Arm CEO says AI chip shortages are delaying medical progress

Global shortages of semiconductor chips are hindering the rapid deployment of artificial intelligence infrastructure and data centres, according to Rene Haas, chief executive of microchip design firm Arm Holdings. Haas warned that these physical supply bottlenecks are delaying technological advancements that could lead to medical breakthroughs, including finding new treatments for cancer.

Speaking on a BBC podcast, Haas argued that computer models will eventually process cellular data and biological markers at scales far beyond human capacity. While acknowledging that mapping how cancer interacts with DNA markers remains difficult for current systems, he expressed confidence that advancing computing power will ultimately solve the issue.

AI’s Potential in Medical Research

Haas, who previously served on the board of pharmaceutical company AstraZeneca until April, predicted that artificial intelligence will help identify cancer cures within our lifetimes. He noted that as data models grow more sophisticated and hardware improves, computers will solve biological problems that are currently impossible for human researchers to unravel.

However, cancer researchers have offered a different perspective on the hardware requirements for medical advances. Professor Chris Bakal, a researcher at the Institute of Cancer Research in London and chief executive of Sentinal4D, stated that the future of medical AI does not depend solely on massive computing power or giant data centres. Bakal emphasized that training AI on specialized data gathered directly from patient samples in laboratories is what drives progress. He noted that this approach could save years in drug development, arguing that the sector will “belong to whoever has the right measurements” rather than whoever operates the largest supercomputer.

Chip Shortages and Manufacturing Realities

The primary barrier to expanding AI capabilities remains the supply of physical chips required to construct data centres. Haas pointed out that ambitious projects, including multi-gigawatt facilities in France and the United States as well as proposed orbital data centres in space, are constrained by manufacturing capacity. He stressed that significant additions to semiconductor manufacturing facilities, known as fabs, are necessary before such projects can become reality.

Despite supply discussions between the British government and the technology industry, Haas expressed skepticism about establishing domestic chip manufacturing in the United Kingdom. He noted that chip fabrication requires immense capital expenditure, specialized technical labor, and vast natural resources, pointing out that global manufacturing remains heavily concentrated in existing hubs such as Taiwan’s TSMC.

Market Expansion and Business Growth

Arm Holdings, which designs the central processing unit architectures used in billions of smartphones and consumer electronics, has seen rapid commercial expansion during the AI sector boom. The company’s technology is currently utilized in roughly half of all AI data centres globally. Following a request from Meta, the firm developed the Arm AGI chip, which has generated more than $2 billion in demand since launching in March.

Beyond medical applications, Haas predicted that AI will drive widespread deployment of humanoid robots within five years. Within a decade, he expects self-learning machinery to handle complex physical tasks in manufacturing, construction, cleaning, and security. Addressing concerns regarding automated labor, Haas described predictions of widespread job displacement as overstated, arguing that emerging opportunities will balance workforce changes. He also dismissed fears of a valuation crash among AI companies, citing sustained long-term demand for computing infrastructure.

Background

Arm Holdings is the largest technology company headquartered in the United Kingdom, based in Cambridge, where it remains the area’s largest employer. Half of its workforce is located within the UK, though the firm was acquired by Japan-based SoftBank Group in 2016 and subsequently listed its shares on New York’s Nasdaq exchange rather than the London Stock Exchange.

The company primarily licenses microchip designs rather than manufacturing physical silicon itself. Its business model relies heavily on third-party foundries, predominantly Taiwan Semiconductor Manufacturing Company (TSMC), to produce physical hardware for tech clients across consumer and enterprise markets.