NVIDIA CEO Jensen Huang told Yahoo Finance on October 21, 2026, that a gigawatt-scale AI data center represents roughly $50 billion in investment. Speaking with Deirdre Bosa at GTC Berlin, Huang connected that enormous capital requirement to deployment delays, GPU resale value and how lenders might finance the infrastructure.

Key Takeaways

• Huang’s $50 billion per gigawatt figure is an executive estimate, not a universal construction quotation.
• He suggested NVIDIA computing could remain economically productive for eight years or longer.
• Faster deployment could bring forward revenue, but power availability and permits remain obstacles.

An AI Factory That Cannot Earn Money Yet

In Yahoo Finance’s published interview and transcript, Huang said conventional systems can take four to six months to become operational after computers arrive at a data center. During that time, operators tie up money without the intended computing revenue.

Huang argued that factory assembly could shorten this waiting period. NVIDIA’s official DSX POD announcement targets deployment in as little as 14 days at an already prepared site. That excludes the time required to secure land, utility power, permits and other site infrastructure.

Can a GPU Really Last Eight Years?

Huang challenged the four-year productive-life assumption often applied to AI hardware, saying its usefulness could extend eight years or more. When asked about Blackwell’s seven-to-eight-year lifetime, he responded that it was probable. This is a forecast from the supplier, not proof that every accelerator will retain commercial value for that long.

A longer useful life matters because lenders estimate how much collateral will be worth if a borrower defaults. In a separate Yahoo Finance report, Huang made the case for GPU-backed borrowing. Actual loan pricing still depends on utilization, contracts, market prices and technological change.

Why Power and Financing Matter Together

BitcoinVersus.Tech previously covered NVIDIA’s modular DSX POD, new GPU-backed financial products and Bitcoin miners’ power-infrastructure advantage. This interview connects those subjects: a $50 billion estimate is meaningful only when capacity can be energized, operated and sold.

BitcoinVersus.Tech Editor’s Note: We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support our independent research, please donate 5,000 sats using the QR code provided.

For best QR-code scanning results, scan this donation code from a desktop or larger screen. Mobile screens may prevent camera recognition.