Zan Digital › Companies
Nvidia
Every piece of Zan Digital analysis that covers Nvidia — 61 pieces, each figure traced to a named source.
- 001AI Capex 2026: The $725B Bet With No Payback MathFour hyperscalers plan roughly $725B of capex in 2026, up 77% on 2025. Here is where the money goes, what it earns today, and the three numbers to watch.$725B
- 002AI Circular Deals: 10 Arrows, Only 3 With a Cash FigureWe mapped 10 money flows between Nvidia, OpenAI, Microsoft and AMD. Only 3 carry a cash figure, led by Microsoft's $10.0B of OpenAI revenue in one quarter.$10.0B
- 003AI Circular Deals: Who Actually Pays Whom, and How MuchNvidia's up to $100B commitment to OpenAI was finalised at $30B. A map of the AI circular deals: what is signed, what is only intent, and why it matters.$100B → $30B
- 004AI Revenue Gap: Sequoia's $600B Question Rerun at $1.4TRerun Sequoia's $600B AI revenue gap math on Nvidia's $89.0B data centre quarter and the bill is $1.42T. OpenAI and Anthropic run rates cover about 8% of it.$89.0B
- 005AMD OpenAI Warrant: 160M Shares, $600 and 6 GigawattsAMD gave OpenAI a warrant for 160M shares at $0.01, vesting on 6 GW of GPU buys and a $600 price target. With Meta's twin, 320M shares, or 16.4% diluted.160M
- 006Anthropic Valuation vs OpenAI: Higher Mark, Lower MultipleAnthropic's $965B mark beat OpenAI's $852B, yet it priced at 20.5x run rate against OpenAI's 35.5x. Revenue, losses and compute commitments, each one cited.$965B
- 007Broadcom AI Revenue: $16.7B, Up 221%, vs Nvidia's $89BBroadcom AI revenue hit $16.7B, 18.8% of Nvidia's data center line, and its top 5 end customers reached 55% of sales. Here is what the filings disclose.$16.7B
- 008Circular Financing: How Auditors Test AI Vendor RevenueAMD will book its 160 million share OpenAI warrant as a cut to revenue. How ASC 606 and related-party rules treat an AI vendor's equity stake in its customer.160M
- 009Cognition Valuation: $48B Is 53x Disclosed Run RateCognition was reported in talks at $40B and closed at $48B. Against its disclosed $900M run rate that is 53x, about double what SpaceX agreed to pay for Cursor.$48B
- 010CoreWeave Customer Concentration: 62% From Two BuyersCoreWeave's top two customers paid 62% of Q2 2026 revenue. We tabulate Microsoft, OpenAI and Meta across filings and set them against the $35.6B debt stack.62%
- 011CoreWeave Customer Concentration: Top Client 71% to 36%CoreWeave's top customer fell from 71% to 36% of quarterly revenue in a year. Filing by filing, with Microsoft, OpenAI and Meta set against $35.6B of debt.36%
- 012CoreWeave Debt: $35.6B and Where the Stress SitsCoreWeave carries $35.6B of debt against a $104B backlog. The spread ladder from SOFR plus 225 to plus 550 shows exactly where lenders price the risk.$35.6B
- 013Crusoe Valuation: What 6 GW of Contracted Capacity MeansCrusoe's Series F priced it at $30.9B, about 3x its 2025 mark. Only 1 GW of its 6 GW is running. How to tell booked, contracted and planned capacity apart.$30.9B
- 014Datadog AI-Native Customers: 12% of Revenue, Then SilenceDatadog's AI-native cohort hit 12% of revenue in Q3 2025, then the share went undisclosed. Filing math puts it near 16% by Q2 2026, with one customer the risk.12%
- 015Dot-Com vs AI Bubble: Six Metrics Where It Breaks DownEveryone compares 2026 to 2000. On six metrics, from P/E ratios to profitability, the comparison mostly fails. Here is the one place it does hold.60x vs 23x
- 016DRAM Shortage 2026: Who Pays the AI Memory PremiumDRAM contract prices rose 93 to 98% in one quarter and Micron's gross margin hit 87%. We trace who absorbs the AI memory premium: Nvidia, clouds or PC buyers.87.0%
- 017Firmus IPO: Why Public Investors Refused a 3x MarkupFirmus pulled its ASX float at a US$30.6B valuation, nearly 3x its August round. What public buyers refused: US$30B of debt, leased sites and few customers.US$30.6B
- 018GPU Depreciation: The $176B Accounting Question in AIHyperscalers depreciate AI hardware over 5 to 6 years. One investor puts the resulting overstatement at $176B through 2028. Both sides of the argument, checked.$176B
- 019H100 Rental Price Index: $3.00 and Three Sources ComparedSilicon Data put the H100 rental price at $3.00 on 8 November 2026, up 50% in 11 months. Ornn read $3.37 that day. Both tested against 6-year useful lives.$3.00
- 020IMF AI Deals: $950B Raised Against $1.4T PromisedThe IMF counted 14 AI deals: $300B of equity and $650B of debt for $1.4T of promised revenue. We rebuilt the list and found $440B of it in company filings.$1.47
- 021Is the AI Bubble Real? 7 Tests, 2 Already TrippedIs the AI bubble real? We scored 7 falsifiable tests on Q3 2026 filings. 2 tripped: Alphabet capex hit 121% of operating cash flow. Demand tests still pass.121%
- 022Microsoft OpenAI Equity Method Loss: The $4.1B 10-Q LineMicrosoft booked a $4.1B OpenAI equity-method loss in one quarter and a $10.0B gain the next. Here is the 10-Q line, and how to read it on October 28.$4.1B
- 023Mistral Series D: Chipmakers and the AI Funding TableSamsung led Mistral's €3B Series D at over €21B. We tabulate 5 AI rounds from 2026 and find strategic money is a quarter or more, but rarely disclosed.€3B
- 024Nvidia Customer Concentration: 4 Buyers, 61% of SalesFour direct buyers took 61% of Nvidia's revenue in Q3 FY26. Two quarters on, one passed 10% while five held 70% of receivables. Four filings, in dollars.61%
- 025Nvidia Earnings: Who Is Carrying the $108B Q3 GuideNvidia guided $108B for Q3. Hyperscale made $48.7B of Q2 data center revenue and ACIE $40.3B, but ACIE grew 25% and now drives the guide. Who those buyers are.$108B
- 026Nvidia OpenAI Investment: $100B Promised, $30B WiredNvidia pledged up to $100B for OpenAI in September 2025. A year of SEC filings shows $30B of equity wired and a $105B lease guarantee. Here is the gap.$30B
- 027Nvidia Q3 FY27 Data Center Bar: $100.8B Implied by GuideNvidia reports Q3 FY27 on 17 November with no Data Center guide. Its $108B outlook implies $100.8B, a bar cleared in 7 of 8 quarters on the same method.$100.8B
- 028Nvidia Receivables and Inventory: The $279B Supply BillNvidia's receivables hit 60 days and supply commitments reached $279B, 2.9x quarterly revenue. Eight quarters of DSO, inventory days and commitments, computed.$279B
- 029OpenAI Compute Commitments: $720B Against $50B RevenueSix disclosed OpenAI compute contracts total about $720B. Spread over their terms they cost $117B to $136B a year, against a $50B revenue run rate in September.$720B
- 030OpenAI Valuation: $1.4T Ask, $1.25T Price, $11B RaisedOpenAI's round slid from a $1.4T ask to $1.25T pre-money, with $11B wired by 6 November. Its run-rate multiple fell from 35x in March to about 25x.$1.25T
- 031OpenAI Valuation: $1.4T Is 35x Run Rate, 107x RevenueOpenAI's reported $1.4T target is 35x its $40B run rate and 107x its $13.07B 2025 revenue. The multiple against Anthropic and the top 5 public software names.$1.4T
- 032Semiconductor Stocks 2026: Why Nvidia Trailed an 86% RallyThe chip index is up 86% in 2026 and Nvidia only 28%, despite $96.2B of quarterly revenue. Memory earnings carried the rally. Here is where the money went.86%
- 033Server Useful Life: $33B to $41B of Big Tech ProfitMicrosoft, Alphabet, Amazon and Meta depreciate servers over 5 to 6 years. At the 4-year life all four once used, operating income falls $33B to $41B.$2.92B
- 034Sovereign AI Funds: $30B and the Buyer of Last ResortSovereign AI spending pushed $30 billion through Nvidia alone in fiscal 2026. Here is what state programmes actually buy, and what it changes for vendors.$30B+
- 035The AI Revenue Leaderboard: Who Has Real RevenueCommitments, run rates and bookings get reported as if interchangeable. A ranking by revenue quality rather than headline number, with what each figure means.~$47B
- 036The IMF Warned on AI Valuations. The Optimists AnsweredThe IMF flagged correction risk in AI-exporting economies. Supporters argue compute scarcity makes the spending rational. Both cases, and what would settle it.Jul 2026
- 037TSMC Q3 2026 Earnings: What the Foundry Says About AITSMC's Q3 2026 revenue hit $46.7B, up 41% in dollars and 51% in NT dollars, with HPC at 68% of sales. What the filing says about AI chip demand in 2027.$46.72B
- 038Vendor Financing in AI: The Lucent and Nortel LessonSuppliers lending customers money to buy their products inflated telecom revenue before 2001. Similar structures now appear across AI. The mechanics compared.2001
- 039Claudeforce at Dreamforce 2026: What Salesforce ConcedesSalesforce shipped Claudeforce with 37 sales skills, an open beta and no price. What shipped, what is pilot, and what naming Claude concedes on agent value.37
- 040ServiceNow Stock Drop: What the Market Priced InsteadServiceNow passed $1B in AI ACV and raised guidance, yet fell 17% after Q1. We rebuild the bear case from its own filings: margins, M&A and on-prem timing.$1B+
- 041GPU Cost Allocation: AWS, Azure and Google Tools ComparedGPU cost allocation from vendor docs: Bedrock bills by IAM role per day, Vertex AI caps each label at 1,000 values, Azure splits by deployment. 7 of 21 native.1,000
- 042State of FinOps 2026: 98% Manage AI Spend, Tools LagState of FinOps 2026: 98% of teams manage AI spend, up from 31% in 2024. We read the sample and map which cost tools see tokens, GPUs and agent runs.98%
- 043Factory AI Review: Droids, Pricing and the Router ClaimFactory AI hit a $5B valuation in September 2026. We test Droids, plan pricing and the 63% router saving against Factory's own docs and the fine print.$5B
- 044Harvey AI Review: Products, Customers and Pricing SignalsHarvey raised $550M at a $15.5B valuation. We review its products, named customers and every disclosed pricing signal, and mark each gap as a gap.$15.5B
- 045MI355X vs B200 Rental: $7.70 vs $9.65 Per GPU-HourCrusoe lists the MI355X at $7.70 and the B200 at $9.65 per GPU-hour. On MLPerf v6.0 Server throughput, that is $0.171 vs $0.207 per 1M tokens at full load.$7.70 vs $9.65
- 046Model Routing Savings: 31.7% Measured, Not 85%The largest independent routing benchmark measured 31.7% cost savings, not 85%. Here is the quality gap across five enterprise tasks, and where it hides.31.7%
- 047Small Language Models Priced Per Task: Phi, Gemma, MistralA frontier model costs 30x to 37x more than Phi-4 per 1,000 tasks. The test, exact prompts and published rates for Phi-4, Gemma 3 and Mistral Small.30x
- 048AI Agent Adoption Surveys: Why 70% and 97% Don't Add UpCreatio says 70% are past AI experimentation. WRITER says 97% of executives deployed agents. Different samples, sponsors and definitions, side by side.70%
- 049AI Consulting Revenue: Who Discloses It and Who StoppedTCS puts AI at over 10% of revenue, Infosys at 8.2%, HCLTech at 5.9%. Accenture and IBM stopped publishing theirs. A firm-by-firm tally of what each counts.$3.1B
- 050AI Infrastructure ROI: What Gartner's 28% Actually CountsGartner's 28% counts AI use cases inside IT operations, not GPU buildouts. What the 782-leader survey measured, and what 5% GPU utilization adds to the story.28%
- 051AI ROI Survey 2026: NVIDIA 88% vs McKinsey 37%, ExplainedNVIDIA says 88% saw AI lift revenue. McKinsey says 37% saw any EBIT impact. Sample, wording and threshold compared, plus the one difference behind the gap.88%
- 052Dell AI Server Backlog: $95B and the Customer Mix Behind ItDell's AI server backlog hit $95B on $60.9B of quarterly orders. Six quarters of filings show who is buying: enterprise adds customers, neoclouds add dollars.$95B
- 053Enterprise AI Adoption 2026: 73% Use It, 10% Call It CorePublicis Sapient found 73% use AI regularly but 10% call it core. Deloitte found 66% see productivity gains, 20% revenue. Why the two measures diverge.73%
- 054Enterprise AI Spend: Menlo's $177B vs a16z and RampMenlo puts enterprise AI deployment spend at $176.8B, but a16z and Ramp publish no total. Compare each method, sample and scope, and what to cite for what.$176.8B
- 055GPU Utilization Break-Even: On-Prem vs Cloud CostOwning an 8-GPU B200 server beats on-demand cloud above roughly 37% utilization. The formula, every published input, and the places where the model breaks.$195.94
- 056Hybrid Search RAG: 5 Benchmarks vs Pure Vector SearchA reranker lifted BM25 by 11% across 18 BEIR datasets, and hybrid fusion beat both retrievers on 9 of 9 sets. What 5 benchmarks say about hybrid search RAG.+11%
- 057McKinsey State of AI 2026: Agent Scaling 40% vs 22%McKinsey's 2026 survey: 40% of $1B+ firms scale AI agents, up from 27%, while smaller firms held at 22%. What the sample can bear and what it never measures.40%
- 058MIT 95% AI Pilot Failure Study: What 52 Interviews ShowMIT NANDA's 95% figure rests on 52 interviews and 153 conference surveys. Of firms that piloted custom AI tools, 25% reached production. The method, read.52
- 059RAG Chunking Strategies: What 5 Benchmarks FoundChunk size moved RAG Recall@1 by up to 67 points in one study. We compare fixed, recursive, semantic and late chunking across 5 benchmarks, with defaults.67 points
- 060Sovereign AI Rules Compared: India, EU and Saudi ArabiaIndia binds data to Indian soil, the EU caps its gigafactory stake at 17% of capex and wants EU owners, Saudi Arabia sells compute abroad. What each requires.Rs 10,371.92 cr
- 061H-1B $100,000 Fee: 3 Lawsuits, 211,600 RegistrationsThe $100,000 H-1B fee met a 38% fall to 211,600 FY 2027 registrations and 3 lawsuits. A dated ledger of USCIS data, rulings and employer responses.211,600