Google DeepMind is launching its first Accelerator program in Asia Pacific, and it’s pointed straight at environmental risk. According to Google DeepMind, the inaugural “AI for the Planet” track is a three-month sprint built for startups, research teams, and nonprofits using frontier AI to tackle problems in nature, climate, agriculture, and energy. The program kicks off with an in-person bootcamp in Singapore.
The timing isn’t accidental. Google DeepMind frames APAC as a global engine of economic growth that’s also unusually exposed to climate change, and points to a recent report showing green tech isn’t scaling fast enough to match rising environmental risks. This is Google DeepMind’s attempt to put its model stack directly into the hands of teams already working the problem.
What participants get
- Expert mentorship from Google AI specialists. Selected organizations work directly with Google’s AI experts over the three months, not just attend talks.
- Tailored support for their specific project. The program is built around the org’s actual product or research, not a generic curriculum.
- Help integrating frontier AI and science AI models. This is the headline benefit. Teams get hands-on assistance plugging Google’s frontier and science-focused models into their workflows.
- In-person bootcamp in Singapore. The program opens with a physical kickoff before moving into the longer mentorship phase.
Who it’s for
- Startups building climate, energy, or agriculture products that need model muscle to scale.
- Research teams working on nature, climate science, or environmental modeling.
- Nonprofits running on-the-ground programs that could benefit from frontier AI capabilities.
The common thread: you’re working on environmental solutions in the Asia Pacific region and you need access to better AI tooling to grow your impact.
Why this matters
What stands out here is the distribution play. Frontier models from Google DeepMind, including its science-focused systems, have largely lived inside research papers and enterprise pilots. An accelerator pointed at climate orgs in APAC is a direct route to deploying those models against problems where the science actually moves the needle: weather prediction, crop yields, energy grid optimization, biodiversity monitoring.
It also signals where Google DeepMind sees applied AI heading. The big labs have spent the last two years racing on general capability. Vertical programs like this suggest the next phase is matching specific model strengths to specific real-world domains, with the lab acting as a partner rather than just a model vendor.
For APAC climate teams, the practical read is simple. If you’ve been bottlenecked on AI talent or model access, this is a chance to skip the queue and get senior Google AI engineers helping you ship.
How to apply
Google DeepMind is taking expressions of interest now. The program runs three months and starts with the Singapore bootcamp. Full eligibility and selection details are on the original Google DeepMind announcement page.
Watch this one closely. If the first cohort produces shippable climate tools, expect similar AI for the Planet tracks to roll out in other regions next.