The potential deal is worth over $1.5 billion, though the talks are ongoing and the details could change, according to the report.
Google is pursuing the deal to boost the coding abilities of its AI models, per the report.
Neither Google nor Mechanize immediately replied to PYMNTS’ request for comment.
Mechanize builds environments and evaluations for frontier coding agents, enables models to carry out software engineering work in the environments, and scores the model’s performance to support reinforcement learning and evaluations, according to its website.
Mechanize announced its launch in an April 2025 press release and described the company as “a startup focused on developing virtual work environments, benchmarks and training data that will enable the full automation of the economy.”
In April, Mechanize said in a press release that it raised $9.1 million at a $500 million post-money valuation.
Mechanize said at the time in a post on X that it aims for full automation of the economy.
“We will achieve this by creating simulated environments and evaluations that capture the full scope of what people do at their jobs,” the company said in the post. “This includes using a computer, completing long-horizon tasks that lack clear criteria for success, coordinating with others, and reprioritizing in the face of obstacles and interruptions.”
It was reported in July 2025 that Google successfully recruited several top executives and researchers from AI code generation startup Windsurf. Unlike a traditional acquisition, Google’s agreement with Windsurf involved paying $2.4 billion in licensing fees to utilize certain Windsurf technologies under a non-exclusive arrangement.
AI coding assistants enable smaller teams to do tasks that would cost a lot more and require additional engineers, and they cut development time for businesses to bring products to market faster, PYMNTS reported in June 2025.
It was reported in May that OpenAI‘s coding agent, Codex, and enterprise AI are becoming more important growth engines for OpenAI.
