Arm CEO Highlights Accelerating AI Demand Despite 4% Stock Dip After Earnings Report

This article does not constitute investment advice. The author holds no shares in any of the stocks discussed.

In a recent interview on CNBC, Arm CEO Rene Haas emphasized that the company continues to experience strong demand for its AI-related products. Despite a significant drop of over 4% in Arm’s shares at the market open today—triggered by investor concerns regarding a potential decline in AI demand—Haas expressed confidence in the future of AI computing.

AI Demand Remains Robust, According to Arm CEO

During the discussion, Haas reassured investors by stating that contrary to fears of diminishing AI demand, Arm is actually witnessing an acceleration in this sector. As the market remains cautious following last month’s significant selloff that erased a trillion dollars from AI stock valuations, Haas’s insights offered a glimmer of hope for investors.

Haas pointed to two critical trends taking place in data centers: the continued adoption of Arm architecture and the growing prominence of general-purpose computing. He highlighted Amazon’s Graviton data center chips, noting that more than half of new installations are now Arm-based, with 90% of Amazon’s highest-volume customers utilizing Arm products.

Amazon's Graviton3 Chip — An Amazon slide describing the Graviton3 chip. Image: Amazon

Focusing on AI operations, Haas explained that these primarily revolve around training and inferencing large language models. He detailed the complementary nature of the products, describing NVIDIA’s Blackwell chips as the “flagship product”where Arm CPUs work harmoniously with NVIDIA GPUs. This symbiosis, according to Haas, positions Arm favorably in the evolving landscape of AI and general-purpose computing.

When questioned about any potential decline in orders for NVIDIA’s products, Haas stated, “I don’t really see a slowdown.”He further elaborated, indicating that the escalating computational needs of large language models require substantial power and resources, which implies a steady or increased level of capital spending in the industry.

Addressing concerns stemming from DeepSeek’s recent demonstration of reduced costs, Haas clarified that this does not diminish the relevance of existing AI models. In fact, he noted, “People need to remember what’s happening is they’re distilling, which basically means building on the frontier models that have been created.” This process emphasizes that the creation of advanced models remains essential while paving the way for rapid enhancements in AI technology.

Lauding the Stargate AI project as an “amazing opportunity, ”Haas posited that this initiative would catalyze significant technological advancements. He expressed optimism that AI would increasingly permeate the consumer market, suggesting that improved efficiency in AI models following DeepSeek would directly benefit Arm by enhancing the performance of power-constrained devices on the endpoint.“More efficient models mean you can run more inference at endpoints, ”he explained, correlating this to heightened demand for Arm’s products.

Source & Images