OpenAI Series E

$157 billion for a company that still can't decide if it's a nonprofit, a capped-profit, or just straight-up begging VCs to let them pivot to whatever makes Altman's lawyers stop sweating. The valuation here is roughly equivalent to slapping a "priceless" sticker on the Mona Lisa if she was also maybe going to be seized by the French government next Tuesday. Sure, ChatGPT prints revenue—reportedly headed toward $4B annually—but you're paying a 40x multiple on a business model held together with duct tape, conditional investment terms, and the legal equivalent of a Jenga tower during an earthquake. Thrive, Microsoft, and Nvidia aren't dumb money, but they're also not exactly neutral observers when two of them literally depend on OpenAI's continued existence to justify their own AI infrastructure plays. This isn't price discovery; it's mutually assured valuation.

The timing screams late-stage froth even as everyone pretends we're still early. October 2024 finds us in this weird moment where AI hype has calcified into dogma but actual consumer behavior remains "I use ChatGPT to write emails I don't want to write." Enterprise adoption is real but commodifying fast—Anthropic and Google are snapping at heels, and open-source models are getting scary good at a fraction of the cost. Meanwhile, you've got regulatory scrutiny mounting globally, training costs that would make a defense contractor blush, and the nagging question of whether AGI is three years away or three decades. Raising $6.6B isn't a flex when you burn through billions like a Formula 1 team that also happens to be racing toward an finish line that may not exist.

Let's talk about the investor composition, which reads like a hostage situation disguised as a cap table. Microsoft's in for obvious reasons—they've already wired $13B+ and can't afford to let this investment crater without taking Azure's AI strategy with it. Nvidia needs OpenAI to keep buying H100s like they're going out of style, which they literally might be once the next architecture drops. Thrive is the only one here not selling picks and shovels, and even they hedged by reportedly negotiating some sweetheart terms that other investors didn't get. This isn't conviction; it's a Mexican standoff where everyone's too deep to walk away. The signaling value is real but increasingly circular: OpenAI is valuable because smart money says so, and smart money can't say otherwise without nuking their own positions.

The fundamentals are simultaneously impressive and terrifying. Revenue growth is genuinely bananas, but so is the cost structure—every breakthrough GPT model requires exponentially more compute, and we're approaching the point where training runs cost more than some countries' GDP. The nonprofit governance structure is a time bomb, the executive drama makes HBO look understated, and the entire enterprise rests on maintaining a technological lead that's already narrowing. Exit potential? Please. No tech company can acquire this without getting Lina Khan's boot so far up their ass they'll taste leather for years. IPO is the only path, and good luck explaining this capital structure to public market investors who think GAAP matters. This deal isn't terrible, but it's not good either—it's just very, very expensive uncertainty.