Stone Fox Capital
Founder, CIO & Portfolio Manager of Stone Fox Capital
Enterprises shifting to cheaper AI models blows away Cerebras's premise that fast inference is always better. Per UBS, via @rohanpaul_ai, 60% of companies watching AI budgets are moving to cheaper models and Chinese open-source models, routing easy tasks to cheap models and saving premium ones for hard reasoning, code and long-context work. x.comÂ
Cerebras and its customers likely accounted for many of the chip limitations alleged in the short report, so the report may be wrong. This is a suspicion without specific insight. Short interest rose 40.7% into the report to 19.86 million shares, or 8.4% of float. x.com x.comÂ
Cerebras looks attractive because the short report's two-year wafer-life thesis rests on unvalidated reliability assumptions, not representative failure data. Energy per useful token, lower warranty utilization and ordinary repair accounting also undercut its financial conclusions. OpenAI has integrated Cerebras into production; AMD and AWS have announced integrations, while Oracle plans 2027 availability for ultra-fast inference, adding commercial validation despite the report. x.com x.comÂ
Cerebras's yield problem is not solved, because the fault tolerance it claims comes at a price. Wafer-scale chips need testing, retesting of repairable defects and repair, and long bypass routes around defects can fail higher-speed specifications. @damnang2 lays out the details in a yield model for the WSE-3 built from Cerebras's public disclosures, though by his own account the figures rest on his assumptions, not Cerebras's actual yield. x.comÂ
Vicor's efficient delivery of key power modules is part of the case for Cerebras meeting and exceeding its 2028-29 numbers. As models improve, more security breaches and deployments should make Cerebras look increasingly like a cybersecurity play. Its revenue ramp looks constrained by capacity, not yet by demand. Bullish on Cerebras but cautious on the broader market, with no short-term earnings prediction. x.com x.com x.com x.comÂ
Better power delivery appears to be helping Cerebras work through wafer-scale power and thermal limits more than expected. CS-4 gets higher clocks and improved I/O from WSE-3T, despite the same core count and SRAM capacity as WSE-3. Oracle, OpenAI and AWS choosing to work with Cerebras carries weight. But higher clocks do not answer Culper Research's allegation that its wafers fail after about two years; how representative those failures are of current systems remains unclear. x.comÂ
Cerebras is a dud that looks set for new lows. It went public into a red-hot semiconductor sector and still failed, and with semiconductors now rolling over, it looks poised to break major support. x.comÂ
Cerebras may be the only company willing to stitch together a true high-speed solution. If so, it retains upside optionality from high-speed pricing and gains a first-mover advantage. x.com x.comÂ
Oracle's year-long evaluation of Cerebras' costs, performance, customer uses and equipment life strengthens the case for its inference technology. Oracle chose to bring the equipment into its data centers; that real-world buyer decision conflicts with the short seller report's criticism. x.comÂ
Cerebras is difficult for retail investors to understand, and management has not done a great job explaining its business since going public. It deserves more attention as the most direct AI IPO of 2026, but going long Cerebras is hard to justify when Nvidia can be owned instead. x.comÂ
Cerebras is now worth buying after its shares fell into the previously stated $60-$120 buying range. The earlier concern was an out-of-control valuation and providing liquidity to insiders at much higher prices; the shares shown are about $117. x.comÂ