AI is no longer a fringe technology—it is rapidly reshaping entire industries, from climate and energy to industrial operations and enterprise software. Yet as competition intensifies and barriers to entry fall, the question facing venture investors is no longer whether to invest in AI, but how to identify companies capable of generating durable, long-term value. In this interview, Jeremy Brown, Principal & Head of Climate at Anthemis, shares his perspective on evaluating AI businesses beyond the hype, drawing on his experience at the intersection of AI, climate, and complex industrial markets, and setting the stage for a broader discussion with fellow investors on where real opportunity—and real risk—now lies in AI venture capital.
In addition, Jeremy Brown will take the stage at 0100 DACH, where he will speak on the panel “Beyond the Hype: Capturing Real Value in AI Venture.” The discussion will bring together leading European investors to examine how the rapidly evolving AI landscape is reshaping venture returns, risk assessment, and long-term value creation. Alongside William Jilltoft, Director of Strategy and Operations at Northzone; Jimmy Fussing Nielsen, Founding and Managing Partner at Heartcore; and Sion Evans, Managing Director at VenCap, Jeremy will contribute a climate- and industry-focused perspective on where AI is delivering genuine, scalable impact—and where investor caution is increasingly warranted.
AI is attracting unprecedented capital and attention. From your vantage point, what are the most reliable indicators that an AI company is creating durable enterprise value, rather than riding a hype cycle?
It will become increasingly important for an AI company to define its core value proposition independent of any single industry. For example, is the platform or product designed for distributed assets, mission-critical operations, or field-based compliance? It’s important to have the operational DNA clear and look at adjacent sectors that demonstrate similar characteristics. There will be more scrutiny on AI-native and AI-powered marketed companies, so this will become increasingly important.
How do you think about AI investments at the infrastructure layer versus verticalized applications, particularly in complex sectors like construction, energy, and climate?
The key question is to what extent is the infrastructure an orchestrator, driving outcomes such that it enhances productivity and cost efficiency for the optimal amount of work? Verticalized applications will be a natural progression as larger industries figure out how they will adopt AI technologies at scale going forward – many infrastructure layers will be enablers of such verticalized applications.
With models and tooling becoming increasingly commoditized, where do you see true defensibility emerging in AI startups today—data, IP, workflow integration, regulatory complexity, or domain expertise?
Defensibility will be one in data & IP as well as an effective distribution strategy and manageable cost structure, which is impacting many AI companies right now. We are hyper aware about this. As such, we launched a not for profit in 2025 out of Cambridge, UK, called CommonAI, that is enabling AI startups to build from shared technical foundational IP protected under a strong IP framework as they build out their “secret sauce”. Think about it as creating ‘Virtual BigCos’ accelerating time to market for AI companies to compete in market faster. We’re also, through the platform, providing greatly reduced GPU costs to tackle the high COGS problem for many startups head on. This, alongside future domain expertise, will drive more defensibility.
From your experience in enterprise software and industrial environments, what are the biggest barriers to real AI adoption, and how should founders design products that get deployed at scale?
A founder really needs to understand his/her Complex, regulated customer ecosystem. Each customer base requires different procurement and compliance processes, which can complicate both organic and inorganic growth efforts over the first few years of product adoption. Also, deployment environments can be complicated given that some sites may be remote of off grid and interfacing with physical infrastructure is a challenge making operational execution extremely important. This elevates the growing need for service-led business models to sustain long-term growth. Customers frequently expect a blend of software and services to support technology implementation. A founder needs to balance the high margins of software with the high-touch trust of services to scale. Additionally, long sales cycles can slow down new product rollouts and market expansion, so resource planning and commercialization must be thought out well. Design flexible business models that can withstand or capitalize on changing market and policy environments.
When it comes to deploying AI in climate and industrial use cases, what are the most significant differences between Europe and the US—in terms of regulation, data access, customer readiness, and speed of adoption?
Both sides of the pond are looking at ways to re-shore manufacturing, and both are dealing with demographic pressures and labor shortages, but these are more acute in Europe. This gives way for immense opportunities in AI-driven industrial automation. Technology uptake has traditionally been stronger in North America, but we hope to see more European-grown technology champions over the next decade to increase speed of adoption. Europe is over-regulated, which stifles adoption and speed to scale. This is what we would like to be loosened over the next year to give European founders more ability to scale in Europe quickly as opposed to moving to the USA.
AI is often framed as an optimization tool in climate-related sectors. Where do you see AI enabling system-level change in the energy transition or industrial decarbonization, and how should investors assess whether that impact is both real and investable?
Surging energy demand, the acceleration of electrification, and the reshoring of manufacturing are driving a massive buildout of infrastructure and technology across the globe. Scale requires more than general SaaS guidelines and rinse-and-repeat playbooks. Facing long sales cycles, specialized use cases, and complex customer structures, companies at this stage need to have a tailored strategy that balances growth with deep customer trust, operational rigor, and durable revenue. Companies that can demonstrate his will pique investors’ appetites.
How do you think rapid AI development—especially in climate and industrial use cases—is reshaping the venture capital model itself, from fund construction to ownership targets and exit expectations?
Companies are staying private for longer – so the venture model is requiring different routes to achieve liquidity beyond a public listing. This includes a growing market for secondaries. The time horizon to exit for many climate and industrial use cases is longer than 10 years, so there are increasingly evergreen structure and longer fund lives (12+ years) to accommodate this reality.
Really sharp take on the defensibility question in AI investing. The CommonAI initiative is clever since it tackles the infra cost problem while building shared IP foundations. I've seen alot of climate tech startups burn capital trying to compete on model training rather than domain expertise. The point about service-led models balancing software margins with deployment trust is spot-on for industrial buyers who need hand-holding more than slick dashboards.