Which Company Is Going To Win The AI Race? - David Friedberg
Speakers
Thomas Laffont
— Managing Director, Coatue Management
Chamath Palihapitiya
— CEO, Social Capital
Jason Calacanis
— Angel Investor / Founder, LAUNCH
David Friedberg
— CEO, The Production Board
Summary
In this All-In clip the hosts and guest Thomas Laffont each name the two companies they believe will win the AI race over the next five years. Laffont picks Nvidia first with Tesla as a dark horse, Chamath picks Tesla then Google, Jason backs Elon's combined Tesla and xAI stack over Google, and David Friedberg ranks Google first as a diversified portfolio of AI options with Tesla second. The debate covers GPU durability, vertical integration, Google's ability to monetize even if search declines, Tesla's humanoid-robot optionality, and a Chinese semiconductor threat to Nvidia.
- Nvidia is defended as the default AI compute layer because all model roads still lead to the GPU.
- Tesla is framed as the vertically integrated physical-AI bet, with Optimus as an embedded call option.
- Chamath and Jason argue Google wins even if search share erodes, via price-per-token monetization and a stronger ad network.
- Google's edge is described as a portfolio of high-beta options, including Waymo, quantum computing and Isomorphic biologics.
- Friedberg flags a low-probability but high-severity China risk to Nvidia after a 1 nanometer process demo and a reported forty-billion-dollar domestic chip investment.
- US export isolation is seen as accelerating a full domestic Chinese chip stack, including DUV and EUV lithography.
- Tesla's valuation is said to already price its robotics optionality, making Friedberg reluctant to pay the premium.
- xAI, Colossus and a possible Tesla and xAI merger are discussed as private-side drivers of Elon's AI position.
Ideas
All AI roads still lead to GPUs
Laffont's number one pick to win the AI race over the next five years is Nvidia, because he does not see the GPU getting displaced. He expects additional compute architectures to come on board, but he thinks they grow the market rather than replace it, since at the end of the day all roads for these models still lead back to the GPU.
Tesla dark horse on vertical integration
His number two is Tesla, framed explicitly as a dark horse: Tesla has the most potential for full vertical integration, owning the stack from the silicon through the model to the physical hardware. He argues that integration may become very important not just in cars but in Optimus humanoid robots.
Tesla owns the vertically integrated physical-AI stack
Chamath ranks Tesla first because it is closest to the vertically integrated stack he favors. Tesla has the best vision models, with xAI it will have one of the best LLM and reasoning models, and he expects those to eventually run on Dojo silicon. That stack then sits inside all of the physical AI people interact with daily: robots, cars and robotaxis.
Google's models, TPUs and distribution win
Asked directly whether Google can win if search declines, Chamath says yes. He boils Google's economic north-star metric down to price per click and argues Google is extremely well positioned to pivot that metric to price per token, because it has the largest pool of users to monetize through YouTube, Gmail, Workspace and a reworked search product. The constraint is willingness to rip the band-aid off, not capability.
Elon wins if Tesla absorbs xAI
Jason's picks are Google or Elon, and he chooses Elon, with Tesla as the listed way to own that bet. After a day at xAI he describes an unmatched magnet for elite talent, and he combines Colossus compute, Tesla's own hardware-plus-software stack with FSD and Optimus, and the real-time data of X. He argues the Tesla and xAI boards should merge the two companies, roughly a trillion-dollar business with a hundred-billion-dollar one, so all the brainpower points in one direction instead of Elon context-switching; if that happens he thinks Elon wins the AI race outright.
Google's ad network improves despite search losses
Jason argues that whether Google loses search share does not matter; what matters is whether its ad network gets more effective. Combining chat queries, Gmail content, browsing behavior in Chrome, Android usage and YouTube consumption gives Google an ad network that performs so much better that ad revenue keeps growing and even accelerates in velocity despite search-share losses.
Tesla's robot optionality already priced in
Friedberg calls Tesla the best place to invest if you want a shot at a massive new industry. On top of the baseline automobile business, he sees the humanoid robot opportunity as mind-blowingly large and believes no company on Earth is better positioned to execute against it, which makes Optimus a low-probability, high-upside call option embedded in the stock. His reservation is price: Tesla already carries a very healthy premium that he thinks has those options priced in, so he is not sure he would pay it, and he ranks it behind Google on a Sharpe-ratio basis.
China tail risk to Nvidia's moat
Friedberg accepts the common view that Nvidia is the most protected business with real durability, but argues there is a low-probability, very high-severity risk to its core coming from China. He cites a demonstration of a 1 nanometer manufacturing process, a reported forty-billion-dollar investment into full-stack domestic semiconductor manufacturing, and new Chinese work on DUV and EUV systems, concluding the lithography IP moat is being crossed and that a DeepSeek-style surprise in semiconductor manufacturing could hit Nvidia's advantage.
Isolation accelerates China's domestic chip stack
Friedberg argues US policy is having the opposite of its intended effect: the more the United States isolates China, the more it emboldens Chinese government and private investment into alternatives to the American chip stack. He points to a reported forty-billion-dollar programme building full-stack domestic semiconductor manufacturing and to new Chinese DUV and EUV lithography development, and expects competitive Chinese semiconductor manufacturing capability to emerge in the near term.
Google is the diversified AI option portfolio
Friedberg's third category is the portfolio solution, and that is Google, which he ultimately ranks number one on alpha and beta-adjusted returns. Inside Google sits a diversified set of high-beta bets, any one of which could be a trillion-dollar outcome: Waymo, quantum computing and the biologics work Demis is doing at Isomorphic, which compensates for risk to the core search business. He also credits Sundar for aggressively re-architecting the search product, and points to research depth well beyond LLMs, including graph-based models and weather forecasting, plus an emerging multi-model, agentic architecture Google should benefit from.
Up Next
This All-In Podcast video, published June 24, 2025, features Thomas Laffont, Chamath Palihapitiya, Jason Calacanis, David Friedberg discussing NVDA, TSLA, GOOG, KSTR. 10 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Thomas Laffont, Chamath Palihapitiya, Jason Calacanis, David Friedberg · Tickers: NVDA, TSLA, GOOG, KSTR