The Index
519 dossiers
A research screener for every company we cover. Search a name or ticker, then sort and triage dossiers by coverage freshness, our conviction and trading relevance.
ai16 names· avg TR 2 | ||||||||
| The market spent two years pricing Alphabet as the AI loser; FY2025 (+15% op income, Cloud margin to 24%, Gemini at 750M users, the antitrust gun mostly de-cocked) proves it owns the full stack — but at $4.5T and a fresh $90B+ capex habit, the easy re-rating is now behind it and the bar is "monetize the buildout." | — | 1 | ||||||
| The neutral arms-dealer of the AI build-out — AWS + Trainium + a stake in both OpenAI and Anthropic is the best-positioned compute franchise on Earth, but free cash flow has been incinerated to ~$1B and a ~$200B 2026 capex bet on 5-year-depreciated silicon is the whole thesis. | — | 1 | ||||||
| The best business in AI on every metric that isn't valuation — enterprise-led, multi-cloud, ~$30B+ run-rate doubling on weeks — wrapped in a ~$965B mark that already prices a decade of flawless execution and rests on a circular ring of vendor-investors funding ~$180B of compute the company hasn't yet earned the margins to pay for. WATCHING, bull on the business / bear on the entry price. | — | 1 | ||||||
| The shelf RE-GROUNDS and partly CORRECTS the prior web-only read. Three filings-level facts reshape the thesis. (1) Concentration is NOT "OpenAI is the new whale" — as of the reported period the top two customers are MBZUAI (62% of FY25 revenue / 77.9% of AR) and G42 (24%), BOTH UAE-affiliated related parties = 86% of FY25 revenue; OpenAI is a forward 750MW contract (deploys 2026-2028) whose revenue has barely begun. (2) FY25 GAAP "net income $237.8M" is an accounting artifact of +$390.7M of non-operating Other income — dominated by a one-time $363.3M non-cash gain on extinguishment of the Series F forward-contract liability — while operations LOST $145.9M and non-GAAP net loss was $(75.7)M. (3) The margin guide-down is substantially a *reported-revenue* effect (customer-warrant amortization that reduces revenue + datacenter pass-through booked gross + rent-back start-up costs), per management a 10-15pt transitional drag, not purely "inference is structurally low-margin." Net: real wafer-scale architectural edge, genuinely circular AI-infra financing now on the balance sheet (~$1.0B OpenAI working-capital loan at 6% + OpenAI & AWS customer warrants), still UAE-concentrated revenue, ~46x FY26-guide sales vs CoreWeave at ~7x. WATCHING, bear-leaning on valuation; the structural short case is weaker than the prior dossier asserted and the *transitional-margin* read is more defensible. | ||||||||
C CoreWeavecatalyst in 44d | The story got bigger AND riskier since Jun 14. RPO leapt $60.7B→$99.4B and the cost of capital genuinely improved (first IG-rated HPC debt, A-, <6%); but net loss widened to $(740)M, capex guidance jumped to $31–35B, ~$8.5B of fresh 8.5–9.75% debt was raised post-quarter, and the prior "regulatory: clean" call is WRONG — multiple securities-fraud class actions are now live over concealed data-center delays. Still a leveraged bet on the AI-capex cycle wearing an infrastructure costume; fairly-to-richly priced at ~$96 (EV ~$86B). WATCH; the de-risking (IG debt, <30% non-IG backlog) is real but does not change direction — want a cheaper entry or a GAAP/FCF inflection before paying up. | — | ||||||
| The lab that broke the cost-of-intelligence narrative and now anchors China's sovereign AI stack — a frontier-class open-weight model trailing the closed leaders by 3-6 months at ~1/50th the unit cost, but state-captured, un-investable to US capital, and monetizing almost none of the value it creates. | — | 1 | ||||||
| The cheapest mega-cap AI franchise — DeepMind's full-stack (TPU → Gemini → 15 products) just proved monetization (Cloud +63%, $462B backlog, AI Overviews monetizing at parity), so the AGI optionality is a free call; the real bet is whether the Dec-2025 Search remedy survives appeal and capex discipline holds. | — | 1 | ||||||
G | The crown jewels and the brain already left for Nvidia for ~$17–20B; what remains is a cash-rich but architecturally-cornered inference cloud whose flagship Saudi demand evaporated — a melting-ice-cube neocloud, not the LPU disruptor the headlines sold. | — | 1 | |||||
| The best-funded, worst-priced AI franchise in megacap — a $200B+ ad engine bankrolling a $135B/yr superintelligence bet the market is refusing to underwrite; the stock pays you to wait while the ad-AI flywheel already prints, but the open-weight thesis is dead and the capex ROI clock is now running loud. | — | 1 | ||||||
| The cheapest mega-cap in the AI complex (~18–20x fwd P/E) compounding ad revenue +33% — but it has converted itself into a single-variable bet on whether $125–145B/yr of AI capex earns its return, and the market won't re-rate until the depreciation wave proves out. | — | 1 | ||||||
| The cleanest AI-infrastructure compounder on the board, now de-rated to ~20x forward on a real fear — Azure deceleration meeting a $190B/yr capex wall — that is more a timing question than a thesis-breaker; the moat (RPO $633B, M365 distribution, ~27% of OpenAI) is intact, the FCF is the variable to watch. | — | 1 | ||||||
| Europe's only credible frontier lab — a sovereignty-and-infrastructure bet, not a model-quality bet; the moat is regulatory/political, the risk is that "good enough, in Europe, on-prem" gets out-priced by Chinese open weights before the €20B mark is earned. | — | 1 | ||||||
| The revenue leader that turned into the cash-burn leader — a $1.4T compute book financed against a P&L losing ~$14B a year, where the only question that matters is whether the IPO window opens before the funding treadmill stalls. | — | 1 | ||||||
| The FY26 10-K turns the bull narrative into audited fact and the bear narrative into a footnote you can now read — $67.4B revenue, $638B RPO, but only 12% converts inside a year, −$23.7B FCF, $129.5B debt, a 13% workforce cut, and a $19B post-quarter purchase commitment; good news kept beating and the stock kept falling (~$166, −52% from peak), which is the whole thesis in one line. | — | 8 | ||||||
| A genuinely great business priced as a perfect one — 85% growth and 60% adjusted margins are real, but ~55x EV/Sales already discounts flawless execution for years, so the asymmetry is to the downside even if the thesis is right. | — | 1 | ||||||
X | A frontier lab fused into Musk's compute-and-distribution empire — Grok is a fast-following #4 model bankrolled by a $30B/yr capex furnace; the real instrument is now SpaceX equity at a $1.25T mark, not a standalone xAI bet. | — | 1 | |||||
| — |
| 8 |
| 8 |