OpenAI's $7 Billion Employee Share Buyback

OpenAI structured the $7 billion sale as a company-funded repurchase of shares from current and former employees rather than routing outside investors into a fresh funding round ^[1]. That let OpenAI relieve near-term liquidity pressure on staff sitting on large amounts of paper wealth while the company stays private ^[1], and it priced the deal at the $852 billion valuation established by the record $122 billion round that closed March 31, 2026, rather than negotiating a new number ^[1][2]. It is the same playbook OpenAI signaled in June, when Sam Altman told staff a tender was coming 'very soon' at $687.69 a share ^[3]- anchor to the last primary-round price and skip the harder work of discovering a fresh one.

This is OpenAI's third major employee liquidity event since late 2024: a SoftBank-backed $1.5 billion tender in November 2024 ^[5], a $6.6 billion sale in October 2025 that valued the company at $500 billion after demand hit $10.3 billion against an initial $6 billion offer ^[4][6], and now $7 billion in August 2026 at $852 billion ^[1]. Between the October 2025 and August 2026 tenders, the valuation climbed 70 percent in ten months, driven mostly by the intervening $122 billion round ^[2]. Roughly every six to ten months OpenAI is opening a new pressure valve for equity-rich staff rather than pushing straight to a public listing.

More than 600 employees took part in the October 2025 tender, with about 75 hitting the $30 million per-person cap ^[7]. J. Thelander Consulting founder Jody Thelander says that reflects how staff at a company of this scale think about a sale: 'With a company of this magnitude, [employees] think the company is going to be worth more than any secondary price' ^[4]. Carta's Peter Walker groups OpenAI with SpaceX and Anthropic as companies where employees behave nothing like a typical late-stage startup: 'There is a world of difference between three companies and everybody else...it's night and day' ^[4]. But Sydecar COO Shriram Bhashyam flags the tension baked into this cycle - the closer OpenAI gets to an actual IPO, the less reason employees have to keep selling into a fixed-price tender: 'There is less incentive to sell into a tender the closer a company gets to an IPO, as the employees can sense a liquidity event on the horizon' ^[4].

Not everyone reads the tender as routine liquidity management. Public discussion of the underlying $852 billion valuation skews openly skeptical, framing the repeated liquidity events as insiders cashing out ahead of a retail-facing IPO rather than as confidence in the business. Some of that skepticism targets the math directly - questioning how a company burning heavily against roughly $20-25 billion in annual revenue justifies the multiple, and whether OpenAI's consumer-first ChatGPT business can hold its edge as Anthropic gains ground with enterprise customers. Prediction markets echo caution rather than certainty: pricing puts only about a one-in-five chance that OpenAI actually announces an IPO within 2026, even as the company completes tender after tender at IPO-adjacent valuations.

Goldman Sachs and Morgan Stanley are competing for the 'lead left' bookrunner slot on OpenAI's anticipated listing, a role IPO expert Jay Ritter says is worth 'multiples of the potential rewards' compared to a secondary underwriting position ^[9]. Prediction markets currently favor Goldman at 73 percent odds ^[9]. OpenAI is reportedly targeting a listing that could raise upward of $60 billion, potentially as soon as fall 2026 ^[9], following its confidential SEC filing in June 2026 ^[8]. The stakes of getting the timing wrong are already visible: reports that the IPO could slip to 2027 sent Goldman Sachs shares down as much as 4.8 percent and Morgan Stanley down as much as 4.1 percent in a single session ^[10]- a reminder that OpenAI's IPO calendar now moves bank stocks, not just its own cap table.