Nvidia CEO Jensen Huang Resigns After Board Opens Probe Into Affairs and an Alleged Chinese Spy

Nvidia CEO Jensen Huang Resigns After Board Opens Probe Into Affairs and an Alleged Chinese Spy

Jensen Huang resigned as CEO of Nvidia on the morning of November 12, 2026, ending a 33-year tenure at the company he co-founded in 1993, after the board opened an investigation into multiple extramarital affairs with a direct report and his ties to a person alleged to be a Chinese spy. The roughly $4.5 trillion chipmaker loses its only public face just as its Rubin platform enters volume ramp, forcing customers and suppliers to re-confirm commitments. Liu Gong's read: the damage comes less

The sequence closed within 48 hours. On the morning of November 12, Nvidia's board informed employees by company-wide email that co-founder, president and CEO Jensen Huang had tendered his resignation, effective immediately. The trigger was an independent investigation opened weeks earlier along two lines: multiple extramarital relationships between Huang and a subordinate who reported directly to him, in breach of the company's code of conduct requiring disclosure of manager-report relationships; and a longtime associate of Huang whom external complainants described as a Chinese intelligence asset. The board retained outside counsel to lead evidence-gathering and formed a special committee of independent directors. Internal emails obtained by The Verge show incoming board chair Harvey Jones telling staff the company faces an 'extraordinarily difficult period' and asking each business

Nvidia is the biggest beneficiary of this AI infrastructure cycle: it crossed a $4 trillion market value in July 2025 and briefly touched $5 trillion in October, with data-centre revenue of $115.2 billion, or 88% of total sales, in fiscal 2025. Scale makes the governance gap more visible. For 33 years Huang was simultaneously CEO, president and the company's only external voice, and the board had no chair until Harvey Jones took the newly created role now. Precedents are close at hand: OpenAI's board removed Sam Altman within hours in November 2023 and reversed within two weeks; Intel's Pat Gelsinger exited under board pressure in December 2024. Both show that at peak AI capex, star-CEO

Liu Gong's take: within six months, Nvidia will most likely rebuild itself around a dual structure of chairman plus CEO, creating a president or co-COO role for the first time. That is not a patch; it is the belated unwinding of a 33-year single-point dependency.
From a technical standpoint, the thing to watch is not the scandal but roadmap continuity. Nvidia refreshes architecture every two years and its product line every year; Rubin wafers and CoWoS capacity were locked into TSMC's schedule two years ago, and the roadmap is carried by a thousand-plus engineers hitting milestones rather than by one person. What the CEO genuinely owns is the non-standardised side: custom customer terms, export-licence conversations with governments, and top-level trust with TSMC and SK Hynix. Every week without a signatory lengthens that queue.
On the business side, orders will not change but pricing will. Hyperscaler purchases are locked quarterly, so near-term cancellations are unlikely; renewals are another matter, and 'management instability' is exactly the lever a procurement team pulls against a 75% gross margin. The larger shift is psychological. Over two years the hardest barrier for AMD's MI line, Google's TPUs and Broadcom's custom ASICs was never peak FLOPS, it was the buyer's fear of migration. That barrier just lost a brick.
For the industry, this becomes a governance case study alongside OpenAI's November 2023 board crisis and Intel's December 2024 leadership exit: star-CEO risk events are now processed as governance, not public relations.
In industrial inspection, the field Liu Gong works in, the pattern is consistent: a vision system takes well over six months from line trial to mass production, and when a supplier's leadership turns over, plants rarely switch vendors; they push the new programme's validation back by roughly a quarter. Expect something similar for Rubin customer qualification.
The falsifiable call: if Nvidia names an internal CEO before the end of Q1 2027, its data-centre share holds through 2027; if the vacancy runs past six months, AMD and custom-ASIC vendors combined could move from under 20% to above 25% within a year.

  • Tenure: Huang served from 1993 to 2026, 33 years, versus a median S&P 500 CEO tenure of roughly 7 years, close to five times as long.
  • Revenue mix: fiscal 2025 revenue of $130.5 billion, of which data centre contributed $115.2 billion, or 88%, up from 78% a year earlier.
  • Profitability: fiscal 2025 GAAP gross margin of 75.0%, 18.1 percentage points above the 56.9% recorded in fiscal
  • – Market value: about $4.5 trillion before the resignation, more than 30 times Intel's roughly $150 billion scale.
  • Position: third-party estimates put Nvidia above 80% of the AI accelerator market, with AMD and in-house ASICs together below 20%.

For the industry, the first three months are about contracts: hyperscalers and OEMs will ask for written supply commitments, procurement teams will write 'management turbulence' into renewal terms, and AMD, Broadcom's custom ASICs and Google TPU get a rare opening. Over twelve months, a smooth transition with an internal CEO likely preserves both roadmap cadence and share; if the investigation reaches China business specifics, export licensing and compliance become a second front larger than the personnel question.
For ordinary users, retail GPU prices and availability stay broadly unchanged in the near term, since consumer lines still queue behind data centre allocation; what transmits

Questions worth exploring:

  • A one-man company for 33 years: why the board only created its first chair now, and what the timing reveals
  • How the alleged spy line interacts with export controls and China sales to form a second risk curve
  • A leadership change during the Rubin ramp: knock-on effects for TSMC allocation and hyperscaler orders

Further reading: 2026 Frontier LLM Comparison: A Practical Guide — a durable, evergreen reference. Bookmark it.

This item comes from The Verge AI (published 2026-11-12 15:48:12). Wenfeng Radar provides a Chinese/English summary and commentary on overseas AI developments; the original content is owned by its author.


Daily aggregation of overseas AI developments with in-depth analysis. Bookmark us to never miss an important signal; back to homepage.

Republishing: feel free to quote or repost with a link back to the original article on Wenfeng Radar. · Business & partnership: see our media kit
Daily briefing + Sunday review · Unsubscribe anytime by replying · Your email stays private

Leave a Reply

Your email address will not be published. Required fields are marked *

中文EN
Contact · Privacy · Disclaimer · AI Disclosure · Weekly Review · Media & Business
Email: wenfengleida@wenfengleida.cn