Mercor co-founder Brendan Foody publicly accused Sequoia of using dual-tranche financing to inflate headline valuations of AI startups, saying he has counted roughly six such rounds in the last six months. Foody, whose AI talent platform was last valued at $10 billion, posted the criticism on X, where Sequoia partner Shaun Maguire responded directly. The exchange opens a rare public window into a pricing mechanic that has quietly spread across the hottest AI deals of 2026.
The mechanic works like this: a lead firm puts the bulk of its capital in at a lower preferred valuation, then writes a much smaller check at a far higher price. The higher figure becomes the press-release number. The blended entry point, which is what the lead actually paid on average, never appears in the announcement.
The Serval round is the cleanest example in the public record. The AI IT-helpdesk startup announced a $75 million Series B at a $1 billion valuation. Sequoia's lowest entry point valued Serval at $400 million, less than half the headline figure. Aaru, which uses AI to simulate user behavior for market research, announced a $1 billion price tag while lead investor Redpoint actually backed the company at $450 million.
Key facts
- 01Mercor co-founder Brendan Foody accused Sequoia of running dual-tranche rounds in roughly six deals over the past six months.
- 02Serval announced a $75M Series B at a $1B valuation, but Sequoia's lowest entry point valued the company at $400M.
- 03At Aaru, lead investor Redpoint backed the company at $450M despite an announced $1B headline price.
- 04Sequoia's Shaun Maguire said dual-tranche structures occurred about five times in his seven years at the firm.
- 05Mercor was last valued at $10B, making Foody one of the more prominent founders to publicly criticize a top-tier VC.
Foody argued the practice harms two groups specifically: employees, who are sold the headline number when they evaluate their stock options, and angel investors, who get pitched the same inflated mark when founders shop subsequent allocations. He acknowledged Sequoia is not the only firm using the structure but said the firm appears to use it most frequently.
Mercor itself was a Sequoia miss, which Maguire conceded in his reply. That context matters: Foody is calling out a firm that passed on his company, and he is doing it from a $10 billion perch. The post landed amid a broader wave of founders and former founders airing VC grievances on X, but Foody's was the one that named a firm.
Maguire framed the practice as a market reality rather than a deliberate maneuver. He said Sequoia is often unwilling to match the multiples other investors will pay for hot AI deals, so it splits its check to preserve the partnership without overpaying for the bulk of the capital. He put the frequency at about five times across his seven years at the firm and said he was not aware of anything shady, adding that venture is a repeated game and misleading people would not pay off.
The question Maguire did not address is what founders tell people who do not know about the lower tranche. Employee 409A appraisals, in theory, should price options off the blended valuation across all tranches rather than the announced figure, according to Jason Woo, a partner in valuation at Armanino. In practice, 409A valuations skew low by design, because a lower strike price means a smaller tax bill for the company. The appraisal meant to protect employees from inflated headlines is structurally incentivized to come in conservative, which means the gap rarely gets corrected by anyone with a fiduciary duty to flag it.
Angels face a sharper problem. There is no independent appraiser between an angel and whatever number a founder shares over a coffee. If the founder cites the headline valuation without mentioning the lower tranche, the angel is buying at a price the lead investor explicitly declined to pay for most of its own capital. That is the misrepresentation Foody is pointing at.
Dual pricing is one of several ways AI-era metrics have drifted from the underlying business. Niko Bonatsos, a longtime General Catalyst partner who recently founded Verdict Capital, described a parallel pattern around annual recurring revenue at an industry event in Athens last month. He recounted founders citing ARR numbers that turned out to be a single strong day multiplied by 365, a practice that has effectively rendered the term meaningless in some pitch decks.
The structural read here is that AI valuations have detached from the normal repair mechanisms of venture pricing. When the lead investor's true entry point is private, the 409A is structurally low, and ARR is reported on a trailing-day-times-365 basis, the public valuation becomes a marketing artifact rather than a price discovery signal. That works fine on the way up. It tends to work badly on the way down, when secondary buyers, acquirers, and IPO bankers reprice the cap table against reality and the gap between the headline and the blended number becomes someone's loss. Foody is not the first founder to notice this, but at a $10 billion valuation he is the loudest, and Sequoia is now the firm that has to answer the question on the record.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
