Inside the moves reshaping Google’s AI leadership and Gemini race

MOUNTAIN VIEW: Google’s AI race is getting a fresh dose of founder pressure, with co-founder Sergey Brin urging key artificial intelligence staff to go all in on the company’s Gemini model as Alphabet looks to close the gap with rivals.

Brin has in recent months pushed senior AI employees to accelerate Gemini’s development, according to two people familiar with his remarks. He also addressed hundreds of employees at a Google DeepMind town hall in April, urging the AI lab to move faster as competitors including Anthropic made gains.

According to a Reuters report, Brin’s intervention comes as Gemini once again faces pressure from rival models developed by Anthropic and OpenAI. While Google’s model briefly overtook competitors last November, more recent releases have put it back in catch-up mode.

Google has also delayed the next flagship version of Gemini by about two months after internal testing showed that its performance continued to lag behind rivals in areas including coding, Reuters reported, citing people familiar with the matter.

The pressure on Gemini comes alongside a major reshuffle at Google DeepMind.

On August 5, Alphabet announced that Demis Hassabis would step aside as head of Google DeepMind and take on a chairman role with fewer management responsibilities. His deputy, Koray Kavukcuoglu, was appointed to lead the division.

At the same time, the two original technical co-leads of Gemini left the company to start a new venture.

The reorganisation also appears to be bringing Google DeepMind closer to the wider Google organisation. During an all-hands meeting on August 6, some teams were told they would move out of DeepMind and into corporate Google, according to two sources cited by Reuters.

The changes represent a further reduction in DeepMind’s autonomy since Google acquired the London-based AI lab in 2014. The company has increasingly sought to use DeepMind’s technology across its products and turn its AI investments into commercial growth.

Seven people familiar with Gemini’s development told Reuters that the latest reorganisation and Alphabet’s commercialisation strategy are closely tied to its ambition to achieve leadership in AI models.

The shift has also strengthened Kavukcuoglu’s influence over Gemini. In 2025, Google CEO Sundar Pichai appointed him as chief AI architect in addition to his role as DeepMind’s chief technology officer, with a mandate to integrate Gemini across Google’s products.

Four people familiar with the developments told Reuters that some DeepMind leaders subsequently saw their influence diminish as Kavukcuoglu became increasingly central to Gemini’s direction.

Under Hassabis, Google DeepMind had pursued a broad portfolio of AI research while developing the Gemini family of models. However, Reuters has previously reported that Hassabis resisted some initiatives that could have generated revenue for Alphabet or improved its position in the AI race.

With Gemini at the centre of Google’s AI ambitions, the latest leadership shake-up suggests Alphabet is increasingly focused on speed, commercialisation and closing the performance gap with its fastest-moving rivals.