Is bigger always better? Cerebras [CBRS] reckons the answer is yes.
The Sunnyvale, California-based firm designs and manufactures the largest artificial intelligence (AI) chip in the history of the semiconductor industry. It is often compared to a dinner plate in terms of size but, at a price tag of a few million dollars, it’s considerably more expensive. Termed a Wafer-Scale Engine (WSE), the Cerebras chip boasts 4tn transistors and 46,225 mm2 of silicon, compared to 208bn transistors and 1,600mm2 silicon for a Nvidia [NVDA] two-chip B200 package.
The company both sells its WSEs to hyperscalers and runs a network of data centres powered by its WSEs, offering AI processing services via its cloud platform, since its chips are too large to fit in regular computer hardware.
The scale of WSEs allows them to store and manipulate large amounts of data without needing information from other components, which, the company claims, results in superior performance and speed for demanding AI workloads, especially inference.
Cerebras CEO Andrew Feldman has hinted at plans to boost the use of Cerebras chips among hyperscalers beyond key partners such as Amazon [AMZN] and OpenAI. He’s not the only one with high hopes in this regard – as the post-IPO quiet period ended in early June, a wave of bullish coverage from no fewer than nine Wall Street brokerages reignited interest in the firm’s investment case.
In this article, CMC Aureon investigates if this pre-revenue custom ASIC maker can meet soaring investor expectations and potentially even compete with the undisputed king of AI chips, Nvidia.
“Everyone but them”
Cerebras is targeting partnerships as a way to make its chips an essential component of AI workloads. It signed a deal with OpenAI in January, in which the model developer will use 750MW of Cerebras WSEs for AI processes, touted as the “largest high-speed AI inference deployment in the world”. The partnership was expanded further in April, with OpenAI agreeing to pay $20bn over the next three years to use servers powered by its WSEs.
On 13 March, Amazon [AMZN] announced it would deploy Cerebras CS-3 systems at AWS data centres. Also in March, during the Q3 earnings call, Oracle [ORCL] CEO Clay Magouyrk named Cerebras as an important chipmaker, prompting speculation that the computing giant could be a new cloud customer for the firm.
It is clear that Cerebras hopes to shake fears it relies on just a few key customers. Speaking at the Bloomberg Tech conference in San Francisco on 4 June, CEO Feldman outlined plans to collaborate with a range of firms in the AI data centre ecosystem, with one notable exception: Nvidia. In fact, the smaller firm seems to be increasingly positioning itself as an alternative to Nvidia, especially for AI workloads such as inference and coding, which require greater speed and token throughput than traditional GPUs can provide.
Boom and bust… and boom?
Cerebras originally teased an IPO in 2024, but delayed plans after increased scrutiny of its reliance on a single client, UAE-based AI firm G42. It debuted on the Nasdaq on 14 May, with shares priced $185, raising $5.6bn in the biggest IPO of 2026 so far – though soon to be left in the dust by SpaceX’s 12 June debut. Shares opened at $350 and rose as high as $385 before closing at $311.07.
The red-hot debut allowed Cerebras to qualify for fast-track inclusion into the S&P Dow Jones Indices, effective as of 25 May. CBRS shares have since slumped – not uncommon post-IPO – and were down over 40% by 5 June. On 8 June, however, following the initiation of coverage from a number of Wall Street analysts, CBRS shares surged 18.32% in a single session, hinting at a potential rebound.
David and Goliath: CBRS vs NVDA vs CRWV
Impeccable timing and hype aside, Cerebras is competing in a market with a single, undisputed leader. Even if its unique product can capture a significant portion of that market, it still has a lot of work to do to even come close to Nvidia.
While Cerebras has yet to release its Q1 2026 earnings, in its IPO prospectus it outlined its performance over the past few years. Figures for Q4 2025 show revenue of $171m, up 96% year-on-year. In comparison, Nvidia reported $81.62bn in revenue for Q1 FY2027, up 85% y/y. That figure, the Wall Street Journal notes, is more than the full-year revenue of any other chip company apart from Taiwan Semiconductor Manufacturing Co [TSM], and is nearly 500 times Cerebras’ revenue. And its growing at a similarly rapid rate.
That said, the market seems poised for alternatives, with hyperscalers looking to both in-house solutions and smaller chip firms to reduce their reliance on Nvidia GPUs, and Cerebras could well benefit as a new go-to provider of inference and high-speed compute hardware. CEO Feldman has emphasised the ease of making the change, claiming that “you can switch from a workload on Nvidia to a workload on Cerebras in about 10 keystrokes” in an interview with the Wall Street Journal in May.
In terms of its cloud platform – offering remote use of its data centre network, powered by WSEs – Cerebras could be compared to CoreWeave [CRWV], which offers software and cloud services from data centres using its stockpile of Nvidia GPUs. In Q1 2026 CoreWeave reported revenue of $2.08bn, up 111.6% y/y, and a net loss of $740m, up from a loss of $315m in the year-ago quarter due to aggressive infrastructure scaling.
Here is how the three stocks compare in terms of fundamentals.
| CBRS | CRWV | NVDA |
Market Cap | $47.30bn | $58.94bn | $4.9tn |
P/S Ratio | 87.65 | 8.77 | 19.79 |
Estimated Sales Growth (Current Fiscal Year) | N/A | 146.89% | 81.38% |
Estimated Sales Growth (Next Fiscal Year) | 239.92% | 96.75% | 40.43% |
Source: Yahoo Finance
CBRS stock: The investment case
The bull case for Cerebras
On 8 June, multiple Wall Street firms initiated coverage of CBRS stock, with the consensus being overwhelmingly bullish. Analysts underlined the chipmaker’s “unique” product and its capacity to cater to inference needs, as well as its partnerships with OpenAI and AWS.
“As AI workloads become increasingly reasoning-intensive, demand for fast, low-latency inference is growing rapidly … we believe Cerebras is well positioned to capture this opportunity,” wrote Morgan Stanley analyst Joseph Moore in a note to clients, initiating coverage with an ‘overweight’ rating and a $250 price target. “This is a unique chance to invest in an AI processor company with a first-mover advantage against NVIDIA.”
UBS and Needham both initiated coverage with a ‘buy’ rating and a $300 price target, while Wedbush initiated with an ‘outperform’ rating and a $270 price target.
Cathie Wood seems convinced. She has been scooping up CBRS shares since the IPO, and holds $29.1m as of 8 June across her ETFs, accounting for 0.51% of total holdings.
The bear case for Cerebras
While CBRS seems headed for an upswing after its post-IPO quiet period, there are still a number of risks that could undermine its long-term prospects. Importantly, Cerebras continues to operate at a loss. In its IPO prospectus it touted swinging from a net loss of $482m in 2024 to net income of $238m in 2025, though much of this reflected a one-time accounting gain tied to a contract with G42. Furthermore, some 86% of its 2025 revenue came from two UAE-based clients.
Additionally, the stock could already be overvalued. Its P/S ratio of 87.65 is well above Nvidia’s, suggesting the market has already priced in years of growth. While it could capture an increasingly significant market share, it is unlikely to replace Nvidia, especially as the larger firm works to improve its low-latency offerings through moves such as the quasi-acquisition of AI chip startup Groq in December 2025.
Conclusion
Boasting a unique product tailored to low-latency, high compute AI workloads and several promising partnerships, Cerebras could emerge as a serious challenger to Nvidia’s dominance of the semiconductor industry in the coming years. Before it does, however, it must ensure a consistent profit and protect its first-to-market edge in a fiercely competitive space. Until then, CBRS stock is likely to see its fair share of ups and downs.
CMC Aureon’s proprietary theme relevance system maps the world’s biggest investing megatrends. For in-depth analyses of stocks with high growth potential, subscribe to CMC Aureon Foresight.
Disclaimer Past performance is not a reliable indicator of future results.
CMC Markets is an execution-only service provider. The material (whether or not it states any opinions) is for general information purposes only, and does not take into account your personal circumstances or objectives. Nothing in this material is (or should be considered to be) financial, investment or other advice on which reliance should be placed. No opinion given in the material constitutes a recommendation by CMC Markets or the author that any particular investment, security, transaction or investment strategy is suitable for any specific person.
The material has not been prepared in accordance with legal requirements designed to promote the independence of investment research. Although we are not specifically prevented from dealing before providing this material, we do not seek to take advantage of the material prior to its dissemination.
CMC Markets does not endorse or offer opinion on the trading strategies used by the author. Their trading strategies do not guarantee any return and CMC Markets shall not be held responsible for any loss that you may incur, either directly or indirectly, arising from any investment based on any information contained herein.
*Tax treatment depends on individual circumstances and can change or may differ in a jurisdiction other than the UK.
Continue reading for FREE
- Includes free newsletter updates, unsubscribe anytime. Privacy policy




