At Google, leaders are anxious about falling behind in the race to offer AI coding tools, especially as rivals like Anthropic PBC offer more effective and popular tools to businesses, according to people familiar with the matter. The search giant is now working to unite some of its coding initiatives under one banner to speed progress and take advantage of a surge in customer interest.
In some corners of Alphabet Inc’s Google, particularly AI lab DeepMind, concerns about the company’s position are mounting, according to current and former employees and executives, who declined to be named because they weren’t authorised to speak publicly. Businesses are just starting to realise that AI coding tools can enable anyone to build products by prompting a chatbot. But Google doesn’t have a clear solution for them. Its Gemini model’s capabilities are sprinkled across half a dozen different coding products with different branding, indicating how the company’s lack of focus and competing internal efforts have hampered success, the people said.
Even internally, some Google engineers prefer to use Anthropic’s Claude Code, they said. More concerning, the people said, are the engineers who are struggling to adopt AI coding at all.
Google has made some effort to reduce the internal confusion over priorities. Chief AI Architect Koray Kavukcuoglu is working with Google’s main engineering team to unite the company’s internal artificial intelligence coding tools in the coming weeks under Antigravity, a platform released last year, according to a spokesperson.
DeepMind is also devoting more resources to AI coding by forming a new team led by research engineer Sebastian Borgeaud, according to a former Google employee. That new team was earlier reported by The Information. John Jumper, who won the Nobel Prize in 2024 alongside Google DeepMind Chief Executive Officer Demis Hassabis, is also at work on AI coding, according to a person familiar with the matter.
Google was widely viewed as ascendant in AI late last year with the release of Gemini 3, a model that appeared to outperform rival services across a range of benchmarks. In recent months, however, Anthropic and OpenAI have gained business momentum by focusing on the lucrative market for products that streamline the process of writing and debugging code to speed up software development.
"Coding is the single easiest way to actually make money,” said Keith Zhai, co-founder of startup TinyFish, which makes web agents. Many engineers in the valley toggle back and forth between Claude Code and OpenAI’s Codex to see which program will give them the best results, but Google often isn’t in the conversation, he added.
Google still has plenty of reasons to feel confident about its position: the company has made big strides in the quality of its foundation models, which underlie coding tools, and it has deep pockets and substantial computing power.
"We’ve seen tremendous adoption of our internal coding tools such as Antigravity and others since introducing them over recent years, and their use has been turbocharging our model and AI tooling development,” a Google spokesperson said in a statement. Meanwhile, Google has been eager to tout the speed of its internal culture change. Alphabet said in February that roughly 50% of new code at the company is written by AI.
But Silicon Valley engineers are embracing AI coding so quickly that even a momentary lag in the market could be consequential. There is a growing conviction in the industry that coding is not just a lucrative early application of AI, but the key to building software that matches human capabilities, said Raj Gajwani, a former Google executive who is now chief business officer of startup OpenArt AI. "From a computer science point of view, if you win at coding this year, you get the raw data you need to win at model capability next year,” he said.
Google’s emphasis on its own technology has also complicated the push to catch up. Most employees are banned from using competing tools such as Claude Code or Codex due to security concerns, but Googlers can request exceptions if they can demonstrate they have a business case, one former employee said. Some teams at DeepMind, including those working on the Gemini model, internal applications, and open source models, use Claude Code, according to three former employees.
"You want the best people to use the best tool, even inside Google,” one of the former employees said.
Anthropic cut off OpenAI’s access to its models last year, Wired magazine reported. Google has invested billions of dollars in Anthropic. A spokesperson for Anthropic did not immediately reply to a request for comment.
In recent years, DeepMind has tried to tighten control over how its AI breakthroughs are woven into Google products. Last year, Google appointed Kavukcuoglu to a new position as chief AI architect, a role in which he is charged with folding generative AI into Google products. Yet confusion about who is leading the charge on AI coding persists. Along with DeepMind, Google Cloud, Google Core, Google Labs and Android are all pushing AI coding in different ways, one of the people said.
Google released its Antigravity platform last year following the acquisition of talent and technology from startup Windsurf in a US$2.4bil (RM9.5bil) deal. It joined a cluttered lineup of Google AI coding tools that includes Gemini Code Assist, Gemini CLI, AI Studio, Firebase Studio and Jules. Kathy Korevec, who oversaw Jules, jumped from Google to OpenAI earlier this month, according to her LinkedIn profile.
In a post on social network X, Korevec wrote that Google had an opportunity to build AI developer tools that "feel cohesive, intuitive, and truly great to use. What I saw more often was fragmentation. Parallel tools. Overlapping surfaces. Smart teams solving similar problems in slightly different ways. That’s not a talent problem. It’s a systems problem.” Korevec didn’t immediately reply to a request for comment.
Within the Googleplex, there is a philosophical clash between AI researchers who want to move as quickly as possible and more traditional senior engineers who have exacting standards for code quality, former employees say. AI usage is factored into performance reviews, according to a former employee. But engineers who try to use internal AI coding tools often hit capacity constraints due to competition for computing power, the former employee said.
One of the executives who oversaw efforts to promote AI coding within Google, Brian Saluzzo, recently departed. Saluzzo did not immediately reply to a request for comment.
Companies are still figuring out how to best incorporate AI into their workflows, and having a variety of products on offer gives Google more chances to see what sticks. But incumbent players like Google have only so much of an edge, said Deepti Srivastava, a former Google executive who is founder and CEO of AI startup Snow Leopard.
"The market is moving too fast for the larger companies to think about it and then move,” Srivastava said. "Speed is your only moat.” – Bloomberg
