AMD CES 2026: Lisa Su Unveils the Yotta Era and AI's Future

The Yotta Era of AI Computing

• 🚀 Dr. Lisa Su announced the "Yotta Era", highlighting an unprecedented demand for computing that requires scaling global infrastructure 100 times.
• 📈 AI adoption has surged from 1 million to over 1 billion active users, with projections to reach 5 billion, making AI as indispensable as mobile phones and the internet.
• 💡 The demand for global compute infrastructure is expected to grow from 1 Zettaflop in 2022 to 10 Yottaflops in the coming years, a 100,000-fold increase from 2022 levels.

Advanced Cloud Infrastructure

• 🧠 AMD introduced Helios, a massive 3,200 kg rack-scale platform for the AI Yottaflop era, integrating the new Instinct MI455 accelerator with 320 billion transistors.
• 🛠️ The MI455 is paired with the next-generation Venice CPU (Zen 6 architecture), specifically co-engineered to double bandwidth and efficiently feed data to the GPUs.
• 🎯 Greg Brockman of OpenAI emphasized the shift from chatbots to AI agents, which can execute complex, multi-day workflows and make decisions, requiring insatiable compute demand.

AI on Personal Devices & Ecosystem

• 💻 AMD's new Ryzen AI 4000 Series and Ryzen AI Max processors integrate NPUs to run 200 billion parameter models locally, enabling private and efficient on-device AI.
• ✅ Liquid AI demonstrated proactive personal assistants running on laptops, capable of joining meetings, transcribing, analyzing emails, and drafting responses locally.
• 🚀 The Ryzen AI Halo development kit provides a pocket-sized AI workstation with 128 GB of unified memory, empowering developers to create local AI agents.
• 🌐 AMD champions an open software ecosystem with RockM, facilitating multimodal general intelligence and simulating the world (e.g., Luma's Ray 3 model).

Transformative Applications of AI

• 🌌 Dr. Fei Fei Li of World Labs showcased spatial intelligence for generating interactive 3D worlds from single images, revolutionizing video game development and architecture.
• 🔬 In health and medicine, generative AI is designing drugs (APSI), accelerating DNA sequencing (Illumina), and reducing drug discovery time (AstraZeneca) to enable preventive care.
• 🤖 Physical AI is demonstrated through Generative Bionics' Jean One humanoid robot, which processes touch, and edge AI is critical for lunar permanence with Blue Origin, enabling real-time decisions in space.

Supercomputing and Scientific Advancement

• ✨ AMD powers the world's fastest supercomputers, including Captain, which are crucial for scientific breakthroughs.
• 🇺🇸 The US Genesis mission aims to integrate supercomputing and AI to double American scientific productivity within a decade, utilizing AI factories for science.
• 💡 A student hackathon showcased how Ryzen AI laptops are lowering the barrier to entry for AI development, fostering creativity and demonstrating the power of an open ecosystem.