At Arkam, we look closely at markets where deep technology, infrastructure and long-term human needs intersect. Climate tech is one such space, and the rise of AI creates a new lens through which founders and investors can rethink how physical innovation is discovered, tested and scaled. In this article, originally published by Bhumika Krishnani on LinkedIn, she explores why AI’s most meaningful impact may come not from another software-first use case, but from helping accelerate climate solutions across energy, industry, agriculture, adaptation and the physical systems that define the path to net zero.
“Why Climate Tech Is Where AI Can Matter Most
What’s Missing from the AI Gold Rush
The past three years in artificial intelligence have brought a promising new wave of innovation, one that is steadily reshaping how consumers and businesses work with technology. Building on successive waves of personal computing, the internet, mobile, and cloud, today’s AI advances are motivating brilliant teams to revisit familiar terrains with fresh possibility: customer support but more conversational, edtech but more personalised, digital dating but better curated, and so on.
These pursuits matter, and many have been solved before or at least earnestly attempted. Yet the sheer talent and energy flowing into them naturally makes you wonder: if AI is this powerful, shouldn’t some of that force be aimed at one of humanity’s true generational challenges?
And yet, when the conversation shifts to climate, the energy subsides. Climate and AI feel oddly out of sync; two transformative forces moving on parallel tracks, rather than intersecting with the urgency or ambition you’d expect.
If you zoom out and look at global climate-tech venture capital funding over the last five years, the arc is unmistakable. The sector peaked in 2021/22, partly fueled by a post-COVID economic environment and a growing realisation that climate innovation couldn’t wait. Since then, funding has dropped by more than 50%. India has held up relatively better, with only a 10–12% decline expected in 2025 versus 2024, but it’s still far from its 2022 peak of $1.6B.
(Data Reference: Global venture funding for Climate Tech and AI)
This isn’t to say we need a return to the frenzy of those years. John Doerr actually offers a more grounded benchmark: roughly $50B a year in global venture investment to make climate-tech solutions viable for real-world adoption. Yet even against that calculated OKR, climate-tech funding has fallen short for the past two years.
On the contrary, AI has pulled in an overwhelming share of investor focus. Most analyses indicate that the climate-tech funding dip stems not from waning interest in climate, but from capital migrating toward AI’s explosive momentum. And although AI and climate tech intersect, that overlap is still modest, leaving the broader climate-tech market in a more sober phase.
Rewriting the Playbook for Climate AI
AI has grown up inside a software-first world. We went from “software is eating the world” to “AI is eating software” and with that came SaaS-style expectations around Q2T3 growth, rapid shipping cycles, and near-instant scalability.
Climate-tech sits in a very different reality. Yes, software plays a crucial role in modelling, reporting, and forecasting, and AI will undoubtedly accelerate all of that. Yet, the biggest and most urgent problems remain fundamentally physical, in energy systems, industrial processes, land use, and adaptation.
This means AI’s impact in climate will be judged on a completely different axis, and capital will need to choose between two competing modes:
Software-native AI that scales fast.
Physical-AI applications that require time, hardware, and real-world deployment to create meaningful decarbonization.
This creates a moment of reflection for both investors and founders. AI may not appear “front and center” in most climate solutions, but it has the potential to become a horizontal enablement layer that accelerates physical innovation. Battery systems, for example, can now use AI-driven synthetic ageing to cut validation timelines in half, reshaping development speed without changing the underlying physics.
This shift also calls for a different evaluative lens, one that recognizes the value unlocked when climate-tech is built in an AI-first world. Traditional metrics like ‘capex to first product’ give way to concepts such as a ‘compute-to-discovery’ ratio, where compute becomes a lever that compresses experimentation cycles. It reflects a world where early learning, not just early hardware, becomes a meaningful indicator of trajectory.
As this lens shifts, the way physical technologies mature is also changing. The development journey, especially for founders working in deeply physical domains, becomes noticeably derisked. Instead of relying solely on multimillion-dollar lab cycles, much of the uncertainty can be pushed upstream into simulation and model-driven exploration. It becomes increasingly possible to say:
"Traditional companies spend $10M on experiments to find one winning battery material. We spend $500k on GPU compute to find the same one. Our ‘lab’ is a server rack."
The Emergence of a Computational Pathway in Climate Innovation
It may be too narrow to assume that every climate-tech breakthrough will be powered by AI, but the technology is increasingly positioned to play a larger role in the sector, especially given the momentum behind physical AI in recent years. The leading models are now performing at frontier levels, creating the foundation for the kind of step-change progress we’ve already seen in fields like healthcare. There is still significant headroom for AI to deepen, accelerate, and broaden climate innovation in ways we are only beginning to explore.
(Data Reference: Performance benchmarks of frontier AI models)
But this new pathway should not become a limiting factor. Climate progress will always depend on a wide spectrum of technologies, many of which may never involve AI, yet move the needle far more meaningfully toward net-zero goals. What AI can change, however, is the structure around the sector: who feels empowered to enter, who sees a credible route to impact, and how capital evaluates the space.
Climate tech has long been treated as a niche; worthy but peripheral, constrained by narrow pools of talent and capital. The debate around 'tourist' investors or builders often reinforces this framing, suggesting that only hardened specialists belong here.
But a more transformative interpretation is possible: if AI lowers the barrier to experimentation, sharpens feasibility, and compresses the distance between idea and proof, it can democratize the field. It can draw in talent that once discounted the sector because the odds felt unforgiving.
And with that shift, climate tech stops looking like a niche. It begins to resemble any other high-potential sector; one where capital flows not out of obligation or impact mandates but out of genuine conviction in the upside. AI does not replace the diverse technologies required for climate progress; it expands the aperture, enabling far more people and far more capital to engage in solving the hardest problems of our time.
The Climate Opportunity Set in an AI-Intensive World
The climate market map is broad, with opportunities across sector-specific segments such as energy, industrials, and climate-response strategies like adaptation and mitigation, all areas where AI can play an enabling role. There are also horizontal opportunities where AI sits at the center, powering solutions that were previously impossible, prohibitively expensive, or too complex to tackle.
And as we’ve often heard, AI has an energy problem. As models scale and data volumes surge, reducing AI’s own energy footprint is becoming a meaningful segment in its own right; and it would be ironic if a technology with the potential to accelerate climate innovation contributed new pressure to the same system it aims to help address.
(Data Reference: Market Map for Climate Tech)
A few global standouts are already showing what this next wave could look like:
Bindwell: Reshaping pesticide discovery with Foldwell, an AlphaFold-inspired model that makes safer, sustainable crop protection not only achievable but dramatically faster to develop.
Emerald AI: Redefining how data centers interact with the grid, using intelligent workload triage to dial down demand when power is scarce and introduce real flexibility into one of the highest-capex, highest-load components of the digital economy.
What is emerging now is a landscape where AI and climate are poised to intersect far more deeply than they have in the past. As real-world deployments grow and the tools mature, the space opens up for founders who want to work on problems with depth, consequence and scale. It is a rare moment when technological capability and planetary need begin to align, and it is one worth building for.
References
Ctvc (2025): new report: Globalization in climate tech, CTVC by Sightline Climate. Available at: [https://www.ctvc.co/new-report-globalization-in-climate-tech/](https://www.ctvc.co/new-report-globalization-in-climate-tech/)
NVIDIA: Emerald ai orchestrates ai factories to help relieve grid stress. Available at: [https://resources.nvidia.com/en-us-energy-utilities/ai-factories-flexible](https://resources.nvidia.com/en-us-energy-utilities/ai-factories-flexible)
Singh, J. (2025): Teen founders raise $6M to reinvent pesticides using ai - and convince Paul Graham to join in, TechCrunch. Available at: [https://techcrunch.com/2025/11/13/teen-founders-raise-6m-to-reinvent-pesticides-using-ai-and-convince-paul-graham-to-join-in/](https://techcrunch.com/2025/11/13/teen-founders-raise-6m-to-reinvent-pesticides-using-ai-and-convince-paul-graham-to-join-in/)
All trademarks and logos displayed in the market map belong to their respective owners. This mapping is a best-effort synthesis based on publicly available data and may contain inaccuracies. AI tools were used in editing the text and generating visual elements.”
For us, the larger signal is that AI can change how climate-tech companies are built, evaluated and financed. It may not replace the hardware, infrastructure or scientific breakthroughs needed for climate progress, but it can compress discovery cycles, improve feasibility testing and bring more talent and capital into a sector that has often been treated as specialised or difficult to underwrite. The strongest opportunities may emerge where AI becomes an enabling layer for real-world climate innovation, helping founders build solutions that combine technological depth with planetary relevance.
