Analysis
I keep hearing founders and board members open IPO-timing conversations with some version of Cerebras's number -- the May IPO that priced at a $56.4 billion valuation and closed its first day of trading up 68%, carried by real revenue and a marquee OpenAI compute deal. I keep giving the same answer: that's not your comp. Accelevation is.
Accelevation priced its IPO months later at $660 million, the bottom of its range and a markdown from its original target, despite carrying $727 million in trailing revenue -- a real, substantial business by most standards. It's been sliding since it started trading, a story we covered here. That's the realistic outcome for almost every company that isn't walking in with Cerebras's specific combination of frontier-AI compute demand and a marquee anchor customer, and it's a far more useful anchor for board conversations than whatever number made headlines in May.
“It's been sliding since it started trading, a story we covered here.”
The mistake I keep seeing is treating 'the IPO window is open' as a single statement about the market, when it's actually two different markets wearing the same calendar. One prices anything that can plausibly claim frontier-AI exposure with a marquee customer attached at an enormous structural premium. The other prices everything else -- infrastructure, industrials, even profitable SaaS -- close to or below where private markets already had it. Confusing which market you're actually in is how a board ends up disappointed by a banker's pitch deck full of the wrong comps.
Room for disagreement: a skeptic would point out that Accelevation is one data point, not a trend, and that post-IPO price action in the first few days of trading is notoriously noisy -- plenty of eventual winners have opened weak. It's also true that being an AI company isn't binary; plenty of non-frontier-lab businesses have real AI-driven revenue growth and could plausibly borrow some of that premium if they tell the story well. I'd push back that the gap between Cerebras's reception and Accelevation's isn't a few points of first-week noise, it's an order of magnitude in how the market is pricing risk, and that's not something a good banker deck closes on its own.
Know which market you're actually selling into before you let anyone show you a comp table.