Nvidia Corp. kicks off its annual GPU Technology Conference Monday in San Jose. After five years of being virtual, the nearly weeklong event returns to an in-person format, and the timing could not be better.
This is arguably the most important GTC in history. What started as a vendor-specific show attended by gamers and data scientists exploded into the world’s largest artificial intelligence-centric industry event. The event is expected to draw a big crowd, close to 20,000 people in person and more than 300,000 online. For those attending or watching, here is what to expect:
An epic keynote
The highlight of every GTC is the keynote speech by Chief Executive Jensen Huang (pictured). Over the years, these have proven insightful, inspiring and sometimes controversial, and I expect this on steroids. Huang has called ChatGPT the iPhone moment for AI and talked about AI being the biggest technology transformation in history. If this is the case and Nvidia is the enabler of this, we should see plenty of vision for what AI can do for us and what we can do with AI.
One indicator of the magnitude of the keynote is that it will be held in the SAP Center, home of the San Jose Sharks. This venue is typically used for sporting events, concerts and other big events. I saw the Eagles there, and Justin Timberlake is coming, which indicates the breadth of interest in Nvidia and AI.
AI use cases
For all the hype and industry chatter around AI, production use cases have been hard to find. Although AI promises to transform every aspect of how we work, learn and entertain ourselves, it’s often hard to understand what’s possible until we see it. Huang was credited once with saying something to the effect of, “The more creativity we bring to AI, the more AI can bring to creativity.”
I agree with this sentiment, but it’s often hard for people to bring the creative aspect to something until there are some examples to work from. This is where, during the keynote and across the event, I hope to see AI use cases that can “prime the pump” and get those creative juices flowing so the AI reality can catch up to the vision. Specifically, within the broader AI universe, I’m expecting a heavy dose of generative AI as that seems to have become the main focal point for organizations.
New technology
What would a GTC be without a flurry of product announcements? Over the years, Nvidia has announced innovations such as NVLink, NVSwitch, Clara for healthcare, Drive kit updates for automotive, and a wide range of graphics processing unit form factors for different use cases. This year, it’s expected there will be details on the B100 Blackwell chip, which will supplant the H200 GPU as the company’s most powerful AI chip.
Industry watchers I have talked to expect the chip to deliver an order of magnitude faster training than its current chip. One interesting data point to look for is the power requirements. Nvidia has provided the world with silicon of all form factors to power every type of accelerated computing, but its chips draw a significant amount of power. Rumors were that the B100 runs so hot that it needs liquid cooling, but earlier this month, Jeff Clark, Dell’s COO, stated while the chip would require 1000 watts, 42% more than the previous chip, it would not require liquid cooling.
The more powerful the silicon is, the more power it requires, and that has implications concerning the plans most organizations have to reduce costs, data center footprint and sustainability. I’ll be curious to see if Nvidia discusses this and if there’s a way to build less power-hungry GPUs. Perhaps AI can find a way?
The societal impact of AI
Even though we all want AI to improve customer experience, help us drive more safely and usher in the era of robots in the workplace, AI has the power to solve some of the planet’s unsolvable problems. One of the industries in which AI can have the most significant impact is healthcare, and GTC has always featured several dedicated sessions. I took a quick look through the agenda, and the sessions that caught my eye were transforming drug discovery, how generative AI can be used in modern medicine and the use of AI-based surgical robots.
Some smart city technologies can help municipalities deal with the crushing weight of traffic problems. One interesting session I’m looking forward to is using computer vision to deal with London’s commuting conditions. Another is how cities and departments of transportation can use AI for road safety.
The societal topic on which AI can have the most impact is sustainability, and that’s well-represented at GTC. Key sessions include transforming agriculture with AI, edge computing for waste contamination detection, and the sustainable futures startup showcase.
The growing AI ecosystem
One of the highlights of any trade show is the Expo Hall, where attendees can see solutions. It looks like there will be about 300 exhibitors at GTC, which is a very large number for a vendor show. One thing I have been trying to understand better is who in the ecosystem will catch above-average tailwinds from AI.
Because we are so early in the AI cycle, Nvidia’s growth hasn’t translated into any other vendor, at least not yet. Arista Networks, Pure Storage and Supermicro are three companies disrupting their industries. Concerning security, there isn’t an obvious cyber beneficiary, although CrowdStrike Holdings Inc. seems like a likely disruptor, as does Zscaler Inc., particularly after its acquisition of Avalor Technologies Inc.
AI will create a rising tide that will lift many boats, but a few will ride the wave — perhaps not as strongly as Nvidia will, but well ahead of their peers. Attendees should use time in the Expo Hall to learn about the ecosystem partners that can help them with their AI journeys.
