Anthropic Hits $965B, and Microsoft Profits Either Way

Anthropic closed a $65 billion Series H round at a $965 billion post-money valuation on May 28, edging past OpenAI to become the most valuable artificial intelligence startup on the planet. The figure sits a rounding error below $1 trillion, and it landed less than four months after the same company was valued at $380 billion.

For anyone running workloads on Microsoft’s cloud, the number that matters is who collects no matter which lab pulls ahead. Microsoft owns roughly 27% of OpenAI, separately committed up to $5 billion to Anthropic, and now sells Claude through its Azure-based Foundry catalog. The rival that just leapfrogged OpenAI is also a Microsoft tenant.

A $585 Billion Jump in Less Than Four Months

The Series H was led by Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital, with a deep bench of co-leads including Capital Group, Coatue, GIC and ICONIQ. Memory and chip suppliers Micron, Samsung and SK hynix joined as well, a sign of how tightly this financing is wired to the hardware that trains the models. You can read the full Series H investor roster on the company’s own site.

Run the arithmetic and the speed is the story. Anthropic raised $30 billion in February at a $380 billion valuation, terms laid out in the company’s February Series G announcement. Three and a half months later it is worth $965 billion, a $585 billion increase booked in a single quarter. That is more paper value created in roughly fourteen weeks than most listed companies carry in total.

The OpenAI comparison is what gives the round its bragging rights. OpenAI was valued at about $852 billion in late March after a record $122 billion raise. Anthropic’s new mark clears that on valuation and, by the company’s own reporting, on revenue too, the first time the upstart has led its larger rival on both counts.

Here is how the three rounds stack up.

Why Microsoft Profits Whichever Lab Wins

Microsoft spent years as OpenAI’s defining backer, and that bet is enormous: a stake of about 27% plus a share of revenue, a position reported to be worth somewhere near $135 billion. On paper, a competitor eclipsing OpenAI should sting.

Except Microsoft hedged. In November it agreed to invest up to $5 billion in Anthropic alongside Nvidia’s larger check, and Anthropic agreed to buy Azure compute in return. Claude reached Microsoft Foundry the same month. So when an enterprise picks Claude over a Microsoft-aligned model, a slice of that spend still routes back through Azure billing rather than walking out the door.

That is the quiet logic of the cloud business. Microsoft does not need to own the winning model; it needs the winning model to run on its infrastructure and clear through its contracts. Three numbers frame the position.

• 27% of OpenAI owned by Microsoft, plus a long-term revenue share.
• Up to $5 billion committed to Anthropic in November, beside Nvidia’s bigger stake.
• $30 billion in Azure compute Anthropic agreed to purchase as part of that deal.

What Azure’s Foundry Customers Get From Claude

Claude has been orderable inside Microsoft Foundry, the renamed Azure AI development platform, since November 18, 2025. The launch line-up covered Claude Sonnet 4.5, Haiku 4.5 and Opus 4.1, and the catalog has since added newer Opus releases as Anthropic shipped them, including the recent Opus 4.6 and 4.8 models. You can see the current Claude models in Microsoft Foundry on Azure in Microsoft’s catalog. The pitch to enterprise buyers is procurement, not novelty: Claude runs as a serverless deployment billed through existing Azure agreements, so a company can skip a separate Anthropic contract.

• Deployment through Foundry’s own APIs and tools, with Anthropic managing the underlying serving infrastructure.
• Billing that counts against a Microsoft Azure Consumption Commitment (MACC, the spend pledge large customers sign), removing a separate vendor invoice.
• Authentication via Microsoft Entra, with software kits for Python, TypeScript and C#.
• Access to successive Claude Opus updates through the same catalog entry instead of a fresh integration each time.

For a Windows-and-Azure shop, that last point is the draw, and the documented Claude deployment options inside Foundry spell out the steps: the model competing hardest with OpenAI’s GPT line shows up in the same console a team already uses to deploy everything else.

The Compute Bill Behind the Valuation

A near-trillion-dollar price tag is really a bet on compute, and Anthropic has signed for staggering amounts of it across every major supplier. The company runs Claude on three chip families at once: Amazon’s Trainium, Google’s tensor processing units (TPUs, Google’s custom AI accelerators) and Nvidia’s GPUs (graphics processors, the workhorse silicon for training large models).

The contracts read like national infrastructure budgets, and Anthropic’s multi-gigawatt compute commitments keep expanding.

• Amazon Web Services: up to five gigawatts of Trainium capacity, tied to a commitment exceeding $100 billion over ten years.
• Google Cloud: access to as many as one million TPUs, bringing well over a gigawatt of capacity online this year.
• Microsoft Azure: $30 billion of compute, with Claude slated to run on Nvidia Grace processors and Blackwell and Vera Rubin accelerators.

That spending is the counterweight to the valuation. Investors are underwriting a company that has promised to consume more electricity and silicon than many utilities plan for, on the wager that enterprise demand for Claude keeps compounding faster than the bills arrive.

The $47 Billion Run Rate, and the IPO Question

Valuations this size need a revenue story underneath, and Anthropic’s is the part even skeptics concede. The company says its annualized run-rate revenue crossed $47 billion in May, up from $30 billion earlier in the year and roughly $10 billion across the prior twelve months.

Where the Revenue Comes From

Most of the growth traces to Claude Code, Anthropic’s coding assistant, with enterprise accounts making up more than half of that product’s revenue. Named customers include Netflix, Spotify, KPMG, L’Oreal and Salesforce, the kind of logos that turn a model into a line item finance teams defend rather than cut.

Krishna Rao, Anthropic’s chief financial officer, tied the raise directly to that adoption.

Claude is increasingly indispensable to our growing global community of customers, and we work tirelessly to make tools like Claude Code and Cowork more helpful, more powerful, and more adaptable to their needs.

The pitch to investors writes itself from there: spend the fresh capital on safety research, on more compute, and on widening an enterprise base that already pays real money.

An IPO Without a Filing

The Series H is widely read as the last private raise before a public listing. Reporting has floated a possible offering as soon as the fourth quarter of this year, with bankers suggesting a raise above $60 billion, which would rank among the largest technology debuts on record.

The caution is that none of it is filed. Anthropic has not lodged an S-1 (the registration document a company submits before a US listing), named an underwriter or set a date, and as recently as December its communications chief said there were no immediate plans to go public. If the listing arrives on the rumored timeline, Microsoft will watch a company it partly funds price itself against the larger lab it partly owns. If it slips, the valuation set this week becomes the reference point every later round and rival is measured against.

Frequently Asked Questions

Is Claude available on Microsoft Azure?

Yes. Claude has been available through Microsoft Foundry, Azure’s AI development platform, since November 18, 2025, as a serverless deployment billed through existing Azure agreements. The catalog now includes recent Claude Opus releases alongside the original Sonnet, Haiku and Opus models.

Did Microsoft invest in Anthropic?

Yes. In November 2025 Microsoft agreed to invest up to $5 billion in Anthropic, alongside a larger commitment from Nvidia, and Anthropic agreed to purchase $30 billion of Azure compute as part of the same arrangement.

How much is Anthropic worth after the Series H?

Anthropic is valued at $965 billion post-money after raising $65 billion in its Series H round, which closed on May 28, 2026. That places it just below a $1 trillion valuation.

Is Anthropic now bigger than OpenAI?

By valuation and reported revenue, yes. OpenAI was last valued at about $852 billion in late March, and Anthropic’s new mark clears it on both price and annualized run-rate revenue.

Will Anthropic go public soon?

There is no filing yet. Reporting points to a possible initial public offering as soon as the fourth quarter of 2026, potentially raising more than $60 billion, but Anthropic has not submitted an S-1, named bankers or confirmed a date.