TAIPEI—Nvidia's Computex keynote, leading off the Monday “pregame” day for Asia’s biggest IT trade show (its first major in-person edition since 2019), was a tour de force of AI and data center innovations. And, if you take Nvidia CEO Jensen Huang's word for it, these moves will touch every business, and every industry, in the years to come. (Wall Street would seem to agree.) That's because generative AI is poised to infiltrate companies and disciplines untouched by AI initiatives to this point. The buzzy new tech is poised to transform software development as we know it, too.
Nvidia's hardware and other initiatives are key underpinnings in the changes happening in the AI space today. Accelerated computing is a theme Huang returned to time and again in the keynote. Nvidia defines accelerated computing as "the use of specialized hardware to dramatically speed up work, often with parallel processing that bundles frequently occurring tasks." It's used in demanding fields such as machine learning, data analytics, and simulations, and GPUs are a robust, common accelerator that puts Nvidia right in the heart of the market.
Huang noted that this presentation was his first live in-person keynote since the onset of the COVID-19 pandemic (“Wish me luck!”) And he had plenty pent up to present: The keynote ran a full 2 hours and would require more unpacking than an entire football team’s trailer full of gear. We boiled it down to eight highlights.
Huang gave some perspective for the rise and potency of GPU compute as it relates to AI models and the capabilities that are now in-market. Given the insatiable demand for compute power to train AI models, he used $10 million as a benchmark. Whereas not long ago, that money would buy you
Read more on pcmag.com