For weeks after Cerebras Systems (CBRS) went public on May 14, Wall Street analysts largely stayed silent.
That changed this week, which started June 7, when the post-IPO quiet period ended, unleashing a wave of coverage from several brokerages, including Morgan Stanley, which issued its first-ever price target on the AI chipmaker’s stock.
The AI chipmaker went public and soared nearly 70% on the first day, according to The Wall Street Journal. Later, it gave back a chunk of those gains, and then, the moment the post-IPO quiet period expired, several major brokerages hit the tape simultaneously with the same message: buy.
Morgan Stanley led the charge with a price target and a thesis on Cerebras that gets straight to the point, according to a note shared with TheStreet.Â
[It’s] one of the most differentiated AI infrastructure companies.
This is what Morgan Stanley called Cerebras, saying it offers something even Nvidia (NVDA) can’t replicate.
That’s a striking claim. Let’s carefully unpack the reasoning behind it.
Morgan Stanley sets $250 price target on Cerebras stock
Morgan Stanley analyst Joseph Moore, who ranks 152 out of 12,279 Wall Street analysts, initiated coverage with an overweight rating and a $250 price target in a note shared with TheStreet.
The core of his argument isn’t just that Cerebras makes a fast chip. It’s that Cerebras makes the only chip of its kind that’s actually deployed commercially at scale.
More AI:
- Micron sits at the center of a red-hot chip rally
- IBM CEO sends blunt message on AI and quantum computing
- Anthropic CEO makes shocking admission about AI
“We view Cerebras as one of the most differentiated AI infrastructure companies, built around the industry’s only commercially deployed wafer-scale processor,” Moore wrote.Â
“This is a unique chance to invest in an AI processor company with a first-mover advantage against NVIDIA, and offers substantial upside as the category evolves.”
Artificial intelligence (AI) workloads are shifting. The industry spent years focused on training — the process of building models. Now the action is moving toward inference — the moment a model actually generates an answer for a user. That shift rewards speed and low latency above all else.
Moore noted that Cerebras enters this inflection point backed by a large contracted backlog and a 750-megawatt committed capacity agreement, according to Cerebras’ press statement.Â
My review of the data suggests that a combination of proprietary hardware plus locked-in demand is exactly what institutional investors want to see in an early-stage infrastructure play.
Nagle/Bloomberg via Getty Images
What makes the Cerebras Wafer-Scale Engine unlike anything Nvidia ships
To understand why analysts are this excited, you have to understand what Cerebras actually built.
The Cerebras Wafer-Scale Engine (WSE) is the largest commercial chip on the planet, according to Cerebras. Where a standard graphics processing unit (GPU) is a discrete piece of silicon connected to other GPUs through complex networking, the WSE consolidates everything onto a single wafer-sized chip.
Related: Cerebras stock faces sharp reality check after massive $5.5B IPO debut
According to Cerebras, the WSE delivers inference speeds roughly 15 to 20 times faster than traditional GPU-based solutions. The key reason: All model weights live on-chip, eliminating the communication bottlenecks that slow down standard GPU clusters.
That speed advantage is most valuable in real-time applications, including coding assistants, research agents, and anything where a user is waiting for an answer. As AI moves deeper into everyday products, that category is growing fast.
Several other firms joined Morgan Stanley at the coverage launch, all with the equivalent of a buy rating. Price targets ranged from $250 to $340, according to a report by Investing.com.Â
Cerebras’ IPO numbers, and the OpenAI concentration risk investors can’t ignore
The story gets more complicated here, and it’s where you should pay attention.
The IPO itself was historic. Cerebras priced at $185 per share in an offering that raised $5.55 billion, the company statement shared.Â
Shares opened at $350, nearly double its IPO price, before pulling back 10% two days later toward the $200 level, my previous Cerebras report indicates.
As of Monday afternoon, June 8, CBRS was trading around $238.33, up more than 18% on the session, according to Yahoo Finance.
The financial backdrop includes:
- 2025 revenue of $510 million, up 76% year over year (YoY), according to Forbes
- A $24.6 billion contracted order backlog as of the end of 2025
- A $20 billion-plus compute deal with OpenAI through 2028, with an option for additional capacity.
- A growing inference collaboration with Amazon Web Services, OpenAI, Meta Platforms, and IBM.
Source: Cerebras Systems
That backlog number is genuinely extraordinary. But a substantial portion of it flows from a single customer relationship with OpenAI.
What Cerebras’ Wall Street debut means for the broader AI chip race
This story matters beyond CBRS itself. Here’s why.
Nvidia (NVDA) has dominated AI chip spending for years. Its GPU architecture became the default infrastructure for training the models that power everything from ChatGPT to enterprise software.Â
But inference, including the faster-growing, more recurring part of the AI compute stack, plays by different rules. Speed and latency matter more than raw processing power when a model is serving millions of live user requests. That’s the gap Cerebras is targeting.
And the fact that these firms launched coverage simultaneously, all bullish, suggests Wall Street believes the gap is real and the window is open.
Morgan Stanley’s $250 target implies roughly 6% upside from current levels. According to Seeking Alpha, the more aggressive targets include UBS at $300, Needham at $300, and Wedbush at $270.
All of them suggest the street sees more room if OpenAI and AWS deployments accelerate on schedule.
Related: Cathie Wood buys $32 million of popular semiconductor stock