CLAI Ventures: 2025 Highlights & Perspectives
As the year comes to a close, we want to thank you for your trust, collaboration, and shared commitment to the intersection of AI, climate technology and energy systems. Our work would not be possible without this remarkable network of innovators, investors, and builders. We wish you a restful holiday season with family and friends, and we look forward to the progress we will collectively drive in the new year. Hope you will enjoy reading our 2025 highlights and reflections on the landscape.
CLAI Ventures: 2025 Portfolio Growth
This year, CLAI Ventures reached a significant milestone in our mission to back the most ambitious founders at the nexus of AI and climate/energy. Our portfolio has now expanded to eight startups, each deploying artificial intelligence to solve a fundamental climate or energy challenge:
Brightband – Advancing AI-native weather forecasting for global resilience.
Powerline – Optimizing grid-scale storage through intelligent digital twins.
Rhizome – Enhancing grid resilience and mitigating wildfire risk for utilities.
Aether – Utilizing AI to streamline operations for solar and roofing installers.
Godela – Developing an AI physics engine to replace traditional, high-latency simulations.
Aigen – Merging AI and robotics for autonomous, herbicide-resistant weeding.
Focal – Applying AI and robotics to the decarbonization of outdoor electrical heating.
Vayuh – Delivering high-fidelity AI intelligence for extreme weather risk.
We look forward to announcing our ninth investment in January - an ambitious venture in the AI-native insurance space, designed to redefine how climate risk is mitigated and insured.
CLAI Ventures Portfolio in the News
We are proud to highlight the recent achievements and media recognition of our portfolio companies:
Brightband: Recognized with an Anthem Award (The Webby Awards Social Impact prize) for “Best Use of Data” in the Sustainability, Environment, and Climate Category. This award celebrates the NNJA-AI dataset, made possible through collaboration with NOAA and NASA.
Rhizome: Formally launched an AI-powered wildfire risk initiative in partnership with National Grid to enhance safety and reliability across their networks.
Aigen: The Element robot made a notable appearance at AWS re:Invent, showcasing the future of physical AI and autonomous agricultural technology.
Vayuh: Announced a strategic partnership with Renew Risk to provide advanced modeling for hailstorm risks, specifically protecting solar farm assets.
Focal: Deployed at 12 Bay Area restaurants including Pacific Catch’s flagship Chestnut Street location, demonstrating the scalability of their AI-driven outdoor heating solutions.
2025 Year in Review
The year 2025 appears to mark a significant inflection point for the AI × Climate/Energy sector. Concepts that once resided at the periphery of research labs have transitioned into substantive, real-world deployment across crucial domains, including weather forecasting, insurance, grid operations, disaster resilience, mining, and data center infrastructure. The year was characterized by two central themes:
AI integrating as a foundational operating layer for climate resilience and energy systems (“AI for Climate & Energy”).
Energy and infrastructure undergoing rapid re-architecting to accommodate AI at scale (“Energy for AI”).
Both sides of this equation experienced measurable acceleration in 2025, and both are poised to define the coming decade of climate-aligned technological transformation.
I. AI for Climate & Energy
1. AI for Weather: From Research to Operational Infrastructure
AI-powered weather modeling attained a historic milestone in 2025. With the launch of Google DeepMind’s WeatherNext2, AI-based forecasting has moved beyond the experimental phase to become an operational, scalable, and globally accessible capability. These models utilize physics-informed neural networks trained on extensive climate datasets to deliver expedient, regionally precise forecasts. This paradigm shift carries profound implications for agriculture, energy markets, disaster preparedness, insurance, logistics, and defense.
At CLAI, our investment in Brightband reflects this transition. Brightband applies AI to generate hyperlocal, impact-driven weather forecasts specifically tailored for critical decision-making. The market is now converging on the reality that AI forecasts constitute not merely predictive tools but core financial, operational, and public safety infrastructure.
2. AI-Native Insurance and the Reassessment of Risk
One of the most discernible downstream effects of AI-powered climate modeling is the reconfiguration of the insurance industry. Traditional actuarial models are proving insufficient in the face of climate volatility. In 2025, AI-native insurers demonstrated their capability to underwrite risk in areas where incumbents have struggled.
Companies such as Stand Insurance, which secured a $35M Series B, are showing that machine learning can dynamically and profitably price wildfire, flood, and storm risk. These platforms synthesize satellite data, sophisticated weather models, historical loss patterns, and real-time environmental signals to re-establish coverage in previously uninsurable regions.
CLAI’s investment in Vayuh is central to this focus. By providing AI-driven extreme weather risk intelligence, Vayuh allows both insurers and infrastructure owners to operate with a heightened level of precision. The outcome is not only superior underwriting but the reinvigoration of economic activity in climate-stressed regions.
3. Property Hardening & Disaster Recovery: A Maturing Category
The year 2025 also witnessed a significant increase in startups concentrating on property hardening and post-disaster recovery. This segment sits at the nexus of climate adaptation, insurance, and construction technology.
Notable financings this year included Bright Harbor’s $10M seed for disaster recovery and property resilience and Frontline Fire Defense raising a $48M Series A for wildfire mitigation.
These firms employ AI to assess structure-level risk, streamline retrofitting decisions, deploy early warning systems, and accelerate post-disaster remediation. This represents the evolution of a new vertical we term “Climate Defense Infrastructure.” We are actively exploring several opportunities in this domain, many of which operate in concert with insurance carriers.
4. AI for Sustainable Mining & the Materials Transition
The application of AI for sustainable mining continues to expand, driven by companies seeking to maximize information extraction about deposits from a minimal number of drill holes. The unrelenting demand for materials like copper, essential for the energy transition, propels this sector. Efforts span the spectrum from greenfield exploration to mine planning and intermediate stages.
There is early evidence validating the power of artificial intelligence to deliver verifiable resource discoveries and enhance capital efficiency.
KoBold Metals utilizes advanced AI and machine learning to accurately model and pinpoint critical mineral deposits, such as copper and cobalt. Their strategy is affirmed by major discoveries like the high-grade Mingomba copper deposit in Zambia.
Earth AI leverages proprietary AI-based algorithms coupled with drone-collected geophysical data to identify targets and to direct and facilitate autonomous drilling in Australia.
This direct correlation between cutting-edge data science, reduced environmental impact, and verifiable resource identification is why CLAI Ventures is drawn to the opportunities in this sector and continues to evaluate companies driving the next generation of sustainable mineral supply.
5. Digitizing the Grid: From Analog Infrastructure to AI-Native Networks
Electrical grids globally largely retain a pre-digital foundation, and 2025 further highlighted the stress that legacy grid planning tools impose on the energy transition. A pivotal development this year was the acquisition of Pearl Street Technologies, whose software automates the single largest bottleneck in decarbonization: the interconnection queue.
Pearl Street’s suite of tools, including SUGAR™ and Interconnect®, streamlines complex power-flow modeling and scenario analysis that historically consumed weeks or months for utilities and consultants. Drawing inspiration from chip-design simulation workflows, their platform enables transmission providers, grid operators, and developers to identify viable grid upgrades and assess project feasibility in hours.
At CLAI, our portfolio company Rhizome continues to lead the application of AI to grid resilience and wildfire mitigation. Its 2025 partnership with National Grid across the U.S. and U.K. constituted a major achievement. Rhizome’s gridFIRM platform quantifies long-term wildfire risk associated with utility assets and assists utilities in making cost-effective, preventative investments. As climate risk escalates, gridFIRM directly links AI risk modeling to capital deployment decisions, safeguarding both affordability and reliability.
This category of AI for grid planning, interconnection, resilience, and risk prevention has progressed from conceptual to mission-critical.
6. AI for Distributed Energy Resources: Optimizing the Edge
The substantial proliferation of Distributed Energy Resources (DERs) in solar, wind, and storage necessitates a new stratum of intelligence to maximize value and stability. Managing these volatile, bi-directional energy flows requires dynamic, AI-driven optimization across the grid.
CLAI Ventures views the optimization of utility-scale Battery Energy Storage Systems (BESS) as the most critical point of immediate value realization. Our investment in Powerline directly addresses this. Powerline’s flagship product, Battery Co-Pilot™, is an AI-powered digital twin platform that runs continuous, fleet-wide performance benchmarking and live counterfactual simulations. This system delivers highly precise, real-time intelligence for bidding and dispatch, enabling asset owners to confidently exceed standard operating methods and significantly augment BESS revenue, accelerating the financial feasibility of storage deployment globally.
Beyond storage, this market is increasingly adopting AI to resolve key integration challenges, from utilizing predictive models for solar and wind asset maintenance to leveraging AI for Dynamic Line Ratings (DLR) to safely unlock more capacity from existing transmission infrastructure. We maintain a focus on partnering with founders who are constructing this next generation of energy infrastructure.
II. Energy for AI
If 2025 established that AI can reconfigure climate systems, it also affirmed that energy is now the preeminent constraint on AI itself.
Image credit: “Data Center Infrastructure in the United States, November 2025,” by Billy Roberts, National Renewable Energy Laboratory (NREL). (https://docs.nrel.gov/docs/gen/fy26/98020.jpg)
1. The Emergence of the “AI Factory” and the Efficiency Imperative
AI compute has now reached industrial scale. Data centers are no longer solely IT infrastructure; they are fundamentally energy infrastructure. The $50M Series B secured by Phaidra underscored this transition. Their AI platform now optimizes not only cooling but the entire AI factory compute stack, in collaboration with NVIDIA’s gigawatt-scale reference designs and Omniverse digital twins.
AI efficiency is no longer discretionary as every percentage point of improvement yields:
Faster deployment
Lower capital expenditure
Reduced grid strain
Improved sustainability economics
2. Data Centers as Grid Participants, Not Solely Consumers
The central challenge facing the grid is no longer whether it can support AI, but how adaptable the AI loads can be to cooperate with the existing electrical network.
A seminal 2025 Duke University Nicholas Institute report indicated that a mere 0.25% load curtailment (equivalent to one day per year) by data centers could liberate 76 GW of new U.S. grid capacity. This insight recast data centers from static loads into dynamic grid assets.
Google’s 2025 blog further catalyzed this shift, detailing real-world load shifting across time and geography, moving non-urgent compute during periods of grid stress. As AI workloads expand, demand response is evolving from a utility tool into a core principle of AI infrastructure design.
3. Four Technical Pathways Manifesting in Energy-for-AI Startups
Throughout 2025, we observed four predominant startup vectors:
Identifying spare grid capacity using AI-driven interconnection intelligence.
Peak load shaving by dynamically adjusting compute and clock speeds.
Storage-coupled data centers enabling accelerated grid interconnection.
Colocation with renewables (wind, solar, hydro) to absorb surplus generation.
These approaches address the “speed-to-power” imperative now characterizing hyperscaler competition. However, our view at CLAI is nuanced: not all of these technical solutions naturally translate into durable, venture-scale business models.
Some will be developed internally by hyperscalers. Others may emerge as acquisition-driven platforms rather than independent category leaders. Distinguishing between venture-grade infrastructure software and service businesses is now one of the most crucial underwriting challenges in climate-AI investing. Venture capital must identify genuine product gaps that can scale as independent businesses. We continue to seek such opportunities.
4. Reassessing the Premises Behind Exponential Energy Demand
A final theme of 2025 has been the increasing, healthy skepticism surrounding linear extrapolations of AI energy demand. Many forecasts implicitly assume that LLM scaling alone leads to AGI, and therefore that compute and power must scale exponentially without fundamental architectural change.
In his recent AXIOS talk, Hassabis was forthcoming about the timeline for Artificial General Intelligence (AGI), which he defines as a system exhibiting all human cognitive capabilities, including inventive and creative skills:
Timeline: He estimates AGI to be 5 to 10 years away.
Current Models: He notes that current Large Language Models (LLMs) are “jagged intelligences,” exceptional in some areas (like solving IMO problems) but deficient in others, lacking capabilities such as continual learning, online learning, and long-term planning.
Path to AGI: Hassabis contends that AGI will likely necessitate scaling current systems to the maximum, plus one or two additional major breakthroughs comparable to the invention of the Transformer or AlphaGo.
The market narrative often equates more parameters plus compute with inevitable progress to AGI, and thus uncontrolled energy demand. History suggests that architectural and algorithmic efficiency repeatedly subvert brute-force paths. Innovations in sparsity, retrieval, modular agents, continual learning, and on-device distillation can materially bend energy consumption curves. Edge inference itself can mitigate much of the predicted grid impact. Once models are distilled and widely served, per-query energy diminishes by orders of magnitude. We anticipate robust training demand but do not presuppose proportional inference growth in centralized data centers.
Basing investment solely on raw GW buildouts is a risky proposition if architectures change.
Closing Perspective
2025 made one reality unequivocally clear: AI is now integral to climate and energy systems, and energy is now integral to AI progress.
On the AI for Climate side, advancements in weather, insurance, disaster resilience, mining, DERs, interconnection, and grid planning have transitioned from experimentation into actual infrastructure deployment.
On the Energy for AI side, data center efficiency, grid interconnection, demand response, and hyperscale power logistics have become primary economic constraints shaping the trajectory of artificial intelligence itself.
For CLAI Ventures, this convergence is precisely our operational focus. We maintain the conviction that the next decade will be defined not just by smarter models but by the co-evolution of AI and energy infrastructure.
Thank you for being a part of our journey!
Warm regards,
Ajay Gupta and the CLAI Ventures team
This post was written by Ajay Gupta and the CLAI Ventures team, a Silicon Valley-based fund investing at the intersection of AI and climate. You can reach us at ajay@clai.vc and team@clai.vc.