NVIDIA CEO Jensen Huang appears far more optimistic about the scale of investment flowing into artificial intelligence than even the most bullish forecasts.

During the company’s earnings call, Huang said he believes capital expenditures related to AI could eventually reach as much as $4 trillion.

“Capital expenditures are currently at $1 trillion and growing toward the $3 trillion to $4 trillion range,” he said, referring specifically to investments made by hyperscalers such as Alphabet and Amazon, while excluding other parts of the supercomputing market, including neocloud providers.

NVIDIA CFO Colette Kress was even more specific.

“As analysts now forecast hyperscaler capital expenditures surpassing $1 trillion by 2027, and as agentic AI begins spreading across industries, investments in AI infrastructure are on track to reach $3 trillion to $4 trillion annually by the end of this decade,” she said.

There is, however, one major issue: these projections are far ahead of Wall Street expectations.

Analysis from Needham & Company analyst Laura Martin suggests that market consensus currently places hyperscaler capital expenditures at $1.03 trillion by 2028, roughly one-third to one-quarter of what Huang’s forecast would imply just two years later.

“If Jensen Huang’s prediction is correct, then market estimates will need to be revised upward,” she wrote. “His vision for hyperscalers differs from what the companies themselves are saying in their earnings calls, and it is more interesting.”

Some analysts on Wall Street expect spending to hit $1 trillion by the end of next year, faster than broader consensus projections, but still far behind Huang’s scenario, which would imply spending quadrupling within the next three years.

There is little doubt that increased investment from hyperscalers and others would benefit NVIDIA, which remains the dominant supplier of AI chips. The growing profitability of cloud services and continued advances in frontier AI models are currently reinforcing Huang’s optimism.

Quarterly revenue growth exceeded expectations across the major cloud providers. Alphabet reported growth of 63%, Amazon Web Services grew by 28%, and Microsoft posted 40% growth.

“The world has one billion human users. I believe there will be billions of agents, and every one of them will create sub-agents,” Huang said.

Even so, doubts remain over the long-term impact of AI on productivity and profitability.

JPMorgan Chase estimated in November that for AI investments to generate a 10% return by 2030, the technology would need to produce $650 billion in recurring annual returns, a figure the bank described as “staggeringly large.”

For comparison, global cloud revenues over the last 12 months reached $455 billion.

At the same time, large-scale productivity gains from AI have yet to fully materialize, and economists still lack consensus on the technology’s long-term economic impact.

Source: moneyreview.gr