Cerebras Enters November With Two Cheers From Wall Street and a Fresh Supply of Shares

AI & Compute

Cerebras Faces $228 Million Insider Sales and Cut Price Targets

Wedbush and Morgan Stanley cut Cerebras targets to $190 and $178 while keeping buy ratings, as 71.4% of Class A shares exited lockup and insiders sold $228 million.

By
Grace Kim
Filed
Channel
AI & Compute
Read
3 min read

Wedbush cut its Cerebras Systems price target to $190 from $290 on Monday, and Morgan Stanley lowered its target to $178 from $279 the same day. Both firms kept their buy-equivalent ratings. The same day, roughly 71.4% of Cerebras' Class A shares became freely tradable after contractual lockup restrictions expired, according to The Inference.

The twin target reductions reflect falling valuation multiples across semiconductor and hardware stocks, Wedbush said. Morgan Stanley maintained its Overweight stance while trimming its number. The companies still endorse the stock — they simply value it less generously than before.

By Grace Kim, Chip Dispatch

What did the insider sales disclose?

The lockup expiry arrived alongside a disclosure that Chief Technology Officer Sean Lie and Chief Operating Officer Dhiraj Mallick sold shares worth about $228 million combined ahead of the deadline. The sales ran through pre-arranged trading plans under Rule 10b5-1-style schedules, The Inference reported.

The enlarged float and the insider sales land days before the company's next earnings report. Cerebras reports third-quarter results on November 12 after the US market close.

What has management guided to?

The quarter will be read against a demanding backdrop. Management has guided to:

  • Core revenue of $214 million to $216 million for the third quarter of 2026
  • Gross margin of 38% to 40% in the core business
  • Operating margin of minus 25% to minus 23%

That operating loss is the crux of the cautious case. Revenue at the top of the range only makes a strong argument if margins hold up alongside it, and the guidance currently pairs healthy top-line growth with a business still burning money at the operating line. Investors would do well to weigh the two figures together rather than treating the revenue range as a verdict on its own.

How large is the backlog?

Underneath the quarterly guidance sits a much larger promise. Cerebras put its remaining performance obligations at $25.4 billion as of the end of the second quarter of 2026, according to The Inference. The overwhelming majority of that figure is attributable to OpenAI.

That customer concentration cuts both ways. A backlog of that size can underwrite expansion, but it also ties the valuation tightly to a single counterparty, which makes execution and margin progression the metrics that matter most. Capacity announcements are only half the story; the other half is what those deployments earn.

What is the durability dispute?

How long the hardware actually lasts has become a valuation question in its own right. Culper Research alleged on October 28 that Cerebras wafers have a functional lifespan of just 1.5 to 2.5 years, and it also took aim at the accounting assumptions behind the company's depreciation schedule. Cerebras rejected the claim. Per The Inference, the chief executive said fewer than 0.5% of deployed CS-3 wafers have been retired.

The dispute has now drawn legal attention. Pomerantz said Monday it is investigating possible securities violations by Cerebras and its executives, and its probe takes up the Culper allegations as well. That raises the premium on verifiable disclosure about both the technology and the accounting.

What should investors watch on November 12?

The earnings report offers the next real checkpoint. The yardstick worth applying combines three measures: margin performance, capacity build-out, and a defensible view of equipment life. Those three pieces determine whether the capital being deployed earns an adequate return.

Analyst enthusiasm and a freshly unlocked share count can move the stock in the near term, but neither settles that arithmetic. Cerebras can capitalize on the build-out of AI infrastructure if its systems last and its contracts pay; how much patience the long-term story deserves hinges on exactly that.

Source: newscase.com

Share this article:

More from Grace Kim

Grace Kim

Show full bio

Market editor covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

879 articles

Related articles

« Previous article