Google DeepMind Accelerator: AI for the Planet (APAC) 2026: A Three-Month, Equity-Free Program With Google Cloud Credits, Free Cloud TPUs, and DeepMind Mentorship for Asia Pacific Climate and Nature Startups

Google DeepMind Accelerator: AI for the Planet (APAC) 2026: A Three-Month, Equity-Free Program With Google Cloud Credits, Free Cloud TPUs, and DeepMind Mentorship for Asia Pacific Climate and Nature Startups

This is a historical reference for the Google DeepMind Accelerator: AI for the Planet (APAC) 2026 cycle. Google DeepMind’s official program page lists an application deadline of July 26, 2026, followed by a September bootcamp, virtual support through December, and a December Demo Day. That application deadline has passed, and the official page does not announce a next cycle. Do not read this archived entry as confirmation that applications are currently open.

The program itself was a three-month, equity-free opportunity for APAC-headquartered startups, research teams, and non-profit organizations using artificial intelligence to address environmental challenges. Google DeepMind described access to the Google AI stack, including specialized frontier models, tailored support from Google and industry experts, technical and business training, and an alumni network. Selected projects could also be eligible for Google Cloud credits and free Cloud TPUs after a separate review. No direct cash grant was stated on the official page.

The sections below preserve the published 2026 call and explain what applicants were expected to show. Where the official source does not confirm a detail, this page does not fill the gap with a guess.

Key Details at a Glance

What the Program Offered

The program’s value was access to Google’s people, tools, and technical network rather than a cash award. Google DeepMind said each participant would receive access to the Google AI stack, including specialized frontier models, plus tailored support for three months. The published description names both general models, such as Gemini and Gemma, and specialized models, such as AlphaEarth, Forestry, and Perch, as possible parts of an applicant’s integration plan.

Technical and business support. Participants were to receive training tailored to their technical and business challenges, along with invitations to Google-hosted technology bootcamps. The support phase included one-on-one mentoring and deeper technical guidance from Google and industry experts, with a dedicated relationship manager coordinating the work.

Technical project partnership. The program was intended to be collaborative. Google DeepMind said teams would work directly with Google experts on significant technical challenges and engage with Google DeepMind and Google Research models while developing their solutions. That is more specific than a general lecture series: applicants needed to identify a concrete project problem that could benefit from this partnership.

Product credits and compute. Selected startups or projects might qualify for credits through Google for Startups Cloud or Google Cloud for research, and for free Cloud TPUs. The official page makes clear that these resources are subject to eligibility review and approval. They should therefore be treated as possible support, not as a guaranteed grant or an entitlement for every admitted team.

Equity-free participation. Google DeepMind described the support as equity-free for the duration of the program. The page did not state a direct cash award. Graduating teams would also join the Google Accelerator Alumni Network, which the official page describes as a global community of more than 2,000 startups and non-profits.

Exposure. The in-person Demo Day was designed to give teams an opportunity to present their AI-driven impact to Google teams, mentors, investors, partners, and other global stakeholders. The page presents this as an opportunity to showcase work, not as a promise of investment or partnership.

Who Should Apply

This program was aimed at organizations where AI was central to solving an environmental problem, not where AI was only a minor feature. Google DeepMind listed three applicant types: startups, research teams, and non-profit organizations. All needed to be headquartered in Asia Pacific.

The broad areas were nature, climate, agriculture, sustainability, and energy. Google DeepMind’s accompanying focus-area document gave special attention to AI-enabled nature protection, AI-enabled sustainable agriculture, and AI to protect against deforestation, while noting that projects outside those three examples were still encouraged to apply. The document describes work such as improving the use of nature data, supporting smallholder farmers with sustainable practices, and applying satellite or acoustic data to reduce deforestation blind spots.

You are a good fit if:

• Your organization is headquartered in the Asia Pacific region.
• You have a functional prototype or MVP, with early validation and proven traction.
• AI is the core driver of your current solution, or it is central to your future technical roadmap.
• You can show how the project could use Google AI, including Gemini, Gemma, AlphaEarth, Forestry, or Perch.
• You have an established in-house technical team with deep AI and machine-learning expertise.
• At least two or three founders, CxOs, or primary technical decision-makers can participate actively across the three months, including the in-person bootcamp and Demo Day.
• Your work maps clearly to critical challenges in nature, climate, agriculture, sustainability, or energy.

You were probably not the right fit if you were pre-prototype, if AI was a label rather than the technical engine, if your team could not attend the in-person components, or if your organization was headquartered outside APAC. Google DeepMind runs accelerator programs in other regions and themes, so organizations outside APAC would need to consult the broader accelerator listing rather than this call.

Eligibility Requirements in Detail

Based on Google DeepMind’s published criteria, applicants should be able to answer yes to each of the following:

1. Location. Was your startup, research team, or non-profit headquartered in Asia Pacific? The call required an APAC headquarters.
2. Stage and evidence. Did you have a functional prototype or MVP, with early validation and proven traction? The published criteria were aimed at working projects rather than ideas on paper.
3. AI centrality. Was AI the core driver of your current solution, or clearly central to your future technical roadmap? Machine learning could not be an incidental feature.
4. Google AI plan. Could you explain a credible way to integrate Google AI? The official examples included Gemini, Gemma, AlphaEarth, Forestry, and Perch.
5. Technical team. Could you show an established in-house team with deep expertise in AI and machine learning?
6. Mission. Were you addressing a critical environmental challenge in nature, climate, agriculture, sustainability, or energy?
7. Commitment. Could two or three founders, CxOs, or primary technical decision-makers participate actively throughout the program?

Some benefits — specifically Google Cloud credits and free Cloud TPUs — were described as subject to eligibility review and approval. Meeting the admission criteria did not automatically guarantee those resources; the official page treated them as a separate eligibility decision.

Program Timeline and What Each Phase Involves

The published 2026 cohort was compressed into roughly three months, with this schedule:

• Application deadline: July 26, 2026. The official page lists this as the deadline. It has passed, and the page does not publish a replacement date or next cycle.
• In-person bootcamp: September 7–11, 2026. The week-long bootcamp was scheduled to open the program, with Google mentors, keynotes, workshops, and diagnostic sessions. The in-person requirement made leadership availability and travel planning important.
• Virtual tailored support: September–December 2026. After the bootcamp, participants were to receive three months of virtual support, one-on-one mentoring, and deep technical guidance coordinated by a relationship manager.
• In-person Demo Day: December 2026. The program was scheduled to culminate in a showcase for industry partners, investors, global stakeholders, mentors, and Google teams.

Google DeepMind planned to select 10 to 15 organizations, making this a small and selective cohort. That size suggested a high bar for evidence, technical readiness, fit with the environmental mission, and the team’s ability to use the support during the scheduled program.

How to Apply

The 2026 application was submitted online through https://goo.gle/GDM-Accelerator-APAC-Apply, linked from the official program page at https://deepmind.google/accelerators/ai-for-the-planet/. Google DeepMind also advertised weekly virtual open forums during the application phase for questions from prospective applicants. The deadline has passed, and no next application window is announced on the official page, so this route should be treated as the historical submission path rather than an open invitation.

The public program page does not publish every form field. The published criteria show that a complete application needed to address:

• Organization and location: legal name, APAC headquarters, organization type, and stage.
• Environmental problem and solution: the specific nature, climate, agriculture, sustainability, or energy challenge and how the product or research addresses it.
• AI system: how AI drives the current solution or future roadmap, including the data, models, evaluation, and operating context that make the claim credible.
• Working evidence: the functional prototype or MVP, early validation, proven traction, and any measurable results or usage.
• Google AI integration plan: a practical account of how the project could use a Google model or tool, such as Gemini, Gemma, AlphaEarth, Forestry, or Perch.
• Technical team: evidence that the organization has an in-house team with deep AI and machine-learning capability.
• Leadership commitment: the two or three founders, CxOs, or primary technical decision-makers who would participate and attend the in-person components.
• Use of the program: the technical, product, or business problems that three months of expert support could address.

With only 10–15 places, the application needed to make fit legible quickly. A clear link between environmental outcome, working system, Google AI integration, technical team, and leadership availability would have been more useful to reviewers than a broad statement of ambition.

How to Prepare a Competitive Application

Lead with the environmental outcome, then the AI. Reviewers for a program called “AI for the Planet” were looking for work connected to nature, climate, agriculture, sustainability, or energy. State the real-world result first, then explain why AI is central to achieving it.

Make the AI concrete. Because the program required AI to be the core driver, vague claims would weaken an application. Name the technical approach, the data, the evaluation method, and the limits of the current system. Then connect the proposed work to a specific Google AI integration plan.

Show that the prototype is real. A functional prototype or MVP, early validation, and proven traction were stated requirements. Point to something a reviewer could see working: a demo, pilot results, early users, field trials, or benchmark numbers. Static claims without evidence would not establish readiness.

Be specific about what mentorship and compute would change. A strong application would name a concrete bottleneck — a model that does not generalize, a data pipeline that cannot scale, or a deployment that needs more compute — and explain how three months of Google support and possible TPU access would address it. Cloud resources were subject to approval, so the plan should not depend on an entitlement that the official page did not promise.

Confirm your team can commit. The program expected two or three founders, CxOs, or primary technical decision-makers to participate actively, including at the in-person components in September and December 2026. That commitment needed to be settled before submission.

Use the open forums. Google DeepMind offered weekly virtual open forums during the application window. Attending one could help an applicant resolve questions before submission, but the forums were part of the closed 2026 application phase.

Common Mistakes to Avoid

• Applying without a working prototype. The MVP requirement was explicit, alongside early validation and proven traction. Concept-stage projects were outside the stated readiness level.
• Treating AI as a label. If a reviewer could not see AI as the technical engine of the work, the project would miss a core criterion.
• Ignoring Google AI integration. The call asked for a credible plan to use Google AI, not merely a general interest in machine learning.
• Ignoring the APAC requirement. Organizations headquartered outside Asia Pacific were outside this specific call and would need to review other Google DeepMind programs.
• Omitting the in-house technical team. The criteria called for established internal AI and machine-learning expertise.
• Underestimating the time commitment. The bootcamp, virtual support phase, and Demo Day required active participation from senior leaders over three months.
• Assuming cloud credits and TPUs were automatic. They were subject to eligibility review and approval, so applicants could not treat them as guaranteed funding.
• Using vague impact claims. A statement about helping the planet needed concrete ties to nature, climate, agriculture, sustainability, or energy outcomes.

Frequently Asked Questions

Does the program give cash? The official page did not state a direct cash grant. It described equity-free support, access to the Google AI stack, technical and business training, mentoring, technical partnership, Demo Day exposure, and possible Google Cloud credits and free Cloud TPUs subject to approval.

Does Google DeepMind take equity? No. Support is explicitly equity-free, so you retain full ownership of your organization and intellectual property.

How many organizations are accepted? Google DeepMind planned to select 10 to 15 organizations for the 2026 cohort.

Who could apply? APAC-headquartered startups, research teams, and non-profits with a functional prototype or MVP, early validation, and proven traction, where AI was central to work on nature, climate, agriculture, sustainability, or energy. The call also required a Google AI integration plan, an established in-house AI/ML team, and committed leadership.

When was the deadline? The official page lists July 26, 2026. That date has passed, and no next cycle is announced there.

Is any part in person? Yes. The bootcamp (September 7–11, 2026) and the Demo Day (December 2026) are in person; the support phase between them is virtual.

Where was the application submitted? The 2026 call used https://goo.gle/GDM-Accelerator-APAC-Apply, linked from the official program page. The link is retained as a historical reference; it is not evidence of a current open window.

Is It Worth Applying?

For an APAC-based team where AI genuinely drives an environmental solution and there is already a working prototype, the 2026 call offered a strong technical fit. Its equity-free structure avoided an ownership exchange, while the Google AI stack, expert mentoring, technical partnership, and possible compute support addressed practical needs for applied machine-learning teams. The small cohort, Demo Day, and alumni network added useful exposure, but none of those benefits should be read as a promise of investment or admission.

The main costs were time and focus: a three-month commitment with in-person components and a competitive application for only 10–15 places. For future applicants, the most important readiness signals would be a working product or research system, evidence of traction, an internal technical team, a specific Google AI plan, and leaders who can participate throughout the program. Since Google DeepMind has not announced a next round on the official page, readers should monitor the official accelerator listing rather than rely on this archived deadline.

Official Links and Next Steps

• Program page: https://deepmind.google/accelerators/ai-for-the-planet/
• Apply: https://goo.gle/GDM-Accelerator-APAC-Apply
• All Google DeepMind Accelerator programs: https://deepmind.google/accelerators/

The 2026 submission window is over. For a future cycle, first check the official program page for a new deadline and application link; then confirm APAC headquarters, environmental focus, prototype or MVP readiness, early validation and traction, AI centrality, Google AI integration, in-house technical capability, and leadership availability. Review the official focus-area document and use any advertised open forum for questions. Finally, verify the benefits and dates on the official page before submitting, because a later cycle may change its scope or schedule.