Low employee morale at Google DeepMind is reportedly contributing to delays in the company’s artificial intelligence models as it faces growing competition from OpenAI, Anthropic and Meta. Axios reported that six current and former DeepMind employees linked the slower release schedule partly to burnout, internal leadership issues and concerns over Google’s military work. Google rejected the suggestion that morale problems are affecting its model development.

Low employee morale at Google DeepMind is reportedly contributing to delays in the company’s artificial intelligence models as it faces growing competition from OpenAI, Anthropic and Meta.

Axios reported that six current and former DeepMind employees linked the slower release schedule partly to burnout, internal leadership issues and concerns over Google’s military work. Google rejected the suggestion that morale problems are affecting its model development.

Gemini 3.5 Pro, Google’s most powerful model still under development, is reportedly several months behind schedule.

Google said this week that it is testing Gemini 3.5 Pro with selected partners and plans to release it more widely when it is ready. The company has not provided a firm launch date.

Google instead released three smaller and more efficient models: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber.

The company designed the models to offer lower costs, faster responses and better efficiency for developers and businesses. However, the launches received mixed reactions.

Meta executive Alexandr Wang responded to one comparison involving Gemini 3.6 Flash by writing, “Gemini who?” on X.

gemini who? 🏎️💨 https://t.co/5ANVqtBDkE

— Alexandr Wang (@alexandr_wang) July 21, 2026

The reported model delays come as Google spends heavily on AI infrastructure.

Alphabet reported strong second-quarter results on Wednesday, with total revenue rising 24% to $119.8 billion. Google Cloud revenue increased 82% to $24.8 billion, driven partly by demand for AI infrastructure and services.

Google Search and other related revenue increased 17% to $63.3 billion. However, the result came in slightly below Wall Street expectations, adding to concerns over whether Google’s large AI investments will deliver sufficient returns.

Alphabet’s free cash flow fell to negative $5.9 billion as spending on data centres, chips and other AI infrastructure increased.

The company now expects capital expenditure of between $195 billion and $205 billion in 2026, above its earlier forecast of between $180 billion and $190 billion.

The competition for AI talent has also affected Google.

Noam Shazeer, a Gemini co-lead and one of the industry’s best-known AI researchers, left Google to join OpenAI in June. Google had brought Shazeer back from Character.AI less than two years earlier.

John Jumper, a Nobel Prize-winning scientist and one of the creators of AlphaFold, also left Google DeepMind to join Anthropic.

One DeepMind employee told Axios that Google was “behind” its competitors in model capabilities.

The employee also said Google had focused heavily on protecting its search business from ChatGPT and had not placed enough early emphasis on AI coding agents.

Several former employees have said they resigned because of Google’s work with the US military.

Google reached an agreement in April allowing the Pentagon to use its technology for classified work. The deal reportedly appears frequently in employee exit interviews and has created tension among workers who oppose military uses of AI.

More than 600 Google employees previously signed a letter asking CEO Sundar Pichai to reject classified AI work with the Pentagon.

One source described opposition to the military agreement as a “constant battle” that had caused emotional burnout.

Alex Turner, a former DeepMind research scientist who resigned over Google’s military contracts, said DeepMind CEO Demis Hassabis did not maintain the same visible internal presence that OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei appeared to have with their employees.

Google disputed that employee morale is delaying its AI models or that workers are leaving in large numbers because of the Pentagon agreement.

The company said attrition among its AI employees was lower during the first half of 2026 than during the same period last year.

Google also said more than 90% of people offered AI positions at the company accept them.

“We’re feeling good about this week’s Flash launches, our roadmap and the incredible demand we’re seeing for our models,” a Google spokesperson told Axios.

Leadership in the AI market can change quickly. Google previously held a strong position before Anthropic advanced with its Mythos and Fable models, while OpenAI is also preparing new systems.

Google’s latest Flash models address growing demand for cheaper and more efficient AI. However, the reported Gemini 3.5 Pro delay, rising infrastructure costs and high-profile staff departures are increasing pressure on the company.

Google may be able to recover from temporarily falling behind on AI rankings. The larger risk is losing key researchers and investor confidence while spending hundreds of billions of dollars on the technology.

Get the latest tech news, telecom insights, and product launches wherever you prefer.

Add ProPakistani to Preferred Sources and see more of our stories in Google Search and Top Stories.