The Affair Rally: Why NVIDIA’s Personal Drama Doesn’t Kill the Chip Thesis

Personal scandals have hit tech founders before. The stocks kept compounding. NVIDIA’s AI chip demand does not disappear because of bedroom headlines.

NVIDIA got smashed overnight. Board meeting on Jensen Huang’s future. Two affairs in the headlines. One with a woman intelligence types call a Chinese spy. Another with a senior executive who reports to him. Stock down 8% after hours. Hundreds of billions in market value wiped in a single session.

I’ve seen this movie. So have you.

Personal scandals hit tech founders hard in the short term. Then the business reasserts itself. Demand for the product doesn’t vanish because the CEO had a messy private life. The chip cycle doesn’t care who is sleeping with whom. What matters is whether customers still want the chips, whether the software stack still locks them in, and whether the order book is still growing. On all three of those, the answer has not changed.

History is clear — the stocks kept compounding

Look at the chart doing the rounds. Jeff Bezos and Lauren Sánchez. Amazon up more than 200% since the affair went public. Bill Gates and the Microsoft employee. Shares more than doubled. Sergey Brin at Google. Chey Tae-won at SK Hynix — that one is ridiculous, thousands of percent. These were not clean stories. Divorces. Public messes. In some cases, real governance questions and board pressure.

Historical price performance after high-profile personal scandals involving tech founders and CEOs. NVIDIA is currently the outlier to the downside.

In every case the market sold the scandal first. Then it bought the business. AI infrastructure spend, cloud growth, memory cycles — those things kept going. The founder’s private life was a headline for a few weeks. The product cycle was a multi-year story. Investors who sold the personal drama and missed the compounding paid a heavy price.

NVIDIA is in the same boat. The AI chip demand does not disappear because of this. Hyperscalers still need the GPUs. The software stack still runs on CUDA. Customers are not cancelling orders over bedroom news. The installed base is still locked in. The roadmap (Rubin and what comes after it) is still the one the industry is building around. None of that has changed in the last 72 hours.

The subordinate problem is real. The support is stronger.

One of the affairs was with a direct report. That changes the calculation for the board. Governance, culture, legal exposure — they have to do something. An investigation was already running. Now a second relationship is public. They cannot just wave it through and pretend nothing happened. That is normal board behaviour. It is also why the stock is down.

But look at the numbers. Shareholders voted 99% to re-elect him as a director at the June AGM. A blind employee poll showed 88% wanted him to stay as CEO. The board itself, as recently as 8 November, said it had confidence in Jensen and the leadership team. That is not nothing. It is a very different picture from a company where the founder has lost the room.

Shareholder, board and employee support for Jensen Huang as of early November 2026.

Key-person risk is obvious. Sixty-plus direct reports. No public successor. That structure is Huang’s. Losing him would be messy. Execution could slow. Customers might hedge. The valuation multiple would take a hit before earnings do. Those are real risks and they are already being priced.

None of that means the revenue disappears. CUDA is still the platform. The installed base is still locked in. The product roadmap is still the one hyperscalers are building around. An interim CEO — Colette Kress is the obvious name for stability, Ian Buck for the technical side — can keep the ship moving while a longer search runs. OpenAI fired Altman and reversed it inside a week. Boards can change their minds when the alternative looks worse. The business does not stop because the board has to show it is doing its job.

What the numbers still say

NVIDIA’s last reported quarter showed $96.2 billion in revenue, up 106% year on year. Data Centre was $89 billion of that — 92.5% of the total. Gross margin was 75%. Guidance for the next quarter was $108 billion. Those are not the numbers of a company whose customers are walking away. They are the numbers of a company that is still the default choice for the most demanding AI workloads.

AMD is making progress. ROCm is better than it was. Some inference and standard training workloads can now run on AMD with less friction than before. That is real. It is also not the same as matching NVIDIA on frontier-scale training clusters, full-system economics, and the depth of the software stack that customers have already built around. The gap is closing in places. It has not closed where it matters most.

That is why we still rate NVIDIA 5/5 in the house view. It sits in the AI infrastructure sleeve we want to hold through Slowdown. Cheque-cashers over cheque-writers. The structural demand for accelerated computing does not get cancelled by personal headlines. Margins are high. The moat is real. The stock just got cheaper on news that does not change the order book.

General advice only. Not personal financial advice. Past performance is not a reliable indicator of future results. Do your own work. Allegations are not findings of wrongdoing. Verify all developments before making investment decisions.
MPC Markets Pty Ltd (ABN 33 668 234 562, AFSL 296877).