OpenAI Cleans Its Cap Table Before a 2026 IPO

The buyback, in plain numbers

On August 10, 2026, OpenAI completed a tender offer that let current and former employees sell about $7 billion worth of company stock (roughly GBP 5.5 billion at today's exchange rate). Bloomberg and TechCrunch, both citing people familiar with the deal, reported that OpenAI's valuation held flat at $852 billion (about GBP 673 billion), the same figure the company reached in March 2026 when it closed a $122 billion primary funding round led by Big Tech and venture investors.

The detail that separates this tender from the ones before it is who wrote the check. No outside investor bought into this round. OpenAI funded the entire $7 billion repurchase itself, using company cash rather than fresh capital from a new backer, and declined to comment on the transaction when asked by reporters.

A break from OpenAI's own playbook

OpenAI has run employee liquidity events before, and neither prior round looked like this one. A 2023 tender backed by outside investors, including Thrive Capital and SoftBank Group, roughly tripled OpenAI's valuation to $86 billion. A later tender around October 2025 let more than 600 employees sell about $6.6 billion of stock at a $500 billion valuation, with roughly 75 people hitting an individual sale cap near $30 million; that round, too, brought in outside money.

This time OpenAI chose not to add a single new name to its capitalization table. That is a meaningful shift for a company that confidentially filed IPO paperwork with the U.S. Securities and Exchange Commission in June 2026: fewer investors holding a claim on the company going into a listing means fewer competing interests to reconcile once public-company disclosure and governance rules apply.

Anthropic tried the opposite play

OpenAI's closest rival took a different route to the same problem. Anthropic ran its own tender offer in April 2026, targeting $5 billion to $6 billion of employee stock at a $350 billion valuation, with the capital for the purchase assembled by outside investors rather than the company itself. Anthropic, too, has confidentially filed for an IPO in 2026.

Anthropic's tender fell short of its target. Enough employees chose to hold their shares rather than cash out, betting that a public listing would eventually pay more than a $350 billion valuation was worth in April. Between the two companies, employees have now sold well over ten billion dollars of pre-IPO stock in a matter of months, one of the largest waves of private tech wealth creation on record, reached through opposite financing choices.

What it means for the European businesses that depend on OpenAI

For a European company running production workloads on OpenAI's API or Enterprise tier, a cleaner cap table ahead of an IPO is a genuine, if modest, stability signal. Fewer new investors means fewer new voices with a claim on strategy, pricing, or board seats right as the company adapts to public-market reporting and disclosure obligations, the period when governance disputes at newly listed companies most often surface.

Set that against a second, less comfortable fact from the same reporting: earlier this year, the Wall Street Journal reported that OpenAI missed its own internal financial targets, and Sam Altman has publicly acknowledged that the past twelve months fell short of the company's own expectations. A company that just admitted missing its numbers, then chose to spend $7 billion of its own cash on internal buybacks rather than raise fresh capital, is a data point worth weighing alongside the stability signal, not a reason to panic, but a reason to read the eventual IPO prospectus and any contract renewal terms carefully.

What to watch before OpenAI lists

Three things will tell European buyers more than this tender offer alone: whether OpenAI actually files publicly and prices an IPO before the end of 2026 as reported, whether Anthropic's employees keep holding rather than selling in any future tender, and whether other frontier labs preparing to list, xAI among them, follow OpenAI's self-funded approach or Anthropic's investor-backed one.

None of this changes what OpenAI's models can do today. It changes how a finance or procurement team should read the company behind them: one that is choosing, for now, to keep new investors out of the room even as it prepares to answer to public shareholders, a stance worth remembering the next time a multi-year enterprise contract comes up for renewal.

Read next: An AI Agent Faked Identities to Get Code Approved  |  Third AI Lab in Two Weeks Reports Same Sandbox Failure