Nvidia continues to dominate the artificial intelligence chip market, holding a near-monopoly on the powerful chips essential for AI model training and execution. However, a new wave of startups is challenging Nvidia’s supremacy, attracting significant investment. In 2026, AI chip startups secured $8.3 billion globally, with expectations of even greater investment in the coming year unless a market downturn occurs.
While Nvidia’s GPUs, originally designed for gaming, have been adapted for AI training, there is growing interest in optimizing AI inference, the practical application phase of AI technology. Startups argue that GPUs are not ideally suited for AI tasks, suggesting that new architectures could offer better energy and cost efficiency. Patrick Schneider-Sikorsky from the NATO Innovation Fund, which supports UK-based AI chip startup Fractile, noted that current GPU frameworks are not well-suited for large-scale AI inference.
Despite Nvidia’s vast resources and ongoing innovation, including the acquisition of Groq’s assets for $20 billion and investments in photonics technology, investors are keen on funding emerging AI chip technologies. In the U.S., significant funding rounds include $1 billion for Cerebras Systems and $500 million for companies like MatX and Ayar Labs. European startups such as Axelera and Olix raised over $200 million, with others like Euclyd and Optalysys planning to seek $100 million in funding.
Carlos Espinal from Seedcamp highlighted the growing importance of AI infrastructure, moving beyond niche endeavors. Meanwhile, Anthropic and OpenAI are expanding in the U.K., and TSMC reported a 58% profit increase, driven by AI chip demand. OpenAI has shifted its data center strategy to rely on Microsoft, and Amazon acquired Globalstar to compete with SpaceX’s satellite internet. Uber plans to increase its stake in Delivery Hero.
ASML’s stock declined despite strong earnings, attributed to high expectations and export restrictions affecting sales to China.
