
OpenAI completed a roughly $7 billion employee tender on Monday at $852 billion, unchanged from its last primary round, and funded the purchase itself rather than bringing in outside investors. The OCC denied bunq a national bank charter in the same week that Erebor moved toward a raise at roughly $8 billion. And Sacra estimates Notion has grown 82% since the tender that set its current mark.
OpenAI completed a tender offer on Monday letting current and former employees sell roughly $7 billion of stock. The price valued the company at $852 billion, the same figure as its most recent primary round.
The buyer was OpenAI. The company used its own cash rather than bringing outside investors into the tender, which leaves the shareholder base unchanged going into a possible listing. Because the tender was company-funded rather than funded by new outside investors, it did not introduce new third-party purchasers through the tender. OpenAI submitted a confidential draft registration statement to the SEC in June. No offering has been scheduled and none may occur.
Last Tuesday this newsletter described employees at the largest private companies declining to sell into tenders while buyer demand ran past available supply. About $7 billion has now moved. It moved to the issuer.
The distinction is worth holding onto. When outside investors buy into a tender, the price is agreed between a seller and a buyer who has other places to put the money, and the number that results carries some information about what a third party would pay. In a company-funded buyback the issuer sets the terms and is the only bid. The $852 billion figure was produced by that second process, and it matches the last primary round, with no change in the five months since.
That does not make the number wrong. It means a reader placing it next to an indication from a brokered venue is placing two numbers side by side that were arrived at differently, and only one of them involved a third-party buyer.
Reported private-company financials and secondary-market indications may be unaudited, incomplete, non-standard, or based on limited transaction activity. They should not be relied upon as fair value, executable pricing, or a basis for any investment decision.
For more company information, recent coverage, and available market data, visit OpenAI on Augment.

The Office of the Comptroller of the Currency rejected the Dutch neobank bunq's application for a US national bank charter in a decision letter dated August 4, reported on August 7. It was bunq's second attempt; the company withdrew an earlier application in 2024 and has said it plans to reapply.
The stated reasons are specific rather than procedural. According to the decision, the OCC found the business and marketing plans inadequately supported and unrealistic given competition in US unsecured credit cards, found that the applicant had not planned for the expenses of competing without name recognition in the US market, and found that the proposed directors did not demonstrate an understanding of differences between US and European credit markets.
The same agency reached the opposite conclusion earlier this year with Erebor, the bank founded by Palmer Luckey and Joe Lonsdale to serve crypto, AI and defense companies, which is reported to be in talks to raise about $1.5 billion at roughly $8 billion. That round has not closed.
Charter optionality sits inside a lot of late-stage fintech storytelling: the argument that a US banking licence is a matter of time, lawyers and patience. The OCC has now published a denial of a capitalised European incumbent on grounds that go to the substance of the plan. Whether that changes how any particular company is valued is not something public reporting establishes, and the decision addresses one application rather than a sector.
Sacra estimates that Notion reached $865 million in annual recurring revenue in July, up 82% year over year. It puts the company's most recent valuation at $11 billion, set in a January 2026 tender, which works out to roughly 18 times that revenue estimate.
Both halves of the ratio need a caveat. The revenue figure is an outside estimate rather than a company disclosure. The valuation is a real transaction price, but it was set in January, inside a process the company ran, the same way the OpenAI number above was set.
Seven months separate the two figures. A public company would have filed twice in that window. A private company's mark holds until it chooses to produce another one, and a buyer reading $11 billion today is reading a number that described a business with a materially smaller revenue base than the one Sacra now estimates.
Sacra's estimate is not audited, and which direction the gap cuts is not something this newsletter can tell you. It is measurable, which distinguishes it from most of the other inputs a private-share buyer works with.
Reported private-company financials and secondary-market indications may be unaudited, incomplete, non-standard, or based on limited transaction activity. They should not be relied upon as fair value, executable pricing, or a basis for any investment decision.
That is the amount of third-party capital that Nvidia and six alternative managers say they intend to mobilize for AI compute infrastructure under memorandums of understanding announced Monday. Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR are the named partners, and the platforms are described as covering chips, power generation and data centers. The financing for the AI buildout is being arranged by firms that raise from pensions and sovereign funds rather than by the venture funds that seeded the companies doing the building.
Appraisal smoothing is what happens to a series of prices when the prices are appraisals rather than trades. A public stock is repriced by whoever is willing to transact, continuously. A private asset is repriced when someone decides to reprice it — a funding round, a tender, an annual valuation — and in between, the last number stands in for the current one.
The effect is that a series of private marks may look steadier than the business underneath it. Movement that a public market would register the week it happened can show up in a private mark months later, or in a single step when the next transaction resets the number. Some of the smoothness is a feature of the measurement schedule rather than the asset.
Which makes two things worth knowing about any private valuation before the number itself: when it was set, and by what process.
Augment Markets Inc. is a technology company offering software and data services. Brokerage services are offered through Augment Capital LLC, an affiliated broker-dealer and member FINRA/SIPC. Investment advisory services are offered through Augment Advisors LLC, an SEC-registered investment adviser.
Important Disclosures: This material has been prepared for informational purposes only. None of the information provided represents a recommendation, an offer or the solicitation of an offer to buy or sell any security. The information provided does not constitute investment, legal, tax, or accounting advice. You should consult with qualified professionals before making any investment decisions. Investing in private securities involves substantial risk, including the potential loss of principal. Private securities are typically illiquid, have limited pricing transparency, and often require longer holding periods. These investments are available exclusively to qualified accredited investors and offer no guarantee of returns. An IPO or other liquidity event is not guaranteed. Additionally, past performance of private securities does not indicate or predict future results. Share price data are estimates only, based on proprietary data from Caplight and Augment Markets Inc. and its affiliates.