AMD CEO Lisa Su Emphasizes AI Inference Growth at Advancing AI Conference
At AMD’s Advancing AI 2026 conference in San Francisco, CEO Lisa Su highlighted the shift from training to inference as a key inflection point for artificial intelligence. She emphasized that businesses are now deploying AI systems in real-world workflows rather than just testing them. According to internal data cited by CryptoBriefing, monthly global AI token consumption has reached 35 quadrillion tokens—160 times higher over two years. Su also noted the growing importance of CPUs alongside GPUs for agentic AI applications that require orchestration and coordination tasks. AMD revised its server CPU market estimate upward to $220 billion by 2030, up from a previous projection of $120 billion. The company sees total AI compute as a potential $2 trillion opportunity with inference driving most growth. AMD recently launched the EPYC Venice processor based on TSMC’s 2nm architecture and is positioned to benefit from rising demand in both CPU and GPU markets.
Topics: AI AMD CEO Lisa Su inference computing semiconductor industry