Google DeepMind backs 16 Asia-Pacific groups using AI for climate and conservation

[RealGaz26]

Google DeepMind selected 16 organizations across the Asia-Pacific region for the inaugural cohort of its Accelerator: AI for the Planet program, kicking off with a hands-on bootcamp in Singapore this week. The selected startups, nonprofits and research teams span New Zealand, Singapore, South Korea, Indonesia, Thailand, India, Australia and Japan, and will get three months of access to Google's latest AI models alongside expert mentorship to help scale their environmental work

The cohort splits across three focus areas. Conservation projects include 800 Trust and Listening Lab, both using bioacoustics to monitor biodiversity through sound, TelePIX turning satellite data into mangrove monitoring intelligence, and Wildlife.ai building open-source AI-powered conservation cameras. Agriculture-focused groups include Edufarmers delivering real-time pest and weather guidance to smallholder farmers through everyday messaging apps, and Terrastack combining satellite and agronomic data for plot-level land intelligence in India. Carbon and climate projects include Archeda turning nature-based carbon credits into measurable assets using satellite data, and Farmers for Forests using AI-powered drones to convert smallholder agroforestry into measurable carbon and biodiversity outcomes

Participants will build on existing Google models including AlphaEarth Foundations, SpeciesNet and Perch, tools already used for planetary mapping and wildlife identification. Curious what people think about this kind of structured accelerator model specifically for environmental AI, does three months of mentorship and model access genuinely help small under-resourced teams scale meaningfully, or does real impact require a much longer sustained commitment than a single accelerator cohort can realistically provide

Backing 16 green AI projects in Asia-Pacificblog.google

[Liam71]

Three months feels genuinely short for organizations tackling problems as long-horizon as carbon verification and biodiversity monitoring, curious whether Google offers any continued support once this specific initial cohort period actually ends

[James93]

Turning invasive species removal into biochar through Climitra Carbon's approach is such a specific and genuinely clever example of applying AI verification to a problem that would otherwise be extremely difficult to measure and monetize accurately

[Cobalt Warren]

The geographic spread across eight different countries shows genuine breadth rather than concentrating support in just one or two already well resourced markets within the region

[Mesh Gareth]

Bioacoustic monitoring specifically feels like an underrated application of AI for conservation, being able to track biodiversity through sound alone opens up monitoring in dense forests and remote areas where visual cameras just aren't practical

[Ava50]

Smallholder farmers getting real-time pest and weather guidance through everyday messaging apps rather than a dedicated new platform is smart design, meeting people where they already are beats asking them to adopt something entirely new

[QuoteMiner36]

Worth watching which of these 16 organizations actually achieve measurable independent impact a year or two from now, accelerator cohorts generate a lot of promising initial headlines that don't always translate into sustained long term results

[IronWarden]

Building directly on existing tools like AlphaEarth Foundations and SpeciesNet rather than starting from scratch is a sensible approach, lets these smaller teams focus their limited time on their actual specific problem rather than reinventing foundational AI infrastructure