AI chipmaker Cerebras Systems Inc. (Cerebras (CBRS)) made its Nasdaq debut Thursday, and CEO Andrew Feldman used the moment to talk about the long game — both for his company and for the U.S. semiconductor industry. The IPO priced at $185 per share, and by the end of the day, shares were trading at $352.01, up more than 90%. Investors are clearly hungry for AI infrastructure plays.

Feldman told CNBC that going public was a natural step for a company scaling rapidly in the AI computing space. He described the IPO as "the right way to fund our growth," adding that Cerebras sees AI transforming everything from enterprise data processing to broader business operations. In his view, the AI industry is still in the early innings of a much bigger transformation.

Cerebras is positioning itself as a direct challenger to Nvidia (NVDA), the current market leader in AI chips. According to a Bloomberg report, Feldman claims the company's large, data-crushing chips allow its specialized computers to run AI models faster than Nvidia's hardware. The secret sauce? Cerebras builds massive AI computing systems powered by custom silicon designed to process huge models more efficiently. Feldman described the company's processor as one of the largest chips ever developed in the semiconductor industry, allowing it to process more information in less time while using less power.

But building those chips requires a global supply chain. Cerebras has relied on Taiwan Semiconductor Manufacturing Co. (TSM) since its founding for advanced production. Feldman noted that the relationship is critical, and it highlights a broader challenge: the U.S. is far from self-sufficient in advanced chip manufacturing.

Feldman described efforts to expand advanced semiconductor manufacturing in the United States as a long-term geopolitical and infrastructure challenge that could take 10 to 15 years to develop at scale. He said advanced chip manufacturing depends on deeply integrated supply chains, specialized infrastructure, and massive capital investment that cannot be replicated quickly. "Many companies have attempted to replicate large-scale computing architectures and semiconductor ecosystems over the years, but struggled to match established industry leaders," he argued.

Looking further ahead, Feldman floated the idea of space-based AI computing, suggesting that large-chip systems might offer advantages in orbital environments. But he described that transition as another decade-long project. For now, Cerebras is focused on earthbound growth — and the public markets are giving it the fuel to do so.