AI Made Friendly HERE

Nvidia’s Jensen Huang Is About to Give Believers a Look at AI’s Future

  • All eyes will be on Jensen Huang next week.
  • The Nvidia CEO is set to host the “Woodstock of AI,” which will host a who’s who of the sector.
  • Nvidia has been tipped to reveal new details of the next generation of its H100 chip.

Thanks for signing up!

Access your favorite topics in a personalized feed while you’re on the go.

download the app

There’s a running gag that Jensen Huang is to AI fanboys what Taylor Swift is to Swifties. Next week could prove it right.

The Nvidia CEO is set to host the latest installment of his GPU Technology Conference (GTC) — almost 15 years after the first event took place in San Jose’s Fairmont Hotel — at a time when the biker-styled billionaire has become AI’s man of the moment.

Nvidia has been on a roller coaster ride since the start of the generative AI boom, as everyone from Sam Altman to Mark Zuckerberg has been scrambling to secure enough computing power for their AI models by buying up Nvidia’s chips, known as GPUs.

Last month, the company gave a sense of just how relentless demand has been after it announced $22.1 billion in revenue for its latest quarter. That’s 265% more than it earned in the same quarter last year.

At well over $2.2 trillion in market capitalization now — in touching distance of Apple — many are wondering where next for Nvidia and AI at-large. Nvidia’s GTC event, starting on Monday, should answer that.

All eyes on GTC

The hype going into the event hasn’t been lost on Nvidia’s employees. Bojan Tunguz, a machine learning researcher at Nvidia, has gone as far as describing this year’s GTC as the “Woodstock of AI.”

It’s not hard to see why.

For one, the AI community will be turning up to hear from Nvidia boss Huang as he’s widely anticipated to reveal a new generation of technology that changes the limits of what’s possible.

Matt Bryson, an analyst at Wedbush, expects Nvidia to lift the lid on the B100, the next-generation version of its H100 GPU. The “B,” which stands for “Blackwell,” is expected to offer new chip architecture that provides “significant performance gains,” he wrote in a research note.

It’ll be a huge deal if so.

Tech execs who have stockpiled H100 chips amid supply constraints and fears of falling behind competitors without them will likely fall over themselves for an upgrade. Zuckerberg, who said he’ll have around 350,000 H100 GPUs by the end of 2024, will surely want in on the B100 too.

mark zuckerberg pointing

Mark Zuckerberg is stockpiling Nvidia GPUs.

JOSH EDELSON


The conference itself is also stacked with a who’s who of AI today.

Leaders ranging from OpenAI’s COO Brad Lightcap to Arthur Mensch, CEO of OpenAI rival Mistral AI, will be there. So too will Christian Szegedy, a cofounder of Elon Musk’s xAI; Cohere CEO Aidan Gomez; Stanford’s Fei-Fei Li, and plenty more.

Though conference panels typically offer little insight beyond surface platitudes, open discussions among some of the most important people shaping AI could raise key questions about challenges and opportunities.

Are large language models all there is to AI in the future? Is there a solution coming to solve the chip supply crunch? Can AI models eventually reason?

Of course, industry watchers will look for any sign that the Nvidia and AI mania might be about to slow down.

As Wedbush’s Bryson noted, for instance, the last time Nvidia announced an upgrade in its main GPU, it saw “pricing more than double.” If the new chip comes with a similar price hike, it’ll be worth noticing what impact that might have on demand.

That’s especially so at a time when plenty of companies are having a go at rivaling Nvidia too. The likes of Meta, Microsoft, Google, and AMD are all working on rival chip designs to Nvidia’s GPUs.

Expect Nvidia to set the tone for AI’s next phase.

Originally Appeared Here

You May Also Like

About the Author:

Early Bird