A spacecraft can carry an ingenious payload, a clever propulsion system and a business plan written in the confident future tense. It still needs electricity. Somewhere between the rendering and the launchpad, somebody must make a solar array that fits into a cramped volume, survives violence on ascent, unfolds without assistance and keeps working through the particular cruelties of orbit. Philip Keller has spent much of his career in that somewhere.
His current company, Source Energy, lives in northern Colorado, where the vocabulary is deliberately less celestial. Availability. Capacity. Lead time. Cost. Test cycles. These are factory words, and Keller's bet is that they are now space words too. Source exists because he and co-founder Bryan Mazor kept encountering the same irritant from different seats in the industry: solar arrays were slow to buy, costly to customize and difficult to procure at the volumes newer satellite programs wanted.
The diagnosis is almost comic in its ordinariness. A new generation of spacecraft was being held up by the thing that catches sunlight. Keller did not arrive at that complaint from a pitch deck. His public technical trail reaches back to 2007, when he co-authored work on a deployable array for the TacSat-4 standard bus. That design pursued low mass, few parts and an aggressive price. The themes have aged better than most aerospace acronyms.
Act IThe hinge before the headline
In 2013, Keller appeared at the Small Satellite Conference as a presenter on a 50-watt articulating array for a 3U CubeSat. The design used a shape-morphing composite boom. It stored energy, drove deployment and became the structure that held the panels away from the spacecraft. One piece doing several jobs is elegant on paper and demanding in a test bay. The presentation's abstract admitted the conflicting design requirements. Good engineering stories tend to contain that sentence, even when marketing departments later misplace it.
At Composite Technology Development in Lafayette, Keller later established a business unit associated with a U.S. Marine Corps program of record for 2,500 power systems. At Roccor in Longmont, he built the satellite solar-array unit into one of the company's better-performing product lines before Redwire acquired Roccor. By 2020, he was principal investigator on an Air Force small-business award aimed at a solar array for SmallSat missions traveling from geostationary transfer orbit toward low lunar orbit.
This is the useful context for Source. Keller had seen solar power as research, product line, government program and customer constraint. He knew the romance of a deployed wing. He also knew the purchase order behind it.
We are dedicated to being the most dependable supplier in the industry.Philip Keller, on Source's emergence from stealth
Act IIProof before publicity
Keller and Mazor incorporated Source in October 2021. Mazor brought his own particular proof: solar-array work for Crew Dragon and the original Starlink constellation at SpaceX. Together they framed the problem as an assured-supply challenge. Their customers did not need a handsome promise of abundance. They needed hardware, qualified and delivered.
Source says its first qualified primary-power solution shipped 14 months after incorporation. Another followed later that month, then another two months after that. The company stayed quiet until the hardware had a stronger line on its résumé: orbit. Source panels became the primary power system on Impulse Space's Mira orbital transfer vehicle for its LEO Express-1 mission. When the company emerged from stealth in May 2024, it could point upward rather than forward.
The sequence says something about Keller's temperament as an operator. In space hardware, flight heritage is a form of grammar: it changes a sentence from “will work” to “has worked.” He thanked the early customers who accepted that first-use risk. The public celebration arrived only after their satellites did.
The capital followed the proof. Source disclosed a $6.5 million seed round completed in June 2023, led by Alpine Space Ventures. Bulent Altan, the firm's founding partner and a former SpaceX executive, joined the board. In 2024, Colorado awarded Source a $250,000 Advanced Industries grant to help scale its Longmont production operation. Keller described the money in local and practical terms: it would help the company grow in Colorado and meet commercial demand.
Act IIIBorrow the factory, change the orbit
Source's central idea is to borrow aggressively from terrestrial solar. Earthbound photovoltaics have absorbed enormous investment in automated tools, materials and supply chains. Space solar traditionally favors costly multi-junction cells with excellent efficiency. Source asks where less exotic cells, properly engineered and qualified, can produce a better economic answer for the mission.
That question became physical in June 2026, when Source installed automated shingle-matrix equipment in Colorado following development work with Fraunhofer ISE and M10 Solar Equipment. In a shingled module, cell strips overlap like brickwork and connect through conductive adhesive. The geometry allows current to route around localized damage and gives designers flexibility over voltage and current. Qualification results published that summer projected 76 percent of original power after seven years in orbit.
The Source operating thesis
The factory logic extends beyond cells. A solar wing must remain stowed until commanded to open. Traditional release hardware can limit the number of ground-test cycles, forcing swaps and configuration breaks. Source's 2026 partnership with deployment specialist Dcubed centers on a resettable release nut. Keller praised it for directly addressing integration and test challenges. The mechanism is small; the schedule it protects is not.
This is one of the more stealable lessons in Keller's approach. The sticker price of a component is only one line in the customer's misery. Testing interruptions, replacement hardware, waiting and engineering labor also cost money. A supplier becomes more valuable when it removes the surrounding nuisance.
Our mission is to be a disruptive supplier that redefines availability and reliability of space-grade solar arrays.Philip Keller, announcing Source's Idemitsu collaboration
The next panelA portfolio shaped by radiation
Silicon does not solve every orbit. Radiation becomes a harsher judge as missions move higher and last longer. Source's collaboration with Japan's Idemitsu Kosan adds a different material to the map: thin-film copper indium gallium selenide, mercifully shortened to CIGS. The partners announced plans in late 2025 to develop products for high low Earth orbit, medium Earth orbit and other radiation-heavy missions, with availability planned for 2027.
Keller called Idemitsu's cells a way to expand Source's offerings and serve a growing list of flight customers. The statement is revealingly commercial. It does not pretend one material wins everywhere. Silicon serves a set of economics and environments. CIGS opens another. The portfolio follows the physics.
Meanwhile, Source says it has had hardware on orbit since November 2023 and now serves more than 25 customers across communications, defense, Earth observation, in-space services and on-orbit computing. Its public capacity figure is 10 megawatts a year. Those are company numbers, not a victory lap. Capacity is useful only when orders, processes and quality move together.
The Keller patternMake the constraint less interesting
There is a consistent line from the composite hinge in 2007 to the automated module in 2026. Keller keeps returning to the part of aerospace where mechanical cleverness meets repeatable delivery. His career has moved through engineering, product strategy, business-unit building and the chief executive's office, but the question remains recognizable: how do you make spacecraft power less of a bespoke obstacle?
The answer is not to drain the wonder from space. It is to reserve wonder for the mission. A dependable power system should become pleasantly unremarkable. It arrives when promised. It tolerates the test campaign. It survives launch. It opens. It produces electrons while everybody else talks about the payload.
Keller's version of ambition is therefore concealed inside operational language. Affordability allows more programs to fly. Availability keeps them from idling. Reliability gives them a chance to finish. If Source succeeds, its best work will disappear into other people's missions, black panels quietly doing their shift in the sun. For a founder who has spent two decades making things unfold, that may be the neatest deployment of all.