There is a tidy version of the energy transition in which somebody buys a wind turbine, plants it in a breezy field, and waits for electrons. Invenergy lives in the untidy version. A real power project needs landowners, weather data, interconnection rights, environmental reviews, local permits, equipment, tax equity, construction debt, a buyer for the power, a control room, and a crew willing to answer the phone at 2 a.m. The turbine is the photogenic part. The company is built around everything that makes the photograph possible.
Founded in Chicago in 2001 by energy entrepreneur Michael Polsky, Jim Murphy, and a small group of experienced power executives, Invenergy now says it has developed more than 225 projects representing about 38 gigawatts across four continents. It operates more than 25 gigawatts. Its menu runs from natural-gas plants to land-based wind, utility solar, batteries, long-distance transmission, domestic solar manufacturing, clean hydrogen, and a new geothermal push. In April 2026, Murphy became CEO while Polsky moved into the Executive Chairman role.
Its product is credible megawatts
Invenergy’s customers are utilities, municipalities, corporate energy buyers, grid operators, and owners who need somebody else to run an asset. The company reports roughly 90 customers, 22 gigawatts contracted with utilities, and 6 gigawatts with corporations. Meta’s announced agreements with Invenergy reached about 1.8 gigawatts in 2025. Verizon’s deals crossed 1 gigawatt. Honda and Tesla have appeared among other corporate counterparties.
Those companies are not shopping for a handsome solar panel. They want power on a date, at a price, in a particular market, with the contractual attributes needed to support a corporate energy claim. A virtual power purchase agreement can settle financially while electricity enters the local grid. A utility deal may transfer ownership. Another client may want Invenergy to keep the plant and operate it for decades. The output is electricity; the product is confidence that a complicated chain will hold.
That integrated stack separates Invenergy from a specialist developer who sells after permitting, an equipment maker who stops at delivery, or a regulated utility limited to its service territory. NextEra Energy Resources, AES, EDF Power Solutions, Pattern Energy, EDP Renewables, Brookfield Renewable, Clearway, Avangrid, and Apex all compete in parts of the same market. Invenergy’s distinction is not that nobody else can build a solar farm. It is the breadth of technology and lifecycle capability inside a privately controlled company.
The first thing that failed was the romantic story
Invenergy’s first wind project was Buffalo Mountain, a 27-megawatt installation on a Tennessee ridge. It required unusual single-blade turbine erection over former coal-mine terrain and a $31 million loan. Years later, Polsky described the project as disappointing. Wind did not automatically produce attractive economics simply because the resource was clean and the technology exciting.
The early portfolio suggests what changed the founders’ minds: do not build a company around reverence for one machine. Build around commercial judgment and operations. Invenergy’s first major operating asset was actually Hardee, a 370-megawatt Florida gas plant acquired in 2003. In 2004, it bought an unfinished Illinois plant out of bankruptcy with just one of four intended turbines installed. The team found a second use for the equipment logic in an Ontario proposal. Awkward assets became raw material.
“By my nature, I’m a developer.”Michael Polsky
The company later built its own operations platform. Its first remote operations center began in a spare room at a Texas wind site with four computers - one per project. That detail explains the owner-operator loop better than a strategy diagram. People who watch equipment fail, schedule repairs, and measure output can send lessons back into engineering, procurement, financing, and the next contract. In physical infrastructure, operations are customer support and product research at once.
Big plans still meet small choke points
Integration does not make Invenergy immune to vetoes. Wind Catcher, a proposed 2-gigawatt Oklahoma wind farm and dedicated transmission connection, carried a $4.5 billion price tag. Three state approvals and a federal approval could not save it after Texas regulators rejected the plan in 2018. American Electric Power canceled the project. One missing permission defeated enormous engineering and financial work.
Grain Belt Express is the larger, longer sequel. The planned 800-mile high-voltage direct-current line would move up to 5 gigawatts from Kansas through Missouri and Illinois toward Indiana. Its estimated cost has risen to roughly $11 billion. In July 2025, the U.S. Department of Energy terminated a conditional guarantee of up to $4.9 billion for the first phase. Invenergy’s answer was to keep the project alive under a private-financing plan. The episode is not proof that the line will succeed. It is proof that financing optionality matters when projects outlive administrations.
Offshore wind exposed a different choke point: supplier concentration. Leading Light Wind, planned off New Jersey for as much as 2.4 gigawatts, struggled to secure suitable turbines after manufacturers changed products and pricing. Supply-chain and regulatory conditions moved faster than the project. Invenergy canceled it, then agreed in June 2026 to conclude four offshore leases in exchange for $765 million in lease-fee reimbursements. It said it would direct capital toward natural-gas and geothermal projects that could advance on a commercially reasonable timeline.
That was the mind-changing condition made explicit: not “Is this technology good?” but “Can this project reach customers on a financeable schedule?” The reimbursement did not cover interest or incremental development spending, so the exit was hardly free. Yet sunk cost is a terrible fuel. Invenergy preserved the option to revisit offshore wind if market conditions change.
The portfolio is a view of the grid
Invenergy still calls its mission cleaner, more reliable, affordable energy. The ordering matters. Reliability and affordability help explain why a company known for renewables is advancing gas plants and geothermal leases. Solar and wind can be low-cost sources but are variable. Batteries shift energy across hours. Gas responds when needed. Transmission connects distant resources and balances regions. Geothermal could supply steady power without combustion. The portfolio is less a moral ranking of machines than a view of the jobs a stressed grid must perform.
Data centers sharpen the argument. The company is working with NVIDIA and Emerald AI on “flexible AI factories” that could adjust computing activity to grid conditions. If it works, a large load becomes partly dispatchable: intensive jobs move toward periods of greater power availability instead of demanding an identical draw every minute. Invenergy brings the generation and grid knowledge; NVIDIA supplies computing architecture; Emerald AI orchestrates workloads.
This is early, and the limits are real. Some computing jobs are latency-sensitive. Grid programs differ by market. Hardware still needs near-continuous utilization to earn a return. The model is most plausible for divisible workloads, valuable interconnection locations, clear utility incentives, and customers willing to trade a little scheduling freedom for faster power access. Without those conditions, “flexible” becomes a presentation adjective.
What builders can steal
Customers want credible power, not a favorite technology. Organize around their constraint and keep the hardware replaceable.
Operations create information. Feed failures, maintenance costs, and real output back into design and contracting.
Long projects cross elections and rate cycles. Maintain more than one path to capital before the public path disappears.
Define the supplier, permit, schedule, and economics that must remain true. When several break, redeploy capital.
The playbook does not travel everywhere. It requires deep technical talent, patient institutional capital, repeat customers, and a market large enough to reward a broad platform. A young developer that tries to own every lifecycle step may simply multiply overhead and risk. Vertical integration works when each layer improves the next one and when the company has enough projects to keep specialized teams busy. Otherwise, partnership is cheaper.
Invenergy has attracted that patient capital without surrendering daily control. Blackstone announced approximately $3 billion in 2022 and another $1 billion in 2023. CDPQ remained a long-term owner. In 2025, an upsized $2.5 billion revolving credit facility added flexibility for the development pipeline. These are not venture rounds funding a fast experiment. They are reservoirs for equipment deposits, letters of credit, acquisitions, and years of work before a plant earns revenue.
The result is a company that looks less like a clean-energy brand than an industrial translation service. It translates windy land into a utility contract, corporate demand into a financed solar farm, an unfinished plant into a usable asset, and a grid bottleneck into a transmission proposal. Sometimes the translation fails. The interesting part is what happens next: another financing structure, another technology, or a decision to stop. In a business measured in decades, patience is not waiting. It is active capital allocation.