Cerebras completes IPO, priced above range and 20x oversubscribed
Cerebras Systems (wafer-scale AI inference cloud, Nvidia challenger) completed its IPO in the week of 2026-08-03, pricing above its initial range and reportedly 20x oversubscribed on debut, before shares fell over the following days. Confirmed via Stocktwits, The Motley Fool, TradingKey, and renaissancecapital.com. Cerebras's usage-based per-token API pricing (Free Trial, $10 Developer, Enterprise) plus the fixed-price Cerebras Code subscriptions were unchanged at capture time.
Cerebras Systems, the wafer-scale AI inference cloud and hardware company positioned as an Nvidia challenger, completed its long-anticipated IPO in the week of 2026-08-03. Coverage (Stocktwits, The Motley Fool, TradingKey, renaissancecapital.com) reports the offering priced above its initial range and was roughly 20x oversubscribed on its first trading day, though shares subsequently pulled back from the IPO-day pop (The Motley Fool notes an over-58% decline from the first-day high within the same week).
No pricing changes accompanied the listing — Cerebras’s per-token inference API (GPT-OSS-120B, Gemma, etc.), $5 free-trial credits, $10 self-serve Developer tier, and the fixed-price Cerebras Code plans ($50/$200) remain as last captured. This is purely a capital-markets/liquidity event.
Detection note: caught via Layer-0 Google-News corporate-event scanning — an IPO does not touch a company’s live pricing page, so Layer-1 diffing was structurally blind to it.
Consistent with the deprecation date footnoted on the pricing page a month earlier, ZAI-GLM-4.7 ($2.25 input / $2.75 output per million tokens) was removed from Cerebras's public "Developer Tier Pricing" rate card and the docs Model Catalog. Confirmed via a fresh capture on 2026-08-26, which shows the rate card and Model Catalog both down to two models — GPT-OSS-120B (production) and Google Deepmind Gemma 4 31B (Preview). Z.AI's GLM 4.X and GLM 5.X model families remain reachable only through Dedicated Endpoints on custom reserved-capacity pricing; no per-token price moved on the two remaining public-card models.