Jensen Huang

Who they are

Jensen Huang is founder and CEO of NVIDIA — co-founded it in 1993 with $600 in seed capital to solve 3D graphics for PCs, and has run it ever since.

Person

Jensen Huang co-founded NVIDIA in 1993 when the company had roughly $600 in initial capital, three founders, and a bet on 3D graphics for personal computers — he has been CEO every day since. Before that, he was a microprocessor designer at AMD and a design engineer at LSI Logic; before engineering, he worked as a dishwasher, busboy, and waiter at Denny's. He earned his BSEE at Oregon State University in 1984 and his MSEE at Stanford in 1992 — the Stanford degree came after he'd already been in industry, which tracks with someone who builds credentials around a specific problem rather than on a schedule. The through-line across his career is a willingness to make a concentrated, long-horizon bet and hold it: NVIDIA pivoted from graphics to parallel computing to AI infrastructure across three decades under the same founder-CEO. He runs a famously flat organization — reportedly around 60 direct reports, mostly engineers — and publicly prefers group problem-solving over one-on-one meetings. He writes on LinkedIn and the NVIDIA blog about AI as essential infrastructure, open models, and physical AI; he keynoted CES 2025, GTC 2025, Computex Taipei 2026, and delivered the 2026 CMU commencement address; he was appointed to the U.S. President's Council of Advisors on Science and Technology in 2026 and received the IEEE Medal of Honor. He and his wife Lori run the Jen-Hsun & Lori Huang Foundation, which has donated to educational institutions including Stanford and Oneida Baptist Institute.

Company

The most recent move is the June 2026 launch of RTX Spark — NVIDIA's first processor for Windows laptops and desktops, combining a 20-core Arm CPU with a Blackwell GPU, with devices from Dell, Microsoft, HP, ASUS, and Lenovo arriving in fall 2026; this signals NVIDIA pushing from the data center into the PC endpoint. Also in June 2026, Huang unveiled the NVIDIA Constellation campus in Taipei's Beitou Shilin Technology Park, spanning nearly 4 hectares and built for about 4,000 employees. The company reported fiscal 2026 revenue of $215.9 billion and raised its quarterly dividend 25x from $0.01 to $0.25. On the M&A front, NVIDIA completed a $20 billion deal for Groq's assets (structured as a licensing and acquihire to reduce antitrust exposure) and acquired SchedMD, Illumex, and Kumo in 2026, while committing over $40 billion in AI equity investments in the first four months of 2026 — including $30 billion to OpenAI — positioning itself as the largest AI infrastructure financier. The Groq acquisition is under formal Senate inquiry from Senators Elizabeth Warren and Richard Blumenthal for potential antitrust concerns.

Market

NVIDIA holds roughly 90% GPU market share in AI accelerators and data center compute, with CUDA's developer ecosystem — over 5 million developers as of 2025 — acting as its deepest structural moat. AMD (MI450 GPUs), Intel, and custom ASIC builders at Google, Amazon, and Microsoft are the credible challengers, with competition intensifying as AI workloads diversify. US export controls on advanced chips to China, the geopolitical risk of TSMC dependency in Taiwan, and rising regulatory scrutiny of NVIDIA's acquisitions are the main structural headwinds shaping the next phase.

Network

Huang's most visible recent relationship is with TSMC leadership — he has appeared publicly with Y.L. Wang (TSMC Vice President of Operations) and Ray Chuang (CEO of TSMC Arizona) around the first NVIDIA Blackwell wafer produced in the US, reflecting the strategic weight of the TSMC partnership. Sequoia Capital was an early investor in NVIDIA at its 1993 founding and remains a long-standing institutional connection. Huang also joined a presidential delegation to Beijing in May 2026 alongside other US tech leaders in an attempt to unlock stalled chip sales in China.

Y.L. Wang· Vice President of Operations, TSMC
Ray Chuang· CEO, TSMC Arizona

How they likely show up

Founder and CEO since 1993 — over three decades at the helm of the same company — → thinks in decade-long arcs; short-term quarterly pressure is not how he frames decisions.
Runs ~60 direct reports, mostly engineers → he flattens hierarchy deliberately; he won't respond well to bureaucratic escalation framing or org-chart politics.
Publicly prefers group problem-solving over one-on-one meetings → expect him to bring things into the room, not settle them bilaterally; conversations are likely to be more Socratic than transactional.
BSEE then MSEE taken mid-career at Stanford → returns to first principles when stuck; the Stanford GSB talk on first-principles thinking confirms this is a stated method, not a pose.
Active at flagship technical conferences (GTC, CES, Computex) and delivers commencement addresses → comfortable commanding large rooms and shaping narratives at scale; not a behind-the-scenes operator.
Structured the Groq deal as licensing + acquihire specifically to reduce antitrust exposure → reads regulatory and structural risk carefully and engineers around it rather than confronting it head-on.

Conversation tips

→ Come in with a technical or architectural angle — he manages mostly engineers and thinks like one; business-speak without substance will lose him fast.
→ Reference a specific product or platform by name (RTX Spark, Vera Rubin, Blackwell, CUDA) rather than talking about 'AI chips' generically — specificity signals you've done the work.
→ Ask about the tension between NVIDIA's fabless model and geopolitical supply chain risk — it's a live strategic problem he's actively navigating with TSMC and US manufacturing investments.
→ The Groq deal's antitrust scrutiny is a live issue; don't treat it as resolved — he engineered the deal structure to manage it, so it's a meaningful topic if you can speak to it precisely.
→ He's on record valuing adversity and perseverance as core principles — if relevant, acknowledge the long hard periods in NVIDIA's history (early layoffs, the near-death moments before GPU compute took off) rather than only the recent triumph.

Toolbox

Openers

Open on the RTX Spark launch — NVIDIA's first Windows PC processor combining a 20-core Arm CPU with a Blackwell GPU, shipping in fall 2026. It's a direct move into Intel and Qualcomm's territory and a signal about where NVIDIA thinks AI inference is heading.
Reference the Groq deal structure — $20 billion framed as licensing and acquihire rather than a straight acquisition, specifically to reduce antitrust exposure. He made a deliberate legal and structural bet; it's now under Senate inquiry, which makes it a live, consequential decision.
Mention the 1993 founding — $600 in seed capital, three co-founders, a bet on 3D graphics — as a reference point against the $4.6 trillion market cap today. It frames the conversation around long-arc conviction rather than recent momentum, which is how he tends to think.

Discovery questions

With RTX Spark pushing Blackwell into Windows PCs alongside the Vera CPU for data centers, how do you think about NVIDIA's CPU ambitions relative to the software ecosystem lock-in CUDA created in GPUs?
The Groq deal was structured as licensing and acquihire to sidestep antitrust review — now it's under formal Senate inquiry anyway. How do you think about the line between vertical integration and market dominance when you're at 90% GPU share?
You've described running roughly 60 direct reports and preferring group problem-solving over one-on-ones — at NVIDIA's current scale, how do you keep that flat structure from becoming a bottleneck rather than an accelerant?

Avoid

Don't frame NVIDIA's position as a recent overnight success — he's been building this for over 30 years through multiple near-failures, and flattening that into 'you won the AI boom' will signal you haven't done your homework.

Make it yours