Meta draws buyers as 5 AI leaders call demand nearly infinite

AI demand is nearly unlimited, and energy is the only real constraint — a reality that makes Meta Platforms a compelling buy, according to a growing chorus of industry leaders.

Five AI industry leaders told CNBC that demand for computing power far exceeds available supply, directly rebutting peak-chip fears and reinforcing the investment case for companies like Meta Platforms Inc. building AI infrastructure at scale.

"AI demand is nearly unlimited, and energy is the only real constraint," Pat Gelsinger, former Intel Corp. chief executive and now CEO of venture capital firm Playground Global, said in a CNBC interview. "The economic value available through AI enhancement is nearly infinite across every industry you can imagine."

Lumentum Holdings Inc., a maker of optical and photonics components for data centers, has already contracted out all the products it expects to produce over the next five years, Chief Executive Officer Michael Hurlston said. The company's shares have risen roughly sevenfold over the past year. Mark Boroditsky, chief revenue officer of neocloud provider Nebius, said there is "far more demand than we can handle — and that's been the case for some time now."

The comments directly challenge the peak-chip thesis — the idea that computing resources are heading toward oversupply — which has weighed on AI-related stocks. For Meta, which has committed tens of billions of dollars to AI infrastructure, the demand outlook suggests its capital spending is not overbuilding but positioning for a structural shift that could last years.

Industry Leaders Dismiss Oversupply Concerns

Andrew Feldman, chief executive of AI chip maker Cerebras Systems Inc., called Meta's recent suggestion that it might lease out surplus AI computing capacity an "unusual case." He said that across the industry, demand for compute far exceeds available capacity and that data centers are in short supply.

Park Sung-yoon, chief executive of South Korean semiconductor startup Rebellions Inc. — which counts Samsung Electronics Co. and SK Hynix Inc. among its investors — said the momentum for building AI infrastructure is still enormous. On Meta and xAI's recent moves to lease computing power, he said he does not view those decisions as an indication that hyperscalers are overinvesting.

Rationalization Phase Begins

Even those emphasizing enormous demand acknowledged a note of caution has crept into AI investment decisions. Boroditsky said chief financial officers are now starting to scrutinize AI budgets more carefully, but stressed this does not mean they are abandoning AI investment altogether. He described the shift as "value-maxing" — optimizing return on investment — and said the earlier era of "token-maxing" only makes sense when companies are getting a return.

"We're seeing a transition toward more rational spending — that's something we've always seen in past cycles," Boroditsky said. "This rationalization will ultimately make AI demand sustainable."

Feldman said that in the future, advanced AI models will handle more complex tasks while other models take care of the rest — "just like you don't need a big car to go to the corner store."

For investors, the message from industry leaders is clear: AI infrastructure spending is not peaking but entering a new phase of rational growth. Meta, which trades at roughly 23 times forward earnings, has committed more than $35 billion in annual capital expenditures, much of it directed at AI data centers and computing infrastructure. If the demand thesis holds, those investments could generate returns for years as enterprises and consumers increasingly rely on AI-powered services. The next catalyst for Meta will be its third-quarter earnings report, where investors will watch for updated capital expenditure guidance and commentary on AI-driven revenue growth.

This article is for informational purposes only and does not constitute investment advice.