Qualcomm CEO Cristiano Amon delivered a pointed message to AI chip rivals as his company announced its entry into the data center market: the battle for dominance is just beginning, and “it’s going to become a competitive environment very soon.“Speaking at the Fortune Global Forum in Riyadh, Amon acknowledged that while “everybody’s playing to win” in the AI race, “it’s hard right now to declare who the winners are.” His comments came as Qualcomm unveiled its AI200 and AI250 accelerator chips, directly challenging Nvidia’s stronghold on the market for AI semiconductors.

The announcement sent Qualcomm’s stock soaring as much as 20% in Monday trading, signaling investor confidence in the company’s pivot from mobile devices to data center AI infrastructure.

New AI chips target Nvidia’s 90% market dominance

Qualcomm’s AI200, shipping in 2026, and AI250, arriving in 2027, will compete against Nvidia’s GPUs that currently command over 90% of the AI chip market. The chips focus on inference computing—running AI models—rather than training them, and boast 768 gigabytes of memory capacity, surpassing current Nvidia and AMD offerings.Saudi Arabia’s Humain will be the first customer, deploying 200 megawatts worth of AI200 chips at data centers next year. The deal expands a partnership announced during President Trump’s May visit to Riyadh.

Despite the competitive rhetoric, Amon compared today’s AI boom to the dotcom era, suggesting the technology’s long-term potential is “probably underestimated.” He noted that companies investing in efficient architectures will be “very well positioned” as the market matures.His measured optimism contrasts with growing bubble concerns from industry leaders including Ray Dalio and Sam Altman, who have warned of overexcitement reminiscent of the late 1990s tech crash. However, Amon maintained that AI represents a “massive opportunity” whose true scale won’t be clear for years.The AI chip market is expected to drive nearly $6.7 trillion in data center capital expenditures through 2030, according to McKinsey estimates.