Lisa SuChair and Chief Executive Officer, Advanced Micro Devices, Inc. (AMD)
10 recorded appearances
Su and Hugging Face's Clem Delangue make the case that open models and open software stacks are AMD's route into an Nvidia-dominated market, rather than competing purely on silicon.
Su frames current AI compute demand as structurally unusual rather than cyclical, and points to data-centre GPU ramp as the driver of the quarter.
Su launches AMD's end-to-end AI portfolio with ecosystem partners and customers, positioning AMD as a full rack-scale systems vendor rather than a chip supplier.
Su puts a concrete number on the Helios rack-scale platform's generational gain and argues rack-level integration, not the accelerator alone, is where the competition now sits.
MIT commencement address urging graduates to be ambitious about which problems they choose and to "run toward" the hardest ones as AI reshapes engineering work.
Su attributes the next leg of compute demand specifically to agentic workloads, which consume far more inference tokens per task than chat.
Hour-long lecture and Q&A on her engineering career, AMD's decade-long turnaround, and semiconductor manufacturing policy.
Post-results interview covering the data-centre ramp and AMD's server CPU share gains alongside the AI accelerator business.
Su's CES keynote lays out an "AI everywhere" arc from cloud to edge to PCs, with guests including OpenAI's Greg Brockman, Liquid AI's Ramin Hasani and World Labs' Fei-Fei Li.
Su pushes back directly on AI-capex-bubble worries, arguing the spend is justified by demand visibility rather than speculation.