Jensen Huang resigned as chief executive, president and director of Nvidia on Wednesday 12 November 2026, effective immediately, ending a 33-year run in which he turned a graphics-card maker into the company that sells the picks and shovels of the artificial intelligence boom. Chief financial officer Colette Kress is interim chief executive while keeping her finance role; Harvey Jones, the former Synopsys chief, becomes the company’s first-ever board chair; and the board is looking inside and outside for a permanent successor. Quarterly results will still be announced on Monday 17 November.
For India, which has built the most ambitious public computing programme in its history on Nvidia hardware, the question is not whether the chips keep shipping next quarter. They will. The question is what it means to have bet a national AI mission on a single supplier at the moment that supplier loses the man who ran it, and whether the security allegations around his exit make Washington’s export rules tighter for everyone.
How a week ended a 33-year career
The sequence was fast. On Sunday 8 November, after the New York Times reported on his relationship with a businesswoman whom United States authorities had concluded was a Chinese intelligence asset, Huang wrote to employees admitting he had been unfaithful to his wife. The board announced an investigation into whether he had been targeted by a foreign intelligence effort, and said it had no reason to believe national security or company secrets were compromised. Shares fell more than 4 per cent on Monday.
A separate report, cited by Reuters, alleged the same woman had helped route at least $7.5 billion of Nvidia chips to Chinese companies through a network of businesses; Reuters said it could not verify that. By midweek a further media report alleged a second relationship, with a company executive. On Wednesday Huang resigned.
The stock fell as much as 14 per cent during Wednesday’s session before recovering to close down about 4 to 5 per cent, briefly taking Nvidia’s market value below $4 trillion for the first time since July 2025. Across the week the company lost roughly $500 billion in value. Bernstein halved its price target from $400 to $200.
Analysts split along a familiar line. Baird’s Tristan Gerra called it “a major page turned in the history of Nvidia”. StoneX’s Cody Acree saw no change to revenue or profit estimates through at least 2028. Goldman Sachs’s James Schneider said the company has strong functional management but no one who matches Huang’s vision and external influence, which could make the very large, multi-party investments of the AI era harder to put together. Raymond James’s Simon Leopold argued Huang was never a one-man band and the market overreacted.
India runs on Nvidia
The IndiaAI Mission, launched in March 2024 with an outlay of ₹10,371.92 crore over five years, exists to give startups, universities and researchers subsidised access to the computing power they could never buy themselves. By June 2026 it had put more than 45,000 GPUs on its shared platform; by August it had approved 237 projects and sanctioned 93.18 lakh GPU hours.
Almost all of that is Nvidia. The mission’s compute portal lists 44,077 units from 14 bidders. Of those, 35,985, or 81.6 per cent, are Nvidia. The rest is 5,548 Google TPUs, 1,200 AWS Trainium and Inferentia chips, 1,096 AMD accelerators and 248 Intel Gaudi units. The government itself has acknowledged that India’s AI computing ecosystem “currently relies on globally sourced GPUs”.
The private build-out is even more concentrated.
Who | What they have committed | Chip |
|---|---|---|
Yotta Data Services (Shakti Cloud) | More than 20,000 GPUs | Nvidia Blackwell Ultra |
AM Intelligence (Greenko promoters) | Orders for 29,000 GPUs | Nvidia Rubin |
L&T Vyoma | 10,000-GPU AI factory | Nvidia B300 |
E2E Networks | New cluster | Nvidia Blackwell |
Neysa, ESDS and others | Large orders, figures not disclosed | Nvidia |
Counterpoint Research puts Indian commitments to Nvidia-powered AI data centres at $4 billion to $5 billion. Nvidia’s Inception programme supports more than 4,000 Indian AI startups, and the company took direct stakes this year in two of them: Sarvam AI, in a $74 million round in August, and Simplismart, in a $14 million round in October. The partnerships Huang announced personally in Mumbai in September 2023, with Reliance and Tata, were followed within a year by deals with TCS, Infosys, Wipro, Tech Mahindra, Tata Communications and Zoho. Every state competing for data-centre investment, Tamil Nadu among them, prices its pitch on the assumption that Nvidia supply arrives on schedule.
What changes, and what doesn’t
Supply does not change this quarter. Orders placed for Blackwell and Rubin hardware are contracts with a company, not a man. Nvidia’s software stack, CUDA, and its developer relationships are institutional assets that an analyst at FOSS United estimates would take India half a decade or more to work around even if it wanted to. Kress has run the company’s finances through the entire AI boom; the board’s statement that it has “full confidence” in her is the message to customers.
The deal-making may. Goldman’s point is the one that matters for India. The multi-billion sovereign-AI arrangements of the past two years, where a government, a hyperscaler, a power company and Nvidia all sign at once, were brokered on Huang’s personal standing. A caretaker CEO closes fewer of those. Indian projects still at the memorandum stage should expect slower decisions from Santa Clara until a permanent chief is named.
Export rules are the real risk. The allegations that Nvidia chips reached Chinese companies through a diversion network will sharpen Washington’s attention to where every advanced GPU ends up. The 2025 AI Diffusion Rule placed India in a middle tier with import caps; its enforcement was suspended but, according to a May 2026 US government audit, never formally rescinded. Reports say the US is now weighing tighter licensing for shipments to seven countries, including Singapore, Malaysia and Thailand. India is not on that list. A scandal about leakage to China is the kind of event that lengthens such lists.
Diversification gets a hearing. Greyhound Research’s Sanchit Vir Gogoi argues India should diversify “by workload, and with evidence”, starting with inference, the part of AI computing that is least tied to Nvidia’s software. The India Semiconductor Mission, with its ₹76,000 crore outlay, and the indigenous SHAKTI and VEGA processors are years away from a GPU that competes; the realistic near-term alternatives are Google’s TPUs, AMD and AWS, which together make up 18 per cent of the mission’s compute today.
What to watch
Monday 17 November: Nvidia’s third-quarter results, against its own guidance of about $108 billion in revenue, and Kress’s first call as chief executive. The questions Indian buyers need answered are about sovereign-AI demand and whether the succession search has a timeline.
After that, the board’s investigation, the US inquiries into the diversion allegations, and any change to export licensing. The chips India has ordered will arrive. Whether the next tranche arrives on the same terms is now a political question as much as a commercial one.
FAQ
Why did Jensen Huang resign from Nvidia?
He resigned on 12 November 2026, four days after admitting an extramarital affair reported by the New York Times with a businesswoman US authorities had concluded was a Chinese intelligence asset. Nvidia’s board had opened an investigation; a further allegation of a second relationship followed. The company gave no reason in its filing.
Who is running Nvidia now?
Colette Kress, the chief financial officer, is interim chief executive and keeps the CFO role. Harvey Jones, former chief executive of Synopsys, is the board’s first chair. The board is searching for a permanent CEO among internal and external candidates.
How dependent is India on Nvidia chips?
Of the 44,077 compute units empanelled under the IndiaAI Mission, 35,985, or 81.6 per cent, are Nvidia GPUs. Private Indian data-centre operators have committed to more than 59,000 further Nvidia GPUs, with Counterpoint Research estimating $4 billion to $5 billion of investment.
Will India’s AI projects be delayed?
Existing orders are company contracts and are not affected by the change of chief executive. The risks are slower decisions on new large deals during the interim period, and any tightening of US export rules after the allegations about chips reaching China.
What are India’s alternatives to Nvidia?
Google TPUs, AMD and AWS chips already make up about 18 per cent of the IndiaAI Mission’s compute. The ₹76,000 crore India Semiconductor Mission and the SHAKTI and VEGA processors are the domestic route, but they are years from matching Nvidia’s hardware and software.
When are Nvidia’s results?
Monday 17 November 2026, as scheduled. The company has guided to about $108 billion in quarterly revenue.
