At Computex Taipei, one of the world’s largest and most influential technology trade shows, held annually in Taipei (Taiwan), Marvell Technology CEO Matt Murphy said that as AI models evolve into increasingly complex AI agents, the next major challenge facing data centers may no…
nger be computing power or memory capacity, but the ability to move vast amounts of data efficiently between chips and servers. The highlight of the event came when Nvidia CEO Jensen Huang made a surprise appearance and said “ladies and gentlemen, the next trillion-dollar company” as he introduced Marvell, drawing loud applause from attendees and becoming one of the most talked-about moments of the conference
Marvell Technology stock ($MRVL) jumped by more than 20% in pre-market on Huang’s endorsement. The appearance highlighted the increasingly strategic relationship between Nvidia and Marvell as hyperscale cloud providers race to build larger and more powerful AI clusters. While Nvidia dominates AI computing through its GPUs, Marvell has become a critical supplier of custom AI chips, optical networking solutions, and high-speed interconnect technologies that enable those systems to function at scale.
Huang and Murphy argue that connectivity is becoming the defining factor for overall system performance. Modern AI models require tens of thousands of GPUs to work together across massive data center networks. As clusters grow larger, the speed and efficiency with which information moves between processors becomes increasingly important.