How Much of Big Tech's Q1 2026 Performance Came From Anthropic and OpenAI?

When Microsoft, Alphabet, and Amazon reported Q1 2026, a surprising share of the headline numbers traced back to two private companies they don't own outright. Cloud-revenue commitments, equity-stake markups, and AI run-rate disclosures all point to Anthropic and OpenAI as load-bearing inputs to Big Tech's recent earnings. The question of how much of Big Tech's good performance comes from Anthropic and OpenAI does not have a single clean number, but it has several specific ones, and they are larger than most investors realize.

Two channels matter. The first is direct revenue: Anthropic and OpenAI buy enormous amounts of cloud compute, and those purchases show up as Azure, AWS, and Google Cloud revenue. The second is equity-stake markups: Microsoft owns part of OpenAI; Amazon and Alphabet each own part of Anthropic. When those private valuations move up after a funding round, the holding company books the gain through net income. In Q1 2026, that second channel produced one of the most striking single-line items of the quarter.

The Numbers, Briefly

Microsoft + OpenAI: The Biggest Single Cloud Customer Relationship in Tech

Microsoft's exposure to OpenAI runs through three lines: Azure consumption, the equity stake, and remaining performance obligations (RPO). All three are large. OpenAI committed to spend at least $250 billion on Azure services through 2032 as part of the October 2025 recapitalization. That commitment is the single largest cloud-purchase contract in the industry, and it is the anchor of Microsoft's $627 billion remaining performance obligations balance disclosed at Q3 FY2026.

Microsoft's $37 billion AI annual run rate, up 123% year over year, is not all OpenAI consumption; it includes Copilot revenue, third-party AI services, and Azure AI workloads from non-OpenAI customers. But OpenAI's traffic is the foundation that makes the rate possible. When OpenAI's user base or model usage grows, Azure compute consumption grows roughly in proportion. Microsoft's AI revenue line is therefore tightly coupled to OpenAI's product trajectory.

On the equity side, Microsoft's stake in OpenAI's PBC is valued at approximately $135 billion as of the October 2025 recapitalization, representing roughly 27% on an as-converted diluted basis. Microsoft books the OpenAI investment through equity-method accounting. In Q3 FY2026, the OpenAI investment had only a $14 million net negative impact on net income, a sharp improvement from the $583 million net negative impact in the prior-year quarter. The recently restructured partnership ends Azure exclusivity (OpenAI can now sell on AWS and Google Cloud) but maintains the $250 billion Azure commitment and the equity stake.

Amazon + Anthropic: When the Stake Becomes the Story

Amazon's Q1 2026 was the quarter the equity-stake angle moved from footnote to headline. The company reported $16.8 billion of pre-tax non-operating gains tied to its Anthropic stake markup. Per Fortune's analysis of the report, that single line represented more than half of Amazon's Q1 pre-tax income. Amazon's $8 billion-plus original investment in Anthropic is now worth more than $70 billion, with management announcing an additional commitment of up to $25 billion in April 2026.

On the revenue side, Anthropic committed to spend more than $100 billion on AWS technologies over the next ten years as part of an expanded infrastructure deal. That implies roughly $10 billion in annual AWS revenue from Anthropic alone, with up to 5 gigawatts of dedicated Trainium capacity reserved for Claude model training and inference. Amazon's 28% AWS growth in Q1, the fastest pace in over three years, was supported in part by this commitment.

The accounting is uncontroversial in itself. The wrinkle, as one tax-and-accounting consultant noted, is what drives the markup: private company valuations come from whatever a small group of investors agreed to pay in the most recent funding round, not from open market pricing. Each new Anthropic round therefore creates fresh markup gains for Amazon to recognize. The cycle creates a feedback loop: Amazon commits more capacity to Anthropic, Anthropic raises at a higher valuation, Amazon books a gain on its stake, and the cycle continues.

Alphabet + Anthropic: The Newest Investment, Already in the P&L

Alphabet announced a $40 billion investment in Anthropic on April 24, 2026, lifting Anthropic's post-money valuation to $350 billion and locking in 5 gigawatts of dedicated TPU capacity, up to one million Tensor Processing Units. The deal puts Alphabet in the same accounting position as Amazon: a sizeable equity stake in a private AI company whose valuation will be re-marked at every subsequent round.

Per the same Fortune analysis, a meaningful portion of Alphabet's Q1 2026 "AI profits" traced back to the Anthropic stake markup mechanism. Disentangling exact figures from the disclosed financials is harder than at Amazon (Alphabet's segment reporting does not isolate the markup contribution), but the pattern is the same: equity-method or fair-value accounting on a fast-appreciating private holding produces non-operating gains that flow into reported earnings.

On the revenue side, Anthropic's TPU commitment, plus its broader Google Cloud usage, supports Google Cloud's 63% Q1 2026 growth and the segment's expansion to 18% of total Alphabet revenue. As with AWS, the cloud-revenue line and the equity-stake line both reinforce the same theme: Anthropic is now a load-bearing customer and asset for two of the three major US hyperscalers.

Meta and Apple: Different Exposures

Meta does not have a major Anthropic or OpenAI commercial relationship. Meta's AI strategy runs through Llama, developed and trained internally on Meta-owned infrastructure. The 33% Q1 2026 revenue growth is largely organic to advertising and Reels monetization, with AI-driven ranking improvements as a contributor but no comparable equity-stake or cloud-customer line tied to Anthropic or OpenAI. That isolation is a double-edged characteristic: Meta misses the boost that Amazon and Alphabet got from stake markups, but it also avoids the dependency on a single non-owned counterparty.

Apple's exposure runs in the opposite direction. The most prominent Apple-AI partnership of the quarter was the announcement that Apple will use Google's Gemini to power upgraded Siri reasoning. That is a cost line for Apple, not a revenue line. Apple is paying Google for inference; the dollars flow toward Alphabet, not toward Apple's results. Apple's relationship with OpenAI through ChatGPT integration in iOS is similar in shape: it is a partnership that adds product capability, not earnings contribution. Apple's Q2 FY2026 numbers stand essentially independent of Anthropic and OpenAI.

How to Read "AI Profits" After This Quarter

Investors trying to disentangle organic AI revenue from equity-stake gains have two practical questions. First, which companies are showing operating leverage from AI deployment, and which are showing balance-sheet leverage from stake markups? Microsoft's $37 billion AI run rate is operating revenue. Amazon's $16.8 billion Anthropic gain is a non-operating mark-to-market. The two are very different in quality, even if they both flow into reported earnings.

Second, what is the durability of each contribution? Cloud revenue commitments from OpenAI and Anthropic are long-dated contracts; the $250 billion Azure commitment and the $100 billion AWS commitment are scheduled spending over 6-10 years. Equity-stake markups, by contrast, are transient: they reverse if a future round prices flat or down. The 2026 numbers contain large amounts of both.

The shorter version: a meaningful slice of Big Tech's Q1 2026 reported performance came from Anthropic and OpenAI, more so than would have been visible from any single line item. The exposure is highly uneven across the five names. Microsoft and Amazon have the largest combined revenue plus equity exposures. Alphabet has the newest and fastest-growing equity exposure. Meta is largely insulated. Apple is on the buy side, paying for capability rather than booking gains from it.

Side-by-Side Summary

• Microsoft: ~27% OpenAI stake (~$135B), $250B Azure commitment by 2032, $627B RPO. AI run rate $37B (+123%). Q3 OpenAI investment hit was -$14M.
• Amazon: $8B+ Anthropic investment (now $70B+), up to $25B more committed. Anthropic AWS spend $100B / 10 yrs. Q1 non-op gain $16.8B, over half of pre-tax income.
• Alphabet: $40B Anthropic investment (April 2026), 5 GW TPU capacity, up to 1M TPUs. Google Cloud +63%, now 18% of total revenue.
• Meta: No major Anthropic/OpenAI deal. AI strategy via Llama (internal). Q1 revenue +33% largely organic.
• Apple: Google Gemini integration for Siri = cost, not revenue. Q2 results essentially independent of Anthropic/OpenAI.
• Two channels in play: cloud-revenue commitments (durable, long-dated contracts) and equity-stake markups (transient, mark-to-market).

The Question Behind the Question

Anthropic and OpenAI are themselves loss-making private companies, building and training models that consume the majority of the cloud capacity Microsoft, Amazon, and Alphabet are spending hundreds of billions to deploy. The hyperscalers are simultaneously their suppliers, their customers (through stake gains), and their counterparties on multi-year revenue commitments. The economic flow is circular in a way that does not fit cleanly into traditional supplier-customer analysis.

The Q1 2026 reports made the circularity visible. They did not resolve whether the underlying unit economics ultimately work for the AI labs themselves. If they do, the equity stakes that Microsoft, Amazon, and Alphabet hold compound further; if they don't, both the stake values and the cloud revenue commitments come under pressure at the same time. Big Tech's Q1 2026 results lean heavily on those two assumptions remaining intact.