Nvidia vs. Cerebras: Which Chip Maker Offers Better Value Now?

Nvidia and Cerebras Systems both design chips essential to powering artificial intelligence workloads, but they operate from very different positions in the market. Nvidia, in business for more than 30 years, has become the dominant force in AI chips since the boom began, setting the tone for the entire market through its quarterly announcements. The company designs graphics processing units (GPUs), which are the key chips needed for training and inference of AI models. Nvidia has maintained its lead through a focus on innovation—it updates its GPUs annually—and is preparing to ship its Vera Rubin platform later this year. The platform will introduce a stand-alone central processing unit (CPU), opening up what the company estimates to be a $200 billion（約32兆円） market. In its latest earnings report, Nvidia predicted $20 billion（約3.2兆円） in stand-alone CPU sales this year and expressed confidence it would dominate this new category. The company's financial strength is evident: it has delivered quarter after quarter of double- or triple-digit earnings gains, with earnings reaching record levels during the AI boom. This sustained performance reflects ongoing demand and the reality that AI remains in its early stages of real-world deployment.

Cerebras Systems, founded in 2015 and a far less familiar name, is an emerging challenger with potentially disruptive technology. The company has designed a wafer-scale engine (WSE)—a giant chip 58 times larger than Nvidia's B200—that it claims delivers speeds faster than today's GPUs. This size advantage, according to Cerebras, provides massive compute and memory bandwidth, resulting in tremendous speed gains. Specifically, Cerebras reports that in inference—the process an AI system uses to generate answers to a query—its WSE delivers answers 15 times faster than today's top-selling GPUs. This performance advantage has begun to drive real business traction: in the first quarter, Cerebras' revenue soared 92% to $193 million（約310億円）. The company recently signed key deals with OpenAI for compute services and with Amazon's cloud division to make its WSE systems more broadly available. These partnerships represent a potential inflection point for Cerebras, as more customers gain exposure to the technology and evaluate it for their own needs.

Cerebras went public in May 2025, raising $5.5 billion（約8800億円） in what was the biggest IPO of 2025 at the time—before Space Exploration Technologies launched its operation in June and claimed the title of largest IPO ever. Since its first day of trading, Cerebras stock has declined 30%, creating a buying opportunity for investors willing to accept higher risk. However, the company is not yet profitable, a factor that reflects both its early growth stage and the additional risk that an investment in Cerebras carries compared to Nvidia.

The choice between these two companies depends on investor risk tolerance and time horizon. Nvidia is trading at 23x forward earnings estimates, a valuation the article characterizes as attractive given the company's proven strengths, sustained growth, and dominance of the market. For most investors, Nvidia is positioned as the better discount buy. Cerebras, by contrast, appeals to more aggressive investors willing to bet on a younger, unprofitable but rapidly growing company with potentially transformative technology. The article notes that analysts predict the AI market will exceed $3 trillion（約480兆円） in the early part of the next decade, suggesting there is room for multiple successful chip makers rather than a winner-take-all outcome. This context implies that Cerebras need not displace Nvidia to deliver strong shareholder returns, but its path to profitability and sustained growth remains unproven.

The AI chip market is entering a pivotal phase in which two very different players are gaining investor attention. Nvidia, which has dominated for over 30 years and been first to market in AI chips, is now moving beyond GPUs into CPUs—a $200 billion（約32兆円） market the company explicitly aims to dominate. The company's track record of annual GPU updates, sustained double- or triple-digit earnings growth, and record profitability suggest it has the execution capability to succeed. However, Cerebras represents a fundamentally different bet: a younger, unprofitable but fast-growing alternative leveraging a radically different chip architecture. The wafer-scale engine's size advantage (58 times larger than Nvidia's B200) translates to concrete performance gains—15x faster inference—that have already attracted major customers including OpenAI and Amazon. The body notes that analysts predict the AI chip market will reach beyond $3 trillion（約480兆円） in the early part of the next decade, which the article frames as creating room for multiple winners rather than a zero-sum competition. Both companies' recent share declines from recent highs thus present distinct investment profiles: Nvidia offers stability and proven execution at 23x forward earnings; Cerebras offers growth and technological differentiation but at higher risk given its pre-profitability stage and 30% slide from its IPO opening price.