AMD's Lisa Su challenges Nvidia: 'My goal is to lead the AI revolution'

Lisa Su is considered the most powerful woman in the technology industry. In the interview she talks about the significance of artificial intelligence and why AMD can outpace Nvidia in the long run.

AMD's Lisa Su challenges Nvidia: 'My goal is to lead the AI revolution'
Image: AI-generated image

Lisa Su is the most powerful woman in the tech industry. When she took over AMD a decade ago, the company was close to bankruptcy. Today it is Nvidia’s most serious challenger. For a Handelsblatt special report I spent six months researching – and met Su in Austin for the interview.

What is at stake

Nvidia dominates the AI chip market as few companies dominate any key industry. Anyone who wants to train image, language or text models cannot avoid Jensen Huang’s graphics cards. Exactly one person has taken it upon herself to break that monopoly: Lisa Su. In the interview she explains how she intends to do it – and why AMD has the staying power.

The starting position

“We are still at the very beginning of the AI revolution,” Su says. “My goal is for us to lead this revolution.” A striking statement from an executive who usually holds back. AMD board member Ruth Cotter told me how close the company was to the edge before Su took charge – they had to sell furniture to stay liquid. Today AMD is one of only two suppliers of high-performance AI chips worldwide. The MI300 accelerators are in use at Microsoft and Meta; the AI revenue target has been revised upwards several times.

The European move

Remarkable is where Su is expanding: Europe. AMD is acquiring Silo AI, a Finnish start-up and the largest private AI laboratory in Europe, for 665 million dollars. The deal brings more than 300 AI researchers and engineers, and more than 200 AI projects for enterprise and cloud customers. “We will grow strongly in Europe,” Su says. It is AMD’s third AI acquisition in two years, following California’s Nod.AI and France’s Mipsology. Silo AI is also one of the few AI labs that has consistently bet on AMD hardware rather than Nvidia’s.

Why it matters

The AI economy needs a second chip supplier – otherwise the entire industry depends on Nvidia. Su is building an open software platform as a countermodel to Nvidia’s closed CUDA ecosystem. That is the real competition: not silicon against silicon, but ecosystem against ecosystem. And through the Silo AI acquisition, Europe is getting a role that few expected. The decisions of the next two years will show whether AMD can close the gap with Nvidia.

I wrote the full feature and interview for Handelsblatt.

More on Artificial Intelligence

All stories