Elon Musk Calls AI 'Highest ELO Battle Ever' As Gavin Baker Warns Nvidia Blackwell Chips Could Shatter Google's Cost Advantage
On Tuesday, Tesla Inc. (NASDAQ:TSLA) CEO Elon Musk warned about an all-out AI hardware war after investor Gavin Baker outlined how Nvidia Corp.'s (NASDAQ:NVDA) next-generation Blackwell chips could reshape the economics of artificial intelligence (AI).
Elon Musk Calls AI A Hardware Speed War On X
The comments surfaced during the Invest Like the Best podcast, where Baker described the escalating struggle between major AI infrastructure players.
Following the discussion, Musk took to X and said, "AI is the highest ELO battle ever. Speed of deployment of hardware, especially robotics, is the [linchpin]."
The Elo rating system, created by Arpad Elo to measure player skill in zero-sum games such as chess, is now widely used across esports, traditional sports and even to rank large language models, with higher scores indicating stronger performance and points shifting based on match outcomes.
Nvidia Blackwell Delays Created Temporary Advantage For Google
Baker said the battle has centered on Alphabet Inc.'s (NASDAQ:GOOG) (NASDAQ:GOOGL) Google and Nvidia, as both companies pushed competing AI chips.
He explained that Nvidia's Blackwell platform faced major delays because the shift from Hopper chips required radical infrastructure changes, including higher power usage, liquid cooling, heavier data center racks and severe heat management challenges.
"It was by far the most complex product transition we've ever gone through in technology," Baker said, adding that large-scale Blackwell deployments only began over the last few months.
Google's Low-Cost AI Strategy Squeezed Entire Market
Baker said Google gained a short-term edge by becoming the lowest-cost producer of AI "tokens," allowing it to aggressively cut prices.
They've been s***ing the "economic oxygen out of the AI ecosystem," Baker said, describing the strategy as rational but potentially temporary.
Blackwell Chips Expected To Flip AI Economics In 2026
Baker said the first major AI models trained on Blackwell chips could arrive in early 2026 and said that Musk's xAI could be the first to deploy them at scale.
He added that newer GB300 systems are designed to be "drop-in compatible," which could make Nvidia-powered AI systems the future lowest-cost option.
Strategic Shift Could Hit Google Stock And AI Market
Baker said once Google loses its cost advantage, it may no longer be able to run AI at large losses, potentially reshaping competitive behavior across the entire industry.
Baker said that if Google is forced to change its strategy, it could face significant pressure as a higher-cost producer, warning that sustaining deeply negative margins would be painful and could begin to weigh on the company's stock.
He added that such a shift would carry major implications for the broader economics of AI, and said the expected launch of Rubin chips could sharply widen the performance and cost gap versus TPUs and other custom AI chips.
Meta Eyes Google's TPUs As AI Hardware Race Intensifies
Last month, it was reported that Meta Platforms, Inc. (NASDAQ:META) is in discussions to buy Google's tensor processing units for its data centers as soon as 2026, with broader deployment planned for 2027.
Following this, Nvidia praised Google's AI progress while highlighting its own dominance, saying in a social media post that it was "delighted" by Google's advancements but still "a generation ahead" of the competition.
Nvidia has climbed 37.74% year-to-date and Benzinga’s Edge Stock Rankings places it in the 97th percentile for Growth and 92nd for Quality, underscoring its strong competitive standing.
Read Next:
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Photo courtesy: Shutterstock