Cerebras Systems, the US-based designer of wafer-scale AI processors, saw its stock price jump more than 12% in early trading today after the company confirmed a new partnership with a major cloud computing platform. The deal will see Cerebras’ CS-2 and upcoming CS-3 systems integrated into the cloud provider's infrastructure, giving enterprise customers direct access to the firm's specialised hardware for training large language models.
The rally comes amid a broader uptick in AI-related equities, with the Nasdaq Composite climbing 0.8% to 16,742 points. Cerebras, which listed on the Nasdaq in March 2024, has positioned itself as a direct rival to Nvidia in the high-end AI training market. Today’s move adds to a year-to-date gain of roughly 45%, though the stock remains volatile given the competitive landscape.
“Cerebras is capitalising on the insatiable demand for compute power needed to train next-generation AI models,” said Dr. Helena Grant, senior analyst at London-based tech research firm Pixel Analytics. “Their wafer-scale approach offers memory bandwidth advantages that large cloud customers find attractive, especially as Nvidia’s supply constraints persist.”
For UK investors, the rally underscores the growing importance of specialist semiconductor companies in global portfolios. While Cerebras is not listed in London, many British pension funds and wealth managers hold US tech stocks through passive index trackers. The FTSE 100 was little changed today, rising 0.1% to 7,684 points, but the broader technology sector remains a key driver of returns for UK savers with exposure to global equities.
Industry watchers caution that the AI chip market remains fiercely competitive, with Nvidia, AMD, and several well-funded startups all vying for market share. Cerebras’ ability to convert partnership announcements into sustained revenue growth will be closely watched by analysts in the coming quarters. The company is due to report its next quarterly earnings in early August.
Source: Market data from Refinitiv; commentary from Pixel Analytics.