Performance conversations in the renewables industry often start with availability. It shows up in service reports, warranty claims and board presentations as the primary measure of whether an asset is doing its job. But availability, as Gareth Brown sees it, is not the right question.
"It's all about making money," Gareth says plainly. "That's what these assets are all for."
Gareth is CEO and Co-Founder of Clir Renewables, an AI platform built on intelligence from more than 350 GW of wind, solar and battery storage assets. He spent a decade as an independent engineer advising investors, banks and insurers across the UK, India, China, Sri Lanka and North America, repeatedly called in on portfolios that weren't hitting budget. That pattern, consistent across enough assets and enough markets, became the problem Clir was built to solve.
Gareth recently joined Russ Bates on the Clean Energy Edge Podcast (CEEP) to talk about what drives real financial performance on renewable assets and where AI is changing how owners manage risk.
When owners talk about improving performance, the conversation almost always starts with production.
Those are important questions, but they address only one dimension of a financial model that runs across revenue, OpEx and CapEx. "Ultimately these renewable energy assets are financial investments; they're a financial instrument," Gareth says. "The technical component of that is a huge aspect of that risk."
The example that tends to land hardest: Clir has worked with projects running at 97% availability that still lost $80 million in two weeks. The turbines were performing. The problem was the offtake: volume commitments made at times of year the asset couldn't actually deliver against when conditions shifted. "The biggest cause of technical risk on a renewable asset today is the offtake," Gareth says. "It is the ability to link shape and volume of generation to the commitments that you make on your financial project."
He points to Scandinavia to illustrate the exposure. Wind farms across the region commit day-ahead to produce electricity. An ice storm moves in and none of them can deliver. Every operator goes back to the market at once to cover their positions at spot price. These are low-frequency events but a single one can wipe out years of accumulated margin. Texas during Storm Uri followed the same logic: technically sound projects financially exposed because the link between generation profile and contractual commitment hadn't been properly understood.
Getting that link right requires an accurate, current view of what the asset is actually capable of. Preconstruction models don't provide that. Real operating data does.
Demand for clean power is not slowing down. Climate targets, AI data centers and the broader electrification of industry are pulling in the same direction simultaneously. The build rate is accelerating. Turbines now have swept areas measured in hundreds of meters. Solar technology is evolving faster than most operators can track. Battery storage is being deployed with new chemistries on almost every new asset.
In that environment, owners are often pulled toward the next project before the current portfolio is operating at its full potential. Gareth is direct on this: building more and optimizing what's already in the ground are not competing priorities.
"I wouldn't call it an either/or question. I would say doing both is what we have to do as an industry." Every megawatt that underperforms represents financial exposure that compounds each year. The higher the build rate, the larger the fleet of operating assets that needs active management. "We need to be able to kind of chew gum and walk at the same time," he says. "We need to be building faster and faster to meet the challenges of climate change, but we also need to be optimizing faster as well."
Optimizing faster requires tools that scale with the portfolio. A team managing 500 MW five years ago is often now managing multiple gigawatts, with OEM models and technology types that weren't in the market when they started. The in-house expertise to cover that range doesn't exist in the volumes the industry needs.
Most asset owners have SCADA. What it gives them is a data stream: temperatures, status codes, production readings, alarms, etc. What it doesn't give them is context. A turbine that has been stopped could be a scheduled inspection or a serious fault. The SCADA record does not distinguish between the two, and without that distinction the data cannot be trusted to drive decisions.
"AI now allows us to extract all the data out from all these unstructured data sources in a way that we just couldn't do 6 months ago," Gareth says. Every daily report, weekly report, monthly report and quarterly board pack coming off an asset is a data stream. An operator note that says inverter 4 was down for a scheduled fix between 5:30 AM and noon can now be read automatically, cross-referenced with the SCADA record and used to correctly classify the event. "That report coming in can reallocate the SCADA data automatically," Gareth explains.
The enriched data layer produced by that process is what makes everything downstream reliable: energy yield predictions, major component exchange forecasting, budget reconciliation. All of it depends on the data being correctly labeled and carrying the right operational context. Once it is, it doesn't stay inside a performance monitoring tool in isolation. It flows through into the corporate AI systems an owner already runs, whether that's Copilot, Claude or any other internal workflow. "We can integrate this kind of vertical renewable energy AI directly into your company," Gareth says, "so that you can latch on to all the AI workflows in a really quick and efficient way across the entire business."
The renewables industry is young and growing faster than experienced professionals can enter it. Gareth describes this as a structural problem: expertise is hard to hire, people move and service providers change. When that expertise leaves, so does the accumulated understanding of what's happening on the asset.
"The AI platform here is not just improving things to be better, but it's providing protection and resilience for the companies as well."
That resilience is measurable. Clir clients typically see around 2.1% or more annual performance improvement, with returns of 15 to 20 times the cost of the platform. Recommendations are available from day one, before SCADA connections are live, based on the asset's technology type, vintage and the known issue patterns across the broader 350+ GW dataset. As operational reports and SCADA data flow in, the intelligence becomes more specific and the gains compound.
Gareth describes turning off Clir as turning off the lights. The continuous intelligence that has built around what the portfolio is actually doing, and what it needs, stops flowing. For owners managing growing portfolios across multiple technologies, regions and OEM platforms, that intelligence layer is what keeps the operation coherent as it scales.
Clir is an AI platform that takes data off wind, solar and battery storage assets globally and enriches it through AI to provide software solutions to stakeholders across the asset lifecycle. The platform holds what Gareth describes as "the largest operational codified dataset on the planet," deployed across more than 350 GW of assets spanning across the world.
The data runs from granular SCADA through to the operational reports and board packs that hold the real context behind performance numbers. Each stream in isolation is incomplete. SCADA is granular but context-free. Reports carry context but arrive unstructured. Bringing them together at scale is what makes the platform useful on day one for an owner connecting a wind farm in Texas and wanting to know what to pay attention to immediately.
The platform serves owners, investors, insurers and operators with the same underlying intelligence. Every stakeholder is trying to answer the same question: When are these assets going to fail, how much are they going to produce and how much uncertainty is on that estimate?
The owners who can answer those questions with real operating data rather than preconstruction assumptions are the ones managing the full financial risk on their portfolios and not just the availability number.
The next phase of renewable energy is making sure every megawatt already in the ground is producing what it should, and that the people responsible for it can actually tell whether it is or not.
Listen to the full interview on the CEEP podcast.