Cerebras Systems is emerging as a serious challenger to Nvidia's grip on artificial intelligence infrastructure, with major Wall Street firms initiating bullish coverage and projecting the company could capture 40% to 50% of the fast-inference market. The inference chip maker, which went public last month, has already secured partnerships with Amazon and OpenAI, positioning itself as a first-mover in a market analysts estimate could reach $130 billion.

Why Is Cerebras Drawing Such Strong Wall Street Interest?

Two major investment banks initiated coverage of Cerebras on the same day in early June, signaling growing confidence in the company's technology and market opportunity. Citi began coverage with a "buy" rating and a $340 price target, implying 70% upside from the stock's closing price of $201 on Friday. Morgan Stanley launched coverage with an "Overweight" rating and a $250 price target, with analyst Joseph Moore noting that demand for low-latency AI inference solutions is rising.

The enthusiasm reflects a fundamental shift in how artificial intelligence infrastructure is being built. While Nvidia has dominated the market for training large language models, inference represents a different challenge. Inference is the process of running a trained AI model to generate responses, and it requires different optimization priorities than training. Cerebras has designed wafer-scale chips specifically for this workload, offering what analysts believe is a superior approach to traditional graphics processing units (GPUs).

"Demand for low-latency AI inference solutions is rising, and Cerebras is well-positioned with its wafer-scale technology," noted Morgan Stanley analyst Joseph Moore.

Joseph Moore, Analyst at Morgan Stanley

Morgan Stanley's analysis suggests that Cerebras's existing contracts could support roughly $6 billion in revenue by 2028, with potential for even bigger growth beyond that projection. This revenue trajectory would represent extraordinary expansion from the company's current scale, underscoring analyst confidence in the inference market's explosive growth.

How Does Cerebras's Technology Differ From Nvidia's Approach?

The core distinction lies in chip architecture and optimization strategy. Cerebras develops wafer-scale AI chips designed to train and run AI models faster than traditional GPUs. A wafer-scale approach means the company manufactures larger, more integrated chips that can process information more efficiently for specific tasks like inference. In contrast, Nvidia's strategy focuses on discrete accelerators that can be combined in various configurations.

This architectural difference matters because inference workloads have different requirements than training. Inference often prioritizes speed and latency, meaning the time it takes to generate a response. Cerebras's wafer-scale design appears optimized for these constraints, which is why the company has attracted major cloud providers seeking alternatives to Nvidia's dominant position.

Cerebras CEO Andrew Feldman emphasized the company's independence from Nvidia during remarks at the Bloomberg Tech conference in San Francisco. He noted that Cerebras is working with multiple data center suppliers, though notably not with Nvidia. This positioning suggests the company is building a distinct ecosystem rather than competing within Nvidia's existing infrastructure partnerships.

Steps to Understanding the Inference Chip Market Opportunity

• Market Size Context: The fast-inference market is estimated at $130 billion, representing a distinct and growing segment separate from the training chip market where Nvidia currently dominates.
• First-Mover Advantage: Citi analysts believe Cerebras has a first-mover advantage in the fast-inference segment, positioning the company to capture 40% to 50% of total market share if execution remains strong.
• Customer Validation: Cerebras has already secured agreements with Amazon and OpenAI, two of the world's largest AI infrastructure operators, providing proof of concept for the technology's viability.
• Revenue Trajectory: Morgan Stanley projects existing contracts could support approximately $6 billion in revenue by 2028, demonstrating a clear path to substantial scale.

The stock's performance since its public debut tells part of the story. Cerebras shares have declined approximately 35% from their initial public offering price, creating what analysts view as an attractive entry point. The stock was trading over 5% higher in pre-market trading on Monday following the brokerage coverage announcements, reflecting renewed investor interest.

However, retail investors remain divided on the opportunity. While some traders on social media platforms noted the stock's valuation appeared "cheap," others pointed to the company's price-to-earnings ratio of 181.91, which significantly exceeds Nvidia's ratio of 31.41. This valuation gap reflects the market's uncertainty about whether Cerebras can execute on its ambitious growth projections.

The broader context matters here. Nvidia's dominance in AI infrastructure has created a concentration risk that both hyperscalers and investors are beginning to address. Amazon, Google, and other major cloud providers are developing custom chips to reduce dependence on Nvidia supply lines. Cerebras enters this competitive landscape at a moment when customers are actively seeking alternatives and when the inference market is expanding rapidly.

The inference chip market represents one of the few areas where Nvidia's traditional advantages may be less decisive. While the company's CUDA software ecosystem has created powerful lock-in for training workloads, inference optimization is a more specialized domain where purpose-built chips can offer genuine advantages. Cerebras's partnerships with Amazon and OpenAI suggest that major infrastructure operators believe the company's technology delivers meaningful benefits for their specific use cases.

As the AI infrastructure market continues to evolve, the competition between generalist platforms like Nvidia and specialized players like Cerebras will likely intensify. Wall Street's bullish stance on Cerebras reflects confidence that the inference market is large enough to support multiple winners, even as Nvidia maintains its overall dominance in AI chips.