DeepMind and Google CEO have ‘daily’ conversations amid ‘fierce’ AI competition

alphabet The stock price began 2025 as investors questioned whether Google could catch up with ChatGPT maker OpenAI in the AI ​​race. By the end of the year, the stock had recorded its best performance since 2009.

Google has regained its AI mojo. Much of it came from DeepMind, a British company that Google acquired in 2014 for around £400m.

In a wide-ranging interview on CNBC’s new podcast “The Tech Download,” DeepMind founder and CEO Demis Hassabis called DeepMind the “engine room” for Google’s AI efforts, adding that the changes were made to help the tech giant quickly deploy AI products in a “fiercely competitive environment.”

Hassabis said he speaks with Google CEO Sundar Pichai “every day” and emphasized how closely the two executives are working to innovate quickly.

“All of the AI ​​technology is being done by this group and spread across all of these great products across Google,” Hassabis told The Tech Download, which launched Friday.

“And over the last few years, we’ve been building that backbone, so we’ve built out not just the models… but the entire Google infrastructure so that… we can ship these things incredibly quickly.”

This could be key for Google, as it faces competition from OpenAI this year as well as a number of other players from Amazon to Perplexity to Anthropic.

“It’s a fiercely competitive environment at the moment,” Hassabis said. He added that “many” veterans who have been in the tech industry for “20, 30 years” have said this is “probably the toughest environment they’ve ever seen in tech.”

In 2023, Google made a significant change by merging its Google Brain research division with DeepMind, laying the foundation for success with its flagship AI assistant Gemini. Other key changes also played their part, including the promotion of Josh Woodward to Gemini’s executive leadership role.

When OpenAI announced ChatGPT in November 2022, Google was playing catch-up. Along the way, product mistakes, especially in 2024 regarding AI tools, reinforced the impression in the industry that Google was struggling to compete.

Hassabis said the company’s problem is not inventing technology. Transformers, the key architecture behind large-scale language models, were, after all, created by researchers at Google. Hassabis continued that the company’s problem is “probably” that it has been “a little slow in commercializing and scaling up.”

“That’s what OpenAI and others have done very well,” he added.

“I think in the last couple of years we’ve almost had to go back to our roots as startups and entrepreneurs and work harder, be faster, ship things faster and make very rapid progress in some ways,” Hassabis said.

DeepMind CEO said the company “hit the ground running” with the launch of Gemini 2.5 in March 2025. Google launched Gemini 3 in November, and its speed was praised by tech CEOs and users.

Hassabis said the Gemini model being developed at DeepMind can be shipped very quickly to various Google products, including search.

“Over the past year, it’s become a really smooth process, and I think you’ll see that even more over the next 12 months,” Hassabis said.

“We think of ourselves as kind of that engine room.”

Hassabis added that he and Pichai “talk almost every day about strategic things, where to deploy technology, and what Google needs overall,” highlighting how DeepMind is essential to Google’s broader plans and the pace of innovation the company is aiming for.

Hassabis said his conversations with Pichai led to the possibility of “day-to-day” adjustments to roadmaps and plans, while also taking a long-term view to achieving artificial general intelligence (AI considered human-equivalent intelligence, the industry’s holy grail) “first, quickly and safely.”

As tech giants pour hundreds of billions of dollars into building AI infrastructure and stock prices continue to rise, market participants are debating whether the AI ​​boom is a bubble. At the same time, venture capital money was poured into AI startups, with many companies raising money at high valuations but few products.

Hassabis said some parts of the industry “may be in a bubble,” while others probably aren’t.

“AI will be the most revolutionary technology ever invented,” he said. He compared it to the dot-com bubble of the late ’90s and early 2000s. “After all, the Internet was very important, and there were some companies that were founded during that time that spanned generations,” Hassabis said.

“It’s almost inevitable in a way. Once everyone realizes how revolutionary a particular technology is, there will be an overexcitement. Then there will be a liquidation, and then the real thing will survive and thrive.”

Hassabis said the seed funding round in the private market, worth tens of billions of dollars, is “unsustainable in the long term” when there is “very little yet” in terms of product.

“Whether it goes either way, whether it continues to be as rosy and exponential as it is now, or whether there’s some kind of bubble burst, we have to make sure that we win and that we’re in a good position to take advantage of it either way,” Hassabis said.

“Given Google’s underlying business and how AI fits into that, I think we’re well-positioned to benefit from whatever direction we go in the future.”