Anthropic CEO DarioAmodei says even a slight miscalculation in AI investment timing could bankrupt the company, as rivals pour hundreds of billions into data centers.
"I really do believe that we could have models that are a country of geniuses in the data center in one to two years," Amodei said.
He added, "One question is: How many years after that do the trillions in revenue start rolling in? I don't think it's guaranteed that it's going to be immediate. It could be one year. It could be two years. I could even stretch it to five years, although I'm skeptical of that."
Amodei warned that committing massive capital too early could be "ruinous" if revenue projections are even slightly off.
"If my revenue is not $1 trillion, if it's even $800 billion, there's no force on Earth, there's no hedge on Earth that could stop me from going bankrupt if I buy that much compute," he said.
Anthropic plans to spend $50 billion on U.S. AI infrastructure, starting with data centers in Texas and New York, while hyperscalers are budgeting far more: Amazon $200 billion, Alphabet up to $185 billion, and Meta $135 billion.
Futurum Group CEO Daniel Newman said on Fox Business that Nvidia benefited immediately from AI infrastructure demand, even if hyperscalers' returns took longer to materialize.
Deepwater Asset Management'sGene Munster said investors were underestimating how close AI was to real-world utility, despite recent declines in Meta, Amazon and Alphabet shares.
Semiconductor stocks, including Broadcom Inc.(NASDAQ:AVGO) and Micron Technology Inc. (NASDAQ:MU), rose after Alphabet and Amazon increased 2026 capex forecasts.
The Semiconductor Industry Association said global chip sales reached $791.7 billion in 2025 and were projected to grow 26% in 2026, driven by AI demand.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Join thousands of traders who make more informed decisions with our premium features.
Real-time quotes, advanced visualizations, backtesting, and much more.