Pichai pushes back on claims Google is losing ground in AI race

By Deborah Mary Sophia and Kenrick Cai

July 22 (Reuters) - Alphabet CEO Sundar Pichai used Wednesday's earnings call to mount a robust defence of Google's AI strategy, pushing back on concerns that the company has fallen behind rivals after delaying a flagship model and ceding ground in AI coding.

Investors have become increasingly uneasy over Google delaying the release of Gemini 3.5 Pro, a model originally slated for June that was expected to bolster the company's standing in AI coding and autonomous "agent" tasks, two of the industry's most fiercely contested areas.

The latest setback has fueled worries that Google was losing momentum just as OpenAI and Anthropic and a host of Chinese rivals have accelerated the pace of model releases. On the call on Wednesday, Pichai struck an unusually defensive tone as analysts pressed him on the state of Google's frontier models and whether they could still compete at the cutting edge of AI development.

"We've had clearly frontier models. There are many attributes on which we are still at the frontier; there are areas where we've acknowledged we need to improve and coding and agentic coding is an example of that," he said in response to JPMorgan analyst Doug Anmuth, who questioned whether Gemini could remain competitive at the industry's leading edge and noted Google's slower release of models.

Rather than dwell on the Gemini 3.5 Pro delay, Pichai repeatedly highlighted Gemini Flash, Google's cheaper, faster "workhorse" model that powers a growing range of applications, including cybersecurity, customer-service agents, data analytics and enterprise software.

He pointed to Gemini 3.6 Flash, released this week, saying it had improved by more than 10 points on a coding benchmark compared with the previous version while using fewer tokens. The company on Tuesday also unveiled Gemini 3.5 Flash-Lite and a cybersecurity-focused Flash Cyber model, while keeping Gemini 3.5 Pro in partner testing.

"I think people will be pleased" when Google unveils Gemini 4, Pichai said, describing it as a "very ambitious effort." He stressed that Google is training a significantly larger model designed to compete at the frontier when it is released, adding that the company remains "very committed and very confident" about staying at the leading edge of AI.

When Barclays analyst Ross Sandler raised similar concerns about Google's model release pace, Pichai disclosed that Gemini 4's roadmap includes rolling out models "almost at a monthly cadence."

Pichai's vigorous defence underscores a growing challenge facing Google: persuading investors to judge its AI ambitions not by the delay of a single flagship model, but by the scale of an ecosystem spanning cloud infrastructure, custom AI chips and a portfolio of Gemini models embedded across its products and services.

While Google posted another quarter of blistering cloud growth on Wednesday — up 82%, far above an average estimate of 64% — Wall Street is increasingly focused on whether the company can regain leadership in AI coding and frontier reasoning, especially as its capital costs skyrocket.

Alphabet raised its capex plans by $15 billion to a range of $195 billion to $205 billion.

The company's shares fell more than 3% in after-hours trading. Concerns over Gemini's delays, coupled with several high-profile executive departures, have left the stock down about 9% since the end of April.

(Reporting by Deborah Sophia in Bengaluru and Kenrick Cai in San Francisco; Editing by Sayantani Ghosh and Shri Navaratnam)