TradFi Terminal › Markets › CBRS
Cerebras Systems (CBRS) stock perp on Hyperliquid
Trade CBRS 24/7 on Hyperliquid with a live chart, one-click order buttons and a sub-second news feed. Loading the live terminal…
| Price | $109.60 |
|---|---|
| 24h change | -7.35% |
| Funding (1h) | +0.0006% |
| Open interest | $27.6M |
| 24h volume | $25.9M |
| Max leverage | 10x |
Latest CBRS news
- — [twt] jukan05: A very sharp friend of mine sent me a take arguing that Cerebras chips could depreciate extremely quickly. I thought I’d share it here. Cerebras’ wafer-scale engine (WSE) and routing architecture do seem fully capable of solving the yield problem, as the company claims. The uncertainty is around how long the wafer will actually last in operation. Cerebras’ design requires a large number of redundant interconnects to accommodate manufacturing defects. But all that BEOL wiring significantly increases the effective area susceptible to TDDB (time-dependent dielectric breakdown), potentially creating much greater long-term reliability risks. TDDB can cause shorts, which are more dangerous than opens in this design. Dielectric thickness determines the probability of TDDB, and variations in that thickness can be spatially correlated across a wafer. That makes it highly likely that many interconnect failures would cluster in a particular region. It is unclear how Cerebras detects these shorts and avoids using the failed interconnects. Ultimately, the argument is that you have a design that is expensive to begin with and also ages extremely quickly.
- — [twt] culperresearch: 15/15) We estimate $CBRS private investors paid ~$23/share - far lower than todays prices... Insiders have sold $400M+ since the IPO. 19.4M shares unlocked today, and over 50M more will unlock in early November. We are short and think shares are headed lower.
- — [twt] culperresearch: 14) Before that, $CBRS CEO Feldman was at Riverstone, which received SEC charges, while Feldman plead guilty to one felony count of violating internal accounting controls. Riverstone would go bankrupt and return just $1.06 per share to shareholders, 92% lower than its spin-off.
- — [twt] culperresearch: 13) Nearly all of $CBRS co-founders and patent holders, as well as many key employees, came from Feldman’s last company, SeaMicro. AMD acquired the business in 2012, but within three years Feldman had left, while AMD took a $76M write-down and shuttered the business.
- — [twt] culperresearch: 12) $CBRS pitch largely rests on the claim that CEO Feldman and his co-founders solved what the semiconductor industry could not solve for over 75 years. Their track record does not inspire confidence…
- — [twt] culperresearch: 11) $CBRS bulls now view the OpenAI deal as a stamp of credibility. It's the opposite, in our view: $CBRS handed OAI a massive discount on compute, and after accounting for reduced useful lives, we believe the deal will prove value-destructive.
- — [twt] culperresearch: 10) ex $CBRS employees corroborated: "the customers won’t see reduced reliability, but Cerebras will have to put in more equipment as a buffer", "they’ve kind of obfuscated this challenge from the customer." We think $CBRS 60% gross margin targets are fantasy.
- — [twt] culperresearch: 9) Conveniently, $CBRS shift to the cloud hides this problem: $CBRS customers no longer see breakdowns, while in the cloud, $CBRS can quietly replace the broken system and record the cost as capex.
- — [twt] culperresearch: 8) $CBRS has already told on itself. Through Dec 2025, draft IPO filings gave useful lives of 1.5 to 3 years. Four months later, the Company nearly doubled useful lives to 3 to 5 years. The Company gave zero explanation for the change, which also inflated reported margins.
- — [twt] culperresearch: 7) Our own Arrhenius modeling - Company-favorably using only measured power draw - corroborates, suggesting $CBRS wafers age 3x to 3.5x faster than GPUs. If a GPU lasts 5-7 years, $CBRS wafers last just 1.4 to 2.3 years.
- — [twt] culperresearch: 6) As a result, $CBRS wafers break down rapidly: multiple customers told us their systems failed far sooner than expected: e.g., "it was about one year in that one region failed, so they had to replace the whole thing." Another said $CBRS had to replace their system twice.
- — [twt] culperresearch: 5) This jerry-rigged "fix" only swaps one issue for another: $CBRS wafers draw near-peak power constantly. A 2022 paper co-authored by a $CBRS engineer found the CS-2 draws 70% of peak at idle. A 2026 Harvard/Meta paper measured the CS-3 at 80% of peak when idle vs ~20% for GPUs.
Headlines are delayed. Sign in for real-time news.
More markets
SKHX · GOOGL · CL · SP500 · MU · XYZ100 · SKHY · SOXL · SNDK · DRAM · BRENTOIL · SILVER · SPCX · NVDA · SMSN · INTC · META · UNITREE · MSFT · MRVL · KORU · NBIS · AMD · TSLA · all markets