Google DeepMind launches AI for the Planet accelerator for Asia-Pacific startups

Asia-Pacific sits at the collision point of technological possibility and environmental crisis. The region absorbs more natural disasters than anywhere else on Earth, and the numbers are only getting worse. The United Nations estimates that by mid-century, climate-driven catastrophes could cost the region nearly a trillion dollars annually—roughly three percent of its total economic output—if warming reaches two degrees Celsius. Green technology is spreading across the region's startup ecosystem, but the pace of adoption and scaling remains far too slow to match the accelerating threat.

Google DeepMind, the AI research division of Google, is betting that the gap between innovation and deployment can be closed with the right combination of resources, expertise, and structure. The company announced the inaugural AI for the Planet accelerator, a three-month intensive program designed to identify and support between 10 and 15 startups, research teams, and nonprofit organizations headquartered across Asia-Pacific. The selection criteria are specific: applicants must be working directly on environmental challenges using frontier artificial intelligence as their core technology.

The program itself is built around practical support. Selected teams gain access to Google's full AI stack, including its most advanced frontier models. They receive dedicated mentoring from Google's own teams and external AI specialists. There is a week-long in-person bootcamp followed by three months of virtual support. The company offers potential cloud computing credits and free access to Cloud TPUs—the specialized processors that power large-scale AI work. At the end of the program, teams present their work at a Demo Day designed to connect them with investors and industry partners. Participants also join an alumni network that already includes more than 2,000 startups and nonprofits.

The focus areas are deliberately broad but strategically chosen. Google is looking for solutions in nature protection, climate adaptation, agriculture, sustainability, and energy. Within those categories, the company has flagged particular interest in AI-enabled nature protection, sustainable agriculture practices, and forest conservation. The specificity matters: these are areas where AI can plausibly move the needle on outcomes that matter at scale.

The bar for entry is high. Applicants must arrive with a functional prototype or minimum viable product that has already undergone some form of early validation. They need an in-house technical team with genuine depth in AI and machine learning—not consultants, but people who understand the technology at a fundamental level. Most critically, AI must be the core driver of the solution, not an afterthought or a feature bolted onto an existing product. Teams must also demonstrate the ability to integrate Google's specific models: Gemini, Gemma, AlphaEarth, Forestry, or Perch. At least two or three key leaders—founders or C-suite executives—must commit to active participation throughout the program.

The application deadline is July 26, 2026. For teams working on environmental challenges in Asia-Pacific, the window is narrow but real. The accelerator represents a significant bet by one of the world's largest AI companies that the region's most pressing problems can be addressed, at least in part, by putting advanced AI tools into the hands of teams that understand both the technology and the local context. Whether that bet pays off will depend on the quality of the teams that apply and their ability to move from prototype to impact.