Jensen Huang, CEO and co-founder of NVIDIA, has been selling NVDA stock while AI chip demand reaches unprecedented levels. NVIDIA's revenue tripled. The stock 10x'd. And the CEO sold.
Every retail investor has the same question: if AI demand is so strong, why is Jensen selling?
The answer reveals a fundamental misunderstanding of how founder-CEO selling works.
What happened
- Who: Jensen Huang, CEO & co-founder (since 1993)
- What: Sold shares of NVDA through a pre-announced 10b5-1 plan
- When: During 2024-2025, while NVDA stock appreciated 800%+
- Context: NVIDIA revenue went from $27B (FY2023) to $130B+ (FY2025), driven entirely by AI chip demand
Why selling during a boom is rational
1. Concentration risk at founder scale
Jensen Huang's NVIDIA stake has been worth $50-120 billion at various points. Even after selling, he remains one of the wealthiest people on Earth with massive NVDA exposure.
Holding $100B+ in a single stock is not prudent wealth management — it's extreme concentration. Selling a small percentage each quarter reduces this concentration to merely enormous.
2. The 10b5-1 plan was set up in advance
Huang's selling plan was pre-announced and pre-scheduled. 10b5-1 plans are established when the insider does NOT possess material non-public information. The actual trades execute automatically on a schedule.
This means: The decision to sell was made months before each transaction. The timing doesn't reflect Huang's current view — it reflects a plan he set up when the stock was lower.
3. Selling high is better than selling low
If a founder must diversify (and all financial advisors would recommend it at this concentration level), selling during a period of strength is the rational choice. Would you prefer Huang sell during a downturn?
4. Personal liquidity for a 60+ year old founder
Huang is in his early 60s. He co-founded NVIDIA in 1993 — over 30 years ago. Having liquid personal wealth outside of a single stock is basic estate and retirement planning, regardless of NVIDIA's prospects.
The numbers in context
| Metric | Approximate value |
|---|---|
| Huang's peak NVDA stake value | $100B+ |
| Shares sold in recent plans | Hundreds of millions in $ |
| Remaining stake value | Still tens of billions |
| % of holdings sold | Low single digits |
| NVIDIA's business trajectory | Revenue tripling, margins expanding |
He sold a sliver of an enormous position during the greatest AI hardware boom in history. He retains a stake worth more than the GDP of most countries.
The comparison that puts it in perspective
| Founder-CEO | Company | Sold during company strength? | Still running the company? |
|---|---|---|---|
| Jensen Huang | NVIDIA | Yes — during AI boom | Yes |
| Mark Zuckerberg | Meta | Yes — during Reels/AI pivot | Yes |
| Jeff Bezos | Amazon | Yes — throughout 2020s | No (transitioned to Exec Chair) |
| Elon Musk | Tesla | Yes — sold $40B+ in 2022 for Twitter acquisition | Yes |
Every major tech founder sells stock. The question isn't whether they sell — it's the pattern, magnitude, and context.
What WOULD be a bearish signal from Jensen
| Scenario | Signal strength |
|---|---|
| Selling on a pre-scheduled 10b5-1 plan | None — this is the current situation |
| Accelerating sales beyond the plan | Moderate — worth investigating |
| Terminating a 10b5-1 plan and selling immediately | Strong — broke the schedule |
| Selling while other NVDA officers are also selling unusual amounts | Strong — consensus insider concern |
| Stepping down from CEO while liquidating | Very strong — losing commitment |
| Making an open-market purchase | Extremely bullish — would break 30-year sell pattern |
None of the bearish scenarios currently apply.
What actually matters for NVDA investors
Instead of tracking Jensen's routine 10b5-1 sales, focus on:
13F institutional data
- Are hedge funds and active managers adding or trimming NVDA?
- Is NVDA's institutional holder count growing or shrinking?
- How does NVDA weight compare across different fund types (growth vs. value vs. quant)?
Competitive landscape
- AMD MI300/MI400 adoption rate
- Google TPU and Amazon Trainium progress
- Customer concentration (Microsoft, Meta, Google = 40%+ of NVDA datacenter revenue)
AI demand sustainability
- Hyperscaler capex guidance (are they still spending?)
- Inference vs. training demand shift
- China export restrictions impact
The bottom line
Jensen Huang selling NVIDIA stock during the AI boom is founder-level wealth management, not a market signal. He set up a selling plan, it's executing on schedule, and he retains a stake worth tens of billions.
The real signals for NVIDIA live in institutional 13F data, competitive positioning, and AI demand trajectories — not in a 60-year-old billionaire's routine diversification plan.
Originally published at 13F Insight
Top comments (0)