AI race pits domestic chips against global guardrails

Alibaba Takes On Nvidia as the Global AI Race Heats Up

The episode connects chip competition, frontier-AI policy, education, financing, and world-model governance to a widening struggle over who controls AI’s direction.

3 key takeaways

1Alibaba is developing training-focused AI accelerators as China seeks alternatives to Nvidia and Huawei.
2U.S. policy favors frontier-AI development over new international controls, sharpening tensions around AI diplomacy.
3Fei-Fei Li argues world models require ecosystem-wide safety standards and firm human control.

Don't miss

Fei-Fei Li explains how world models move beyond language into pixels, geometry, and controllable simulated environments, while insisting humans remain in control.

The brief

At the United Nations, U.S. AI policy is framed around superintelligence and competitiveness rather than international controls, as China and Nvidia remain central to the debate.

Bloomberg’s Mike Shepard explains how AI diplomacy is entering the UN agenda, while Alibaba develops domestic accelerators aimed at challenging Nvidia and Huawei.

An alternative to college puts AI-native projects, mentorship, people skills, and startup funding at the center of education for ambitious young builders.

Fei-Fei Li distinguishes world models from language models, describing systems that generate pixels, geometry, and controllable 3D or 4D environments.

Li’s central warning is institutional: world models need capability and safety benchmarks shaped by industry, government, academia, open source, and civil society.

What was said on this episode

23 statements · 19 positive · 3 negative · 1 neutral

Fei-Fei Lion Company self-regulation of AIPositive5:40
Company self-regulation can support continued growth and U.S. disinflationary growth
“I think self-regulation from the company, solutions around the problems allow us to continue to grow and continue to see disinflationary growth in the U.S. economy”
Listen at 5:40
Fei-Fei Lion Artificial IntelligencePositive6:47
AI is more powerful than the 1990s Internet and will change how people live
“This technology is much more powerful than the Internet that we had in the 90s, obviously, and also will change the way we live.”
Listen at 6:47
Fei-Fei Lion ChinaNegative7:11
China will not change its behavior
“I don't think we'll see China ever change their behavior.”
Listen at 7:11
Fei-Fei Lion ChinaNegative7:27
China will not self-regulate AI
“I don't expect China to self-regulate.”
Listen at 7:27
Fei-Fei Lion Meta revenue growthPositive7:59
Meta revenue growth will compound at 20%
“We think the revenue growth will compound at 20%.”
Listen at 7:59
Fei-Fei Lion Meta advertising businessPositive8:11
Meta’s advertising business is gaining market share
“the advertising business is taking share.”
Listen at 8:11
Fei-Fei Lion Meta MusePositive8:19
Meta Muse materially improves Meta’s investment outlook
“Muse changes the calculus materially.”
Listen at 8:19
Fei-Fei Lion Meta MusePositive8:25
Muse can help Meta monetize its capital expenditures
“It is a way for them finally to monetize the CapEx spend.”
Listen at 8:25
Peter Elstromon U.S.-China AI competitionNeutral11:10
The United States and China are competing to dominate AI
“those two countries really are locked in competition to try to dominate this technology.”
Listen at 11:10
Peter Elstromon U.S.-China AI meetingNegative13:28
The U.S.-China AI meeting is unlikely to produce major advances
“we are not expecting big advances.”
Listen at 13:28
Peter Elstromon Trump administration economic agendaPositive14:51
AI is central to the Trump administration’s current economic agenda
“AI is central to the administration's economic agenda right now”
Listen at 14:51
Gagan Biyanion AI-native buildersPositive23:14
AI-native builders can use AI to build software, products, or hardware
“they're AI native builders. So they understand how to use the technology, whether that's building software, products, or hardware.”
Listen at 23:14
Gagan Biyanion Horowitz Andreessen AcademyPositive26:03
The academy will help students obtain startup funding
“we're absolutely going to help students at the academy get funding if they want to.”
Listen at 26:03
Gagan Biyanion San FranciscoPositive28:14
San Francisco is an innovation and business center and is becoming a cultural center
“San Francisco is the center of innovation. It's the center of business. And it's now becoming the center of culture for a lot of things.”
Listen at 28:14
Fei-Fei Lion World Labs world modelsPositive32:56
World Labs world models generate pixels and accurate 3D or 4D geometry
“The models generate beautiful pixels, very accurate geometric structure of the”
Listen at 32:56
Gagan Biyanion AtlasPositive33:32
Atlas is World Labs’ most advanced world model
“It is the latest, the greatest world model that World Labs has developed.”
Listen at 33:32
Peter Elstromon AtlasPositive34:06
Atlas outperforms specialized state-of-the-art models on its tasks
“It's better than any state-of-the-art specialized model for these tasks.”
Listen at 34:06
Gagan Biyanion World model (artificial intelligence)Positive35:55
Robust world models can reduce robotics training and evaluation bottlenecks
“A robust world model can help unlock some of this bottleneck.”
Listen at 35:55
Gagan Biyanion Technology development and deploymentPositive37:05
Technology development and deployment should prioritize safety and responsibility
“we have always to put safety and responsibility in the forefront of the development and deployment of technology.”
Listen at 37:05
Fei-Fei Lion AI governance ecosystemPositive39:02
AI should develop through collaboration across public, private, academic, and civil sectors
“We need to understand. and keep this technology, continue to develop this technology in a robust ecosystem where there is public sector, there is open source, there is closed source, there is collaboration with academia, with government, with civil society.”
Listen at 39:02
Fei-Fei Lion Artificial IntelligencePositive40:17
Humans must remain in control of AI
“If we, we must be in control.”
Listen at 40:17
Gagan Biyanion AI evaluation and benchmarkingPositive41:19
AI evaluation and benchmarking should be an ecosystem-wide responsibility
“this problem, this question of evaluation and benchmarking should be an ecosystem level responsibility.”
Listen at 41:19
Gagan Biyanion AI benchmarksPositive42:01
AI benchmarks should involve independent, public-sector, industry, and government bodies
“I continue to believe importance of benchmark from independent bodies and public sector bodies like academia, as well as the shared responsibility of industry, of government together.”
Listen at 42:01

Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.