Jensen Huang on AI: 'This Time Is Different' — Why the Chip Boom Has Years to Run

Constraints preventing overbuilding

Huang cited real physical constraints — chip fabrication capacity, energy supply for data centers, and the sheer engineering complexity of building AI factories — as natural governors against runaway overinvestment. The chip industry must roughly double every few years just to keep up with demand, he argued .

Timeline for continued growth

Huang said "we are in the beginning of this build-out" and projected that the growth cycle will span "many years," with a bubble unlikely before 2030 . In another interview, he described the current moment as "the beginning of a ten-year build out of a new computing platform" .

Projected AI Infrastructure Investment Scale

Huang's estimates have escalated rapidly over the past year:

• $600 billion annually: In mid-2025, Huang said the top four hyperscalers (Amazon, Google, Microsoft, Meta) were spending around $600 billion per year in combined AI capital expenditures, roughly double the prior year .
• $3–$4 trillion by end of decade: He repeatedly forecast that global AI infrastructure spending — data centers, real estate, power, and hardware — will total between $3 trillion and $4 trillion by 2030 .
• $4 trillion per year potential: By Nvidia's May 2026 earnings call, Huang raised the ceiling further, saying annual AI infrastructure spending could soar to $4 trillion — four times the Wall Street consensus at the time . CNBC reported that Huang had "significantly outpaced expectations regarding the forthcoming investments in artificial intelligence" .

By late July 2026, The Motley Fool reported that Nvidia had approximately $1 trillion in confirmed demand through 2027 from customers including hyperscalers and sovereign AI projects .

Hyperscaler Borrowing: What Huang Didn't Address

The available sources do not contain explicit quotes from Huang addressing investor skepticism over hyperscaler borrowing and negative free cash flow. Multiple articles note that hyperscalers are doubling their capital expenditures — Meta allocated $72 billion in 2025 and planned up to $135 billion in 2026, while Google announced plans to invest as much as $185 billion — but Huang's specific rebuttal on debt/bond financing versus free cash flow is not captured in the evidence gathered.

Physical AI Partnerships with Japanese Industrial Leaders

During a two-day visit to Tokyo on July 15–16, 2026, Huang announced a major expansion of Nvidia's "physical AI" collaborations with Japan's industrial giants . The partnerships include:

• Toyota — collaboration to develop next-generation autonomous driving and robotics platforms
• Fujitsu — joint development of physical AI for manufacturing, with CEO Takahito Tokita co-hosting a roundtable
• Fanuc — industrial robotics integration with Nvidia's AI platforms
• Kawasaki Heavy Industries — physical AI for factory automation
• Kioxia — memory/storage integration for AI infrastructure
• Yaskawa Electric — also a participant in the consortium
• SoftBank and Sony — additional technology partners

Huang said Japan's traditional craftsmanship is "essential" to physical AI and framed "Made in Japan" AI robots as a solution to the country's biggest problem: a shrinking workforce . He stated, "The next frontier of AI is in the physical world, and this is a once-in-a-generation opportunity for Japan" .

The Vera Rubin AI Factory Planned for Japan

On July 16, 2026, the Japanese government, industrial leaders, and Nvidia jointly announced the world's first national AI infrastructure project — a Vera Rubin AI factory built in partnership with Noetra Corp. .

Key specifications confirmed by Nvidia's official newsroom :

• 13,750 Vera CPUs and 27,500 Rubin GPUs
• 140 megawatts of data center capacity
• Built on the Nvidia DSX platform
• $6.2 billion total project value, government-backed
• Japan became the first nation to designate a physical AI compute cluster as critical national infrastructure

Construction is expected to begin in April 2027, with operations planned for 2028 .

Huang declared that "Vera Rubin is already in production. Giant amounts of production incoming" during a Tokyo developer event, flatly denying manufacturing delay reports .

Japan's National Robotics Plan

Multiple sources describe that Japan's broader national robotics strategy includes a plan to deploy 10 million AI-equipped robots by 2040 to compensate for its severe labor shortage . The Observer noted that "the government and major industrial players are gearing up to deploy AI in robots and factories" as a national priority .

Note: The specific branding of a "Noetra consortium" for the 10-million-robot-by-2040 plan could not be independently verified from the sources gathered. The Noetra Corp. entity is confirmed as the AI factory operator; the broader national robotics plan of 10 million robots by 2040 is reported in multiple sources.