Bloomberg Talks · Sep 24, 2026
With Vinod Khosla, Founder of Khosla Ventures
Vinod Khosla looked at every quantum computing company in the US for 10 years and turned all of them down. This year he made his first investment in one.
Sam Altman told the United Nations the day before that people must stay at the center of decisions about AI, and an anchor described him as open to slowing development for safety. Khosla, an early investor in OpenAI, said calls to slow down were the wrong message and that the bigger danger is AI in the hands of the country's adversaries.
"I think much more dangerous than AI is AI in the hands of our nation state adversaries."
Vinod Khosla, founder of Khosla Ventures, on Bloomberg Talks, backed OpenAI and Commonwealth Fusion in the same year, 2018, and, as the anchors noted, is buying the Seattle Seahawks.
The full interview is covered here so you can skip it.
Here are the 10 arguments that matter.
Key Takeaways
Power shortages slow the deployment of AI, not the development of models
He expects new geothermal supply from Mazama by 2028
Data center operators should subsidize local power rates, because their economics leave room to do it
Protect the people displaced by AI, not their jobs
He wants capital gains taxed at the same rate as ordinary income
The tech industry fumbled by not leading with AI's benefits, such as primary care for every American at one or two dollars a month
Calls from AI leaders to slow down were the wrong message, given the race with China
He turned down every US quantum computing company for 10 years before making his first investment this year
Live sports will not be affected by AI, which is why he calls the Seahawks a hedge
1. AI in Adversaries' Hands
The segment opened with a clip of OpenAI's Sam Altman at the United Nations, saying AI must remain under human control. The anchors added that President Trump had posted that morning that AI would be a big topic with China and that both countries want to "leave it exactly where it is."
One anchor said Khosla had argued as early as 2023 that the US cannot slow down because of China, and asked about his conversations with Altman, who has said he is open to slowing things down for safety.
Khosla declined to discuss Altman and named the risk he sees as larger
Khosla said he would not go into specific conversations and might "diverge a little."
"I think much more dangerous than AI is AI in the hands of our nation state adversaries." — Vinod Khosla
He said AI is far more dangerous with bad intent behind it than on its own.
