Nvidia CEO Jensen Huang took the stage at the tech giant’s GTC Washington, DC conference wearing his usual black leather jacket, but he might as well have accessorized his go-to outfit with a red MAGA cap.
In a nearly two-hour long keynote at the Walter E. Washington Convention Center here, Huang hyped Nvidia’s current and coming progress in AI computing while announcing multiple new partners, with the Trump administration high on that list.
Huang took a nationalistic tone early on in the conference, which he called “the Super Bowl of AI.” He announced the first in a series of partnerships, a venture with Nokia to build gear for the 6G networks that Nokia predicts will enter service by 2030.
The CEO said that after decades of American innovation, today’s wireless networks are now “largely deployed on foreign technologies,” going a step beyond earlier, bipartisan expressions of concern over Chinese and Russian influence in 5G standards and network hardware.
“That has to stop,” Huang said.
The Nokia partnership, which involves Nvidia investing $1 billion in that Finnish firm, will yield a platform called Arc—an acronym nesting other acronyms: “Aerial RAN (Radio Access Network) Computer”—which Huang described as a software-defined 6G base station.
“For the first time, we’ll be able to use AI technology, AI for RAN, to make radio communications more spectral-efficient,” he predicted.
(The GTC Wi-Fi itself could have used some optimization at that point, with attendees next to me complaining about it not working.)
‘A Fundamental Breakthrough’
Next, Huang turned to a tech frontier he has been skeptical about before—quantum computing and its potential for exponentially more powerful processors if engineers can surmount its error-prone nature. “The industry has made a fundamental breakthrough,” he said, nodding to recent advances in making qubits—quantum’s equivalent of a bit in classical computing, except it’s much harder to work with—”coherent, stable, and error-corrected.”
Nvidia’s latest contribution to that field is a system architecture called NVQLink that can integrate a quantum processing unit with a GPU and a conventional CPU, integrating error correction for the fragile output of a quantum computer.
Nvidia’s description of its partners amounts to ‘I have friends everywhere.’ (Credit: Kent Nishimura/Bloomberg via Getty Images)
Eight Department of Energy labs will be working with Nvidia to experiment with this technology, and Huang further announced that Nokia will collaborate with Oracle to build seven new AI supercomputers for DOE labs to apply that level processing power to some of science’s harder problems. “These scientific supercomputers are not going to run chatbots,” Huang said. “They’re going to do basic science.”
All of this AI infrastructure demands enormous amounts of power, which doesn’t bode well for home electric rates, and Huang took a moment to commend the current administration for its energy policies. “President Trump deserves enormous credit for putting the weight of the nation behind pro-energy growth,” he said.
“If this didn’t happen, we could have been in a bad situation, and I want to thank President Trump for that,” Huang declared.
Get Our Best Stories!
Your Daily Dose of Our Top Tech News
By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy Policy.
Thanks for signing up!
Your subscription has been confirmed. Keep an eye on your inbox!
Huang’s description demands clarification: While Trump has been a vocal supporter of fossil fuels as well as nuclear energy, as seen in his AI Action Plan, his open hostility towards solar and wind power has led to the cancellation of renewable-energy projects that would have added gigawatts of capacity to the electric grid.
Huang got back on message about the virtues of AI, comparing it to conventional software that exists as a tool for use at a specific task: “AI is not a tool; AI is work.” With companies now willing to pay a premium for AI’s ability to advance their work, he said AI had reached the “virtuous cycle” stage. “The more profit generated, the more compute that’s put on the grid,” he said.
Nvidia’s current Grace Blackwell and upcoming Vera Rubin GPU platforms will help that cycle spin faster by making AI cheaper to use, Huang attested.
“The speed difference is incredible,” he said of the company’s Grace Blackwell NVL72, citing SemiAnalysis benchmark tests that found it yielded 10 times the performance per dollar and watt of Nvidia’s older H200 NLV8.
With further advances coming in Rubin, Nvidia projects $500 billion in sales of the two platforms combined over 2025 and 2026.
Recommended by Our Editors
A video highlighted how such manufacturing partners as TSMC, SK Hynix, and Foxconn are now making Blackwell hardware at US facilities, and Huang again thanked Trump for that. “The first thing that President Trump asked me for is, bring manufacturing back,” he said. “Nine months later, we are now manufacturing, in full production, Blackwell in Arizona.”
But bringing hardware manufacturing back isn’t a new plot twist that began under this administration. The Chips and Science Act that Congress passed in 2022 under President Biden included billions of dollars in subsidies and incentives to encourage that industrial shift, with TSMC and SK Hynix among the bigger beneficiaries.
Bring in the Robots
Huang has a big vision for humanoid robots. (Credit: Kent Nishimura/Bloomberg via Getty Images)
After recapping Nvidia’s work in using its Omniverse platform to develop “AI Factory” data centers and new partnerships with the defense-tech firm Palantir and the security firm CrowdStrike, Huang turned to talking up Nvidia’s hopes for robotics.
One is robots that interact with humans, such as Tesla’s humanoid, much-hyped Optimus and Disney’s BDX Droid, a more diminutive robot that evokes the scruffy hero of Pixar’s WALL-E.
This entire category remains vaporware at the consumer level, but Huang voiced a level of optimism that evoked Meta CEO Mark Zuckerberg’s already-busted metaverse predictions: “This is going to be the largest consumer electronics product line in the world.”
Another is robotaxis, which are now a commercial reality in a growing number of cities.
“There’s one robot that is clearly at an inflection point,” Huang said before touted Nvidia Drive AGX Hyperion, a platform that the company is developing to make cars “robotaxi-ready.” Stellantis, the parent firm of such makes as Chrysler and Fiat, Mercedes-Benz, and the EV maker Lucid are along for the ride, he said.
Huang wrapped up the keynote with one more grace note for the president who, like many, keeps mispronouncing the company’s name as “nuh-vidia” by giving a MAGA shout-out to audience members: “Thank you all for your service and making America great again.”
About Our Expert
Experience
Rob Pegoraro writes about interesting problems and possibilities in computers, gadgets, apps, services, telecom, and other things that beep or blink. He’s covered such developments as the evolution of the cell phone from 1G to 5G, the fall and rise of Apple, Google’s growth from obscure Yahoo rival to verb status, and the transformation of social media from CompuServe forums to Facebook’s billions of users. Pegoraro has met most of the founders of the internet and once received a single-word email reply from Steve Jobs.
Read Full Bio
