Backed by 350+ GW of Operational Data and Purpose-Built Renewable Energy Models, Clir's MCP Server Connects Portfolio Intelligence Directly Into AI Agents and Platforms Asset & Investment Managers Already Use
Vancouver, BC – June 11, 2026 — When a portfolio of over 500 assets across solar, BESS and wind was onboarded to Clir Renewables, generic AI answered the portfolio's operational questions with roughly 50% accuracy. With Clir's data and models underneath, that figure went to >99%. Today, Clir announced the launch of its MCP Server, making that analytical foundation directly available inside Claude, Microsoft Copilot, Gemini and ChatGPT.
The Problem It Solves
Renewable energy teams have never had more data, and never spent more time turning it into something they can act on. Every board narrative, budget variance commentary and anomaly escalation still runs through an analyst who pulls the numbers, reconciles them by hand and builds the case. The analytics have grown more sophisticated. The manual work between analysis and action has not moved.
The shortcut most teams reach for, pointing a generic corporate AI at their operational data, makes the problem worse in two ways. The first is accuracy. Without codified, renewable-energy-specific data underneath it, a generic AI answers with whatever it can find. On the renewables portfolio mentioned, that means roughly half the answers were wrong.
The second is cost. Renewable energy executives are approving AI budgets and then getting surprised by the token costs. When an LLM has to reason across unstructured documents to find an answer, it uses far more tokens and processes far slower. That cost adds up quickly and grows with every additional user, query and reporting cycle.
The deeper risk is what happens when a wrong answer sounds right. Clir's AI flags high uncertainty and calls out questions it cannot answer reliably rather than guessing. In an industry where decisions are financial and contractual, a confident wrong answer is more dangerous than no answer at all.
Clir's MCP server closes the gap a different way. The hard analysis, including contractual availability reconciliation, forecasting and budget reconciliation, keeps running on the Clir platform where it stays defensible. What changes is the distance between that output and the people who act on it.
How It Works
With Clir's Model Context Protocol (MCP) server, you ask your AI assistant for what you need and it handles the rest. Want this quarter's variance commentary, an availability figure reconciled against a specific OEM contract or the week's top anomaly recovery actions? Ask in plain language, inside the assistant your team already uses such as Claude, Microsoft Copilot, Gemini or ChatGPT and Clir delivers the answer straight into the conversation.
Behind that simple exchange, the assistant calls the Clir MCP server, which runs your request against your codified portfolio data and predictive models and returns Clir's defensible analytical output. The work that used to take an analyst days exporting data, cross-referencing figures and building the case now happens in minutes, without leaving the tool you're already working in.
Every response links straight back into the corresponding views in the Clir platform, so you can verify any finding and examine the underlying data whenever you need to. The Clir app remains your system of record. The MCP server is simply the fastest way to get its intelligence into the hands of the people who act on it.
A Three-Layer Architecture Built to Compound Over Time
Clir's AI Layer is built across three sequential layers.
Layer 1
The first is data. SCADA systems tell operators that a turbine is offline; they do not tell them why. Clir closes that gap by ingesting the unstructured context surrounding every operational event: daily reports, board packs, OEM communications and field notes, structuring them through AI-driven extraction pipelines and aligning them with live SCADA data.
Extraction runs through an ensemble of multiple models simultaneously, with convergence across results determining confidence. Extractions without clear agreement are flagged for operator review. With a hundred sites come a hundred different contractual definitions of availability and loss factors, and the codification work that aligns those definitions to each specific contract is what makes every downstream output defensible.
Layer 2
The second layer is predictive intelligence. Once data is codified, machine learning and physical models run against it: energy yield forecasts, component failure predictions and performance benchmarks drawn from Clir's 350+ GW operational baseline across wind, solar and battery storage.
The resulting intelligence is proprietary to each client and is not shared across the Clir client base. The model trains on each client's specific operational patterns, contracts and OEM history, compounding in accuracy with every reporting cycle.
Layer 3
The third layer connects the first two to the AI the company already runs on. Most Clir customers already have Copilot, Claude, ChatGPT or another assistant deployed internally. Clir's MCP server makes each customer's codified data and predictive models accessible inside whichever tool their team uses, so a site manager gets an immediate picture of what happened overnight and an investment manager can surface key risks from a board pack before the meeting starts.
Clir's data engineering and renewable energy expert teams continuously verify that operational KPIs, contractual definitions and document extractions flow accurately through the model. That sustained oversight is what makes the outputs defensible when they matter most. The MCP server is only as valuable as the predictive layer feeding it, and the predictive layer is only as accurate as the codified data underneath it. Clir's architecture is built to operate as a complete system, with each layer reinforcing the next.
"The Clir MCP server puts our full analytics platform inside the AI assistants your team already uses. Ask for this quarter's variance commentary, this week's top anomaly recovery actions, or your latest OEM dispute brief and get Clir's defensible analytical output back in minutes. Over time, the intelligence layer trains on your own operations and contracts, becoming a private capability tuned to your portfolio." — Gareth Brown, CEO, Clir Renewables
About Clir Renewables
Clir's AI accelerates renewable energy production, improving the economics of projects and ultimately reducing human impact on the planet by further incentivizing the shift away from fossil fuels.
Combining the industry’s largest contextual operational dataset of more than 350+ GW with powerful AI that is built, designed and supported by decades of renewable energy expertise, Clir provides owners, operators and investors with the insights and tools they need to assess and optimize wind, solar and BESS portfolios.
Founded in 2017, the company works with renewable energy investors and their asset managers across Europe, Africa, the Americas and Asia.
Media Information
Contact: Gareth Brown
Email: gareth@clir.eco
Phone: 604-262-2009
