Nvidia CEO Jensen Huang said Marvell Technology is "the next trillion-dollar company" during a joint appearance in Taipei on Tuesday [1].
The endorsement signals a growing confidence in the AI-driven semiconductor sector, suggesting that the financial windfall currently benefiting Nvidia may expand to other key infrastructure providers.
Huang said the remarks during Computex 2026, where he appeared alongside Marvell CEO Matt Murphy [1, 2]. The two executives discussed the future of AI hardware and the critical role of networking and data processing in scaling artificial intelligence capabilities [1].
Market reaction to the comment was immediate. Marvell shares rose between 22% [3] and 25% [4] in pre-market trading following the announcement. Other reports placed the surge at 23% [2].
Huang said Marvell’s growth prospects were strong, specifically highlighting the company's role in the semiconductor ecosystem that supports generative AI [1, 3]. The valuation prediction suggests that Marvell could eventually reach a market capitalization of $1 trillion, a milestone currently reserved for a small group of the world's largest technology firms [1, 3].
While some early reports associated the remarks with the GPU Technology Conference, verified sources confirm the statements were made at Computex 2026 in Taiwan [1, 2].
Marvell has focused heavily on custom silicon and optical connectivity, both of which are essential for the high-speed data movement required by Nvidia's GPUs [1]. The partnership between the two companies underscores the interdependence of chip design, and networking hardware in the race to build more powerful AI clusters.
“"the next trillion-dollar company"”
The endorsement from Jensen Huang acts as a powerful market signal, effectively validating Marvell's strategic pivot toward AI infrastructure. By identifying Marvell as a future trillion-dollar entity, Huang is highlighting that the AI boom requires not just the processing power of GPUs, but the sophisticated networking and connectivity hardware that Marvell provides to prevent data bottlenecks.
