14 +52% |
| MRVL | FQ1 FY27 | 27 May 26 | Revenue $2.42bn +28%; data centre $1.83bn +27% (76% of total); non-GAAP EPS $0.80; no standalone AI line. Reports FQ2 tonight |
| AMAT | FQ3 FY26 | 13 Aug 26 | Revenue $9.12bn +25%; Semiconductor Systems $7.04bn +27% (record); non-GAAP EPS $3.50 +41%; FQ4 guide ~$10.25bn |
| KLAC | FQ4 FY26 | 28 Jul 26 | Revenue $3.66bn +15.2%; non-GAAP dil. EPS $1.05 post a 10-for-1 split effected 11 Jun 26; FQ1 guide $4.0bn ±$200m |
| LRCX | FQ4 FY26 | 29 Jul 26 | Revenue $6.72bn +30%; non-GAAP dil. EPS $1.82 (record); China 26% of revenue, down from 34%; Sep-quarter guide $8.10bn ±$400m |
| ASML | Q2 CY26 · EUR | 15 Jul 26 | Revenue €9,326.5m +21.3%; GM 54.0%; net income €2,917.6m +27%; 86 systems; FY26 raised to €43–45bn. Net bookings not disclosed |
| ARM | FQ1 FYE27 | 29 Jul 26 | Revenue $1.29bn +22% (record); royalty $715m +22%, licensing $574m +23%; non-GAAP operating margin 41.2% vs GAAP 7.1% |
| TSM | Q2 CY26 · TWD | 16 Jul 26 | Revenue NT$1,270.4bn +36.0% (US$40.2bn); GM 67.7%; operating margin 60.3%; ≤7nm = 77% of wafer revenue; HPC 66% of net revenue. 2026 capex raised to $60–64bn from $52–56bn; 2026 growth guided slightly above 40% in USD terms |
| TSM — monthly | Jul 2026 | 10 Aug 26 | NT$467.58bn, +44.7% y/y, +5.6% m/m; Jan–Jul cumulative NT$2,872.06bn +37.0%. August print due ~10 Sep |
| CDNS | Q2 CY26 | 27 Jul 26 | Revenue $1.58bn +24%; backlog $8.1bn (record); NTM RPO $4.2bn; non-GAAP dil. EPS $2.11 vs $1.65 |
| SNPS | FQ3 FY26 | 26 Aug 26 | Revenue $2.477bn vs $1.740bn; non-GAAP EPS $3.91; both above the high end of guidance; FY26 raised to ~$9.715bn revenue and $15.07 EPS at midpoint. Stock fell on the print |
| Cloud & AI infrastructure | |||
| AMZN | Q2 CY26 | 30 Jul 26 | Revenue $200.6bn +20%; AWS $42.2bn +37% (fastest in 18 quarters); AWS operating income $16.6bn vs $10.2bn; total operating income $27.5bn +43%. CY26 cash capex ~$220bn, raised from ~$200bn on memory prices; $25bn bond placement ~2.5x oversubscribed |
| MSFT | FQ4 FY26 | 29 Jul 26 | Revenue $90.0bn +18%; Microsoft Cloud $59.3bn +27%; Azure +43%; commercial RPO $678bn +84%; capex $41bn incl. finance leases. CY26 capex ~$175bn, down from $190bn on a 15→25yr useful-life change — not a spending cut; FQ1 FY27 >$50bn |
| GOOGL | Q2 CY26 | 22 Jul 26 | Revenue $119.8bn +24%; Google Cloud $24.77bn +82%; cloud backlog $514bn; Q2 capex $44.9bn; CY26 capex raised to $195–205bn from $180–190bn; $84.75bn equity raise Jun 2026. ~75% of revenue is still ads |
| ORCL | FQ4 FY26 | 10 Jun 26 | Revenue $19.2bn +21%; RPO $638bn +363% (incl. ~$75bn of prepayments and customer-supplied hardware); OCI $5.8bn +93%; FY26 capex $55.7bn; FY27 revenue guide $90bn; FY27 net cash capex ~$70bn ex $20–25bn prepayments; ~$40bn debt+equity planned FY27; $43bn debt raised FY26 |
| NET | Q2 CY26 | 6 Aug 26 | Revenue $696.1m +36%; DBNRR 120% vs 114%; 4,698 customers >$100k, +27%, = 73% of revenue; GAAP GM 71.8% |
| VRT | Q2 CY26 | 29 Jul 26 | Revenue $3,274.3m +24%; adj. operating margin 22.6% (+410bp); adj. dil. EPS $1.52 +60%. Orders and backlog no longer disclosed |
| GEV | Q2 CY26 | 22 Jul 26 | Revenue $11.1bn +22% reported / +12% organic; orders $24.2bn +88% organic; backlog $176bn vs $129bn; gas equipment backlog plus slot reservations 116 GW, up from 100 GW |
| CRWV | Q2 CY26 | 11 Aug 26 | Revenue $2,575m +112%; backlog $104.2bn +246%; Q2 capex $6,422m cash PP&E; net loss $(626)m against net interest expense of $(640)m; total debt $35.07bn ($31.4bn recourse + $3.7bn non-recourse). FY26 revenue guide $12.4–13.2bn, capex raised to $35–39bn. $2.6bn DDTL closed 10 Aug at SOFR+5.50%, ~5yr vs ~3yr contracts |
| SaaS & enterprise software | |||
| CRM | FQ2 FY27 | 26 Aug 26 | Revenue $11,345m +10.8% (incl. $456m Informatica); cRPO $33.5bn +14%; operating income $2,331m, flat y/y; GAAP EPS $4.29 / non-GAAP $5.90, of which $2.43 / $2.53 is strategic-investment gains; FY27 guide raised to $46.1–46.4bn; Agentforce ARR >$1.5bn on a widened definition |
| ADBE | FQ2 FY26 | 11 Jun 26 | Revenue $6.62bn +13%; total ending ARR $27.10bn +12.5%; RPO $22.27bn +13%; AI-first ARR >$500m. Digital Media / Digital Experience segmentation dropped |
| NOW | Q2 CY26 | 22 Jul 26 | Revenue $3,987m +24%; subscription $3,877m +24.5%; cRPO $13.20bn +21%; 658 customers above $5m ACV |
| SNOW | FQ1 FY27 | 27 May 26 | Revenue $1.39bn +33%; product revenue $1.334bn +34%; NRR 126%; RPO $9.21bn +38%. FQ2 due 2 Sep |
| MDB | FQ1 FY27 | 28 May 26 | Revenue $687.6m +25%; Atlas $512.5m +29% (~75% of subscription); 67,700+ customers; 2,895 above $100k ARR. FQ2 due 1 Sep |
| HUBS | Q2 CY26 | 5 Aug 26 | Revenue $911.7m +20%; subscription $894.0m +20%; 306,446 customers +14%; ASRPC $11,800 +4% |
| DDOG | Q2 CY26 | 6 Aug 26 | Revenue $1.12bn +36%; ~4,720 customers above $100k ARR, +23%; FCF $279m. GAAP operating income $5m (~0%) against 23% non-GAAP |
| WDAY | FQ1 FY27 | 21 May 26 | Revenue $2.54bn +13.5%; subscription $2.35bn +14.3%; cRPO $8.81bn +15.5%; non-GAAP operating margin 31.8%. Reports FQ2 tonight |
| SAP | Q2 CY26 · EUR | 23 Jul 26 | Revenue €9.878bn +9% (+11% cc); cloud revenue €6.281bn +22%; current cloud backlog €22.9bn +27% |
| PANW | FQ3 FY26 | 2 Jun 26 | Revenue $3.0bn +31%; NGS ARR $8.1bn +60%; RPO $18.4bn +36%. CyberArk and Chronosphere contributed $388m of the quarter — not organic. FQ4 due 1 Sep |
| CRWD | FQ2 FY27 | 26 Aug 26 | Revenue $1,470.9m +26%; net new ARR $332.8m +51% (record, vs $330.7m prior high); ending ARR $5.84bn +25%; FCF $377.4m; first GAAP profit since Jul-2024 ($5.3m); FY27 ARR guide $6,603–6,612m. 4-for-1 split 2 Jul 2026 breaks per-share series |
| Ads & adtech | |||
| META | Q2 CY26 | 29 Jul 26 | Revenue $60,801m +28%; advertising $59,363m; Family DAP 3.60bn +3%; impressions +14%, price per ad +12%; CY26 capex $130–145bn incl. finance leases, narrowed from $125–145bn. 26 Aug: $16.7bn/10yr settlement with 51 state AGs plus minor-usage injunctive terms |
| ZETA | Q2 CY26 | 4 Aug 26 | Revenue $443m +44% reported but +28% organic; adj. EBITDA $92m at 20.7% margin; 197 super-scaled customers +17% at $1.8m ARPU |
| TTD | Q2 CY26 | 6 Aug 26 | Revenue $715m +3%; adj. EBITDA $241m at 34% margin; retention >95%; Netflix joined the marketplace and Samsung home-screen inventory opened |
| ROKU | Q2 CY26 | 6 Aug 26 | Revenue $1.35bn +22%; platform $1.22bn +25% (advertising $673m, subscriptions $548m); 37.9bn streaming hours; net income $164.2m. Household count discontinued since Q1 2025 |
| NFLX | Q2 CY26 | 16 Jul 26 | Revenue $12.56bn +13%; operating income $4.19bn at 33.4% margin, down from 34.1%; FCF $1.53bn with FY26 guide ~$12.5bn. Membership counts not disclosed |
| Consumer platforms & commerce | |||
| ABNB | Q2 CY26 | 6 Aug 26 | Revenue $3.61bn +17%; Nights and Seats Booked 148.3m +10% (metric renamed); GBV $27.2bn +16%; ADR $184 +5%. 52-week high $187.30 on 21 Aug |
| BKNG | Q2 CY26 | 4 Aug 26 | Revenue $7.4bn +8%; room nights 325m +5%; gross bookings $51.0bn +9% |
| UBER | Q2 CY26 | 5 Aug 26 | Revenue $14.19bn +12% against gross bookings $58.02bn +24% — a wide as-reported divergence worth checking in the 10-Q; adj. EBITDA $2.82bn +33%; MAPCs 208m +16% |
| SHOP | Q2 CY26 | 5 Aug 26 | Revenue $3,583m +34%; GMV $115,567m +32%; FCF $654m at 18% margin, up from 16% |
| WMT | FQ2 FY27 | 20 Aug 26 | Revenue $187.9bn +5.9%; US comps +2.6% ex-fuel, transaction-led; global eCommerce +23%; global advertising +38% (Walmart Connect US +43% ex-VIZIO) |
| Hardware & devices | |||
| AAPL | FQ3 FY26 | 30 Jul 26 | Revenue $109.4bn +16%; iPhone $54.3bn +22%; Services $30.7bn +12% (June-quarter record); GM 50.1% including ~2pp of tariff-refund benefit; dil. EPS $2.02 +29%. Product event 9 Sep, 10:00 PT |
| DELL | FQ1 FY27 | 28 May 26 | Revenue $43.8bn +88%; ISG $29.0bn +181%; AI-optimised server revenue $16.1bn +757%; AI orders booked $24.4bn; AI backlog $51.3bn; FY27 AI-server guide ~$60bn. FQ2 due 1 Sep |
| CSCO | FQ4 FY26 | 12 Aug 26 | Revenue $17.3bn +18%; product orders +35% (+25% ex-hyperscalers); AI infrastructure orders $4bn in the quarter and $9.3bn for FY26, ~4.5x FY25; RPO $46.7bn +7% |
On the calendar
- Today, 27 Aug · 17:45 BRTMarvell (MRVL) FQ2 2027, after the US close (call 13:45 PT). The custom-silicon read on Nvidia’s supply-constraint claim.
- Today, 27 Aug · after closeWorkday (WDAY) FQ2 2027. The seat-based-SaaS counterpoint to Salesforce’s consumption story.
- Tue, 1 Sep · 17:30 BRTDell (DELL) FQ2 2027 (15:30 CDT). Watch AI server backlog against $51.3bn, and gross margin against memory cost — Dell has no pricing power to absorb it.
- Tue, 1 Sep · 17:30 BRTPalo Alto Networks (PANW) FQ4 and FY2026. Reads across to CrowdStrike’s net-new-ARR share claim.
- Tue, 1 Sep · 18:00 BRTMongoDB (MDB) FQ2 2027 (17:00 ET). Three covered names collide on this date.
- Wed, 2 Sep · 18:00 BRTBroadcom (AVGO) FQ3 2026 (14:00 PT). The $16.0bn AI-semis guide and any refresh of “>$100bn in FY2027”.
- Wed, 2 Sep · 18:00 BRTSnowflake (SNOW) FQ2 2027 (14:00 PT), quarter ended 31 July.
- Wed, 9 Sep · 14:00 BRTApple “Surprise and shine.” (10:00 PT, Steve Jobs Theater) — iPhone 18 Pro line and foldable, 2nm silicon.
- ~Thu, 10 Sep · 02:30 BRTTSMC August monthly revenue (13:30 Taipei). The highest-frequency read on AI demand anywhere in coverage.
- ~Thu, 10 SepAdobe (ADBE) FQ3 2026, after the close — third-party listing only, not confirmed by Adobe IR.
- Sep, date TBCOracle (ORCL) FQ1 2027 — Oracle has not issued its scheduling release; third-party sources split between 8 and 14 September. The single most important print for the financing thread.
- Wed, 30 Sep · 17:30 BRTMicron (MU) FQ4 2026 (14:30 MT). The direct test of the memory-pricing thesis Nvidia and Amazon both endorsed yesterday.
Story worth owning
Who owns the residual? On 10 August Nvidia signed MOUs with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to mobilise more than $500bn of third-party capital into AI compute, explicitly framing GPUs as an investable asset class alongside toll roads and commercial real estate. Separately from that release — in a post by Huang on X the same day, not in the announcement itself — came the detail that matters most: Nvidia may provide residual-value support of up to 25%, assessed deal by deal. That is not the blanket “$125bn backstop” now circulating; it is something more interesting, and more specific.
Because a residual-value guarantee is a statement about a price curve nobody has published. The sell-side depreciates AI hardware over five to six years because that is what the hyperscalers disclose — but the hyperscalers are the buyers, not a market, and a useful life is an accounting policy, not a resale price. Microsoft just demonstrated exactly how movable that policy is, extending datacenter life from 15 to 25 years and cutting $15bn off its own capex line without changing a single spending decision. Meanwhile the one arm’s-length datapoint available, the tenor gap in CoreWeave’s own facility, has lenders underwriting roughly two years of residual on hardware whose successor architecture Nvidia has already put into full production.
The wider pattern is that the marginal AI dollar keeps being moved off the buyer’s balance sheet, in a different instrument each time. CoreWeave guides 2026 capex of $35–39bn against revenue of $12.4–13.2bn — roughly 3x capex-to-revenue, outspending Micron and Intel on capital from a $13bn revenue base, with a $640m quarterly interest bill against a $49m operating loss. Oracle funds its build partly with about $75bn of customer prepayments and customer-supplied hardware. Marvell handed Google a warrant over roughly 59m shares vesting in 240 tranches tied to Google’s own purchases. Intel raised $20bn of common equity in mid-August. Alphabet raised $84.75bn of equity in June. Debt, vendor equity, prepayments, contract-backed loans, third-party platforms, plain dilution — six instruments, one move.
The piece writes itself from work already on the shelf: the AI Bubble Radar tripwires, the CoreWeave debt schedule, the Data Centre Atlas capacity data, and the depreciation-life disclosures in the Amazon, Microsoft, Alphabet and Meta workbooks. The angle is not “is it a bubble” — that question is exhausted, and Nvidia’s 70% guide has largely answered it for the next four quarters. It is narrower and genuinely unowned: the AI trade is being refinanced from equity onto credit, that credit is being written against a residual value nobody has estimated in public, and the vendor has now agreed to guarantee part of it.
The deliverable is a single taxonomy chart: rows are financing instruments, columns are covered names, and each cell carries the disclosed amount and — the column nobody else builds — who holds the residual: shareholder, bondholder, private-credit fund, customer, or vendor. Roughly $225bn of hyperscaler and related bond issuance has been counted in 2026 to date, about a third of capex debt-financed, but bonds are the visible part. The prepayments, warrants and contract-backed loans are not, and that is where the concentration sits. An independent estimate of the GPU residual curve, built bottom-up from useful-life disclosures, contract tenors and successor-architecture timing, would be the first of its kind. CoreValue has the master files for almost every name in that grid; nobody else does.
Sources
- NVIDIA Corporation — Form 8-K, Exhibit 99.1, Q2 FY2027 results — SEC EDGAR, 26 August 2026.
- NVIDIA Announces Financial Results for Second Quarter Fiscal 2027 — NVIDIA Newsroom, 26 August 2026.
- Nvidia’s 70% sales growth projection for 2028 wows analysts on earnings call — CNBC, 26 August 2026.
- NVIDIA Q2 FY2027 earnings call transcript — Investing.com, 26 August 2026. (Source of the “cloud industry backlog now greater than $2 trillion” and the ~$800bn / $1.3tn hyperscaler capex figures.)
- NVIDIA expects ~70% fiscal 2028 revenue growth as it resets gross margin outlook to 72–73% — Seeking Alpha, 26 August 2026.
- Nvidia pops 4.7% after hours as earnings lift futures — Benzinga, 26 August 2026.
- NVIDIA Partners With Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to Mobilize Over $500 Billion of Third-Party Capital — NVIDIA Newsroom, 10 August 2026.
- Nvidia lines up $500 billion in financing as Huang calls his chips an ‘investable asset’ — CNBC, 10 August 2026.
- CoreWeave — Form 8-K, Exhibit 99.1, $2.6bn DDTL 5.5 facility — SEC EDGAR, 10 August 2026.
- CoreWeave Q2 2026 earnings report — CNBC, 11 August 2026.
- Hyperscalers Tap External Financing as AI Capex Outruns Cash Flow — FactSet Insight, 23 July 2026.
- Salesforce — Form 8-K, Exhibit 99.1, FQ2 2027 results — SEC EDGAR, 26 August 2026.
- CrowdStrike Holdings — Form 8-K, Exhibit 99.1, FQ2 2027 results — SEC EDGAR, 26 August 2026.
- CrowdStrike Q2 FY2027 earnings — shares pop more than 11% in extended trading — CNBC, 26 August 2026.
- CrowdStrike Holdings — Form 8-K, Exhibit 99.1, FQ1 2027 results — SEC EDGAR, 3 June 2026. (Prior guidance, used for the split-adjusted comparison.)
- Synopsys Posts Financial Results for Third Quarter Fiscal Year 2026 — Synopsys, 26 August 2026.
- Micron Announces Leadership Appointments — GlobeNewswire, 26 August 2026.
- Attorney General Bonta Secures Settlement With Meta — California Office of the Attorney General, 26 August 2026.
- Microsoft FY2026 Q4 earnings materials — Microsoft Investor Relations, 29 July 2026.
- Amazon.com — Form 8-K, Exhibit 99.1, Q2 2026 results — SEC EDGAR, 30 July 2026.
- Alphabet Q2 2026 earnings call — Alphabet Investor Relations, 22 July 2026.
- Meta Platforms — Form 8-K, Exhibit 99.1, Q2 2026 results — SEC EDGAR, 29 July 2026.
- Oracle — Form 8-K, Exhibit 99.1, FQ4 and FY2026 results — SEC EDGAR, 10 June 2026.
- TSMC — Q2 2026 results and guidance — SEC EDGAR, 16 July 2026.
- Broadcom — Form 8-K, Exhibit 99.1, FQ2 2026 results — SEC EDGAR, 3 June 2026.
- Broadcom to Announce FQ3 2026 Results on 2 September 2026 — Broadcom Investor Relations.
- Snowflake to Announce FQ2 2027 Results on 2 September 2026 — Snowflake.
- Dell Technologies Delivers FQ1 2027 Results — Business Wire, 28 May 2026.
- Apple Events — “Surprise and shine.”, 9 September 2026 — Apple.
- CoreWeave Reports Second Quarter 2026 Results — CoreWeave Investor Relations, 11 August 2026. (Revenue, backlog, interest expense, debt, FY26 capex guidance.)
- Marvell’s stock pops on an AI chip deal that lets Google buy up to $12.2 billion in shares — CNBC, 19 August 2026.
- Intel Announces Upsize and Pricing of $20 Billion Common Stock Offering — Intel Investor Relations, 10 August 2026.
- TSMC July 2026 Revenue Report — TSMC Investor Relations, 10 August 2026.
- Microsoft jumps as it boosts capital spending plans — CNBC, 29 July 2026. (Useful-life change, ~$175bn vs $190bn.)
Figures are as reported by companies, their investor-relations releases, SEC filings and the news sources listed above, gathered on the morning of 27 August 2026 and cross-checked in a separate verification pass. They have not been independently re-derived from audited financial statements, and where sources disagree the disagreement has been stated in the text rather than resolved silently. Prepared for internal research use at CoreValue Research. This is not investment advice, not a recommendation to buy or sell any security, and no representation is made as to accuracy or completeness.
