the Nvidia founder and CEO was still a centibillionaire when this week started. Yet the biggest single-day plunge in a stock ever recorded cost Huang $9.8 billion, according to estimates by Bloomberg and Forbes that track the fortunes of the ultra-rich.

Nvidia’s $279 billion drop in market cap on Tuesday eclipsed even Meta’s $251 billion from February 2022, driving heavy losses in the broader equity market which notched its worst day since August 5, a session so brutal it was compared with 1987’s Black Monday crash. 

Huang’s loss in wealth is not likely to end there.

Shares are expected to open lower on Wednesday following a Bloomberg report that broke after the market closed saying the U.S. Department of Justice has escalated an investigation into possible antitrust behavior by Nvidia.

When reached by Fortune for comment, Huang’s spokespeople did not deny the report citing anonymous sources that the company had been served a subpoena.

“Nvidia wins on merit, as reflected in our benchmark result and value to customers, who can choose,” it said in a statement sent on request late on Tuesday.

this marked the top for Nvidia, $NVDA pic.twitter.com/jKAIK5yvDi

— wallstreetbets (@wallstreetbets) August 28, 2024

Nvidia dominates lucrative AI training chip market

Huang is considered a rock star in the tech community. The Taiwanese immigrant started from humble beginnings before founding Nvidia in a Denny’s, where he once worked washing dishes. Initially, he set out to make superior processors for video game graphics by performing calculations in parallel rather than in sequence.

Since Nvidia’s graphics processing units (GPUs) crunched vast amounts of data more quickly and efficiently than rivals’ central processing units (CPUs), it became the favored choice of Bitcoin miners and AI companies training large language models like OpenAI’s GPT-4.

Huang’s company currently controls roughly 90% of the AI chip market. Any company serious about developing artificial general intelligence, like OpenAI or Elon Musk’s xAI, needs his chips to realize their ambitions. That’s why former Google CEO Eric Schmidt recommended last month that no investor take a pass on Nvidia, since it will vacuum up the lion’s share of hardware spending in the space.

Whether this domination is the result of illegal anticompetitive behaviour or the legitimate fruits of superior manufacturing is a judgement call Lina Khan’s Federal Trade Commission and the U.S. Department of Justice will now have to make.

“If they can’t show that the consumers are being harmed, then they really don’t have a case,” said Silicon Valley-based analyst Ray Wong of Constellation Research in an interview with Bloomberg Television. 

‘Clock is ticking’ on AI hype

Nvidia briefly became the world’s most valuable company this summer amid a frenzy in AI stocks. But sentiment has begun to wane following overextended valuations and disappointing fiscal third-quarter guidance and delays in its next generation Blackwell AI chip architecture.

Two Wall Street buy-side analysts fueled fears on Tuesday that near-term productivity gains may not justify the gold rush sentiment in the sector.

The global head of research at BlackRock, the world’s largest asset manager, warned recent research had cast some doubt on whether revenues from AI would catch up with capital investments. 

“We think it will take some time for big tech company revenues to reflect their AI capital spending. The AI buildout will take years—not quarters—to complete,” wrote BlackRock’s Jean Boivin.

Similarly, JPMorgan worried that MIT economist Daron Acemoglu may be right with his warning not the believe the AI hype. The latter predicted that productivity gains from AI could be a meager 0.06% per year, far from what’s needed to justify current spending.

“It’s probably too soon to worry that there’s no killer generative AI app yet,” Michael Cembalest, the chairman of market and investment strategy at the bank’s asset management arm, wrote on Tuesday. “But the clock is ticking.”