Cerebras sinks below its $185 IPO price after its October 1 lockup expiry

At roughly $165, a buyer at the $350 opening print keeps about 47 cents of each dollar paid ^[16], and a buyer at the $311 first-day close keeps about 53 cents ^[15]. The less-than-half figure depends on the day, though. At $177, the top of the early-October range, the open buyer keeps about 51% ^[17].

Cerebras itself was paid in May. It sold its 30 million shares at $185 ^[1][2], so the decline moves money between shareholders and leaves the May proceeds where they were. Judged by the first day, the offer was cheap. Judged by early October, it was about $20 a share too high ^[6].

Crypto Briefing says timing played a part, pointing to the October 1 lockup expiry and about $84 million of insider sales around that date ^[7]. That is a small sum for a company this size. Dividing the $56.43 billion fully diluted valuation ^[3] by the $185 offer gives roughly 305 million shares ^[20], worth about $50.3 billion at $165 ^[21]. The insider sales come to about 0.17% of that ^[22]. The account dates the slide only to early October and does not give a September price, so it cannot show how much of the fall came before the lockup ended.

At $50.3 billion, the stock trades at about 99 times 2025 revenue ^[23] and about 212 times 2025 net income ^[24]. The same share count at the $350 open implied about $107 billion, or roughly 209 times revenue ^[25]. Net income improved by about $720 million in a year when revenue grew by about $220 million ^[26]. Profit, in other words, moved more than three times as far as sales did. The Motley Fool, as cited by Crypto Briefing, said "Cerebras is about as big today as Nvidia's data center business was nearly a decade ago," and argued that the similarities mostly end there ^[11].

One reading of $165 is that the first-day premium drained away over five months and says little about the business. A second is that insiders sold the first shares they were allowed to sell, and that the timing is the signal. A third is that buyers are pricing in the two concerns Crypto Briefing lists, customer concentration and competition from Nvidia ^[12]. We think the first explains most of the move, because $84 million of selling is too small to push a $50 billion company down by itself ^[22]. The counter-case is that size is the wrong test: insiders selling at the first chance tells the market something at any volume. Heavier insider selling from here, arriving alongside further declines, would prove us wrong.