For most of the last decade, solar occupied a peculiar, second-class place in the American energy conversation. When Silicon Valley placed its energy bets, the glamorous money went to nuclear, including Sam Altman’s Oklo, Bill Gates’s TerraPower, and fusion rounds with Nvidia and Google on the cap table. Solar was seen as far gone, evocative of dumb panels on rooftops, manufactured at scale in China, and sold on thin margins. Chinese oversupply had crushed module prices to nearly nothing; solar’s entire business model became a matter of commodity pricing rather than technological innovation. Reasonably, the "solar is dumb" view was nearly consensus.

That's starting to change. Katie Miller, wife of White House deputy chief of staff Stephen Miller, former DOGE aide, and a reliable weather vane for MAGA sentiment, posted that solar is "the energy of the future." The sun, she added, is the "giant fusion reactor up there in the sky." Fabrizio, Lee & Associates, Trump’s own chief pollster, found that 70% of Republican voters back utility-scale solar when the panels are American-made; Kellyanne Conway’s firm found 75% of Trump voters across Arizona, Florida, Indiana, Ohio, and Texas want solar used to strengthen the US energy supply. In 2025, 73% of all new US solar capacity went up in red states. Even Energy Secretary Chris Wright (who told a House subcommittee in June 2025 that an intermittent source like solar is "just a parasite on the grid") insists he's "pro-solar," and says the technology should compete on its own merits.

Solar's new trendiness has an obvious explanation. Data centers now need over 100 gigawatts of new capacity by 2030, which is roughly the output of 100 large nuclear reactors. Residential electricity rates climbed about 10% nationally over the past year, with more than a dozen states seeing double-digit jumps. China installed 315 gigawatts of solar in a single year — more than most countries have ever built. Solar made up 85% of new capacity added to the American grid in Q3 2025; solar paired with battery storage is projected to be responsible for 79% of new capacity in 2026.

Solar's case is being made on deployment speed, cost, and geopolitical urgency — all real concerns — but not on solar as a technology. The underlying electrical architecture inside a solar installation has barely changed since the early 2000s. Worse, utility-scale farms routinely underperform their own production models by eight percent on average, leaving billions in lifetime value on the table.

One company has spent nine years betting that solar can be smart. Optivolt, founded by Thiel Fellow Rohit Kalyanpur, embeds aerospace-grade power electronics directly inside panels to intelligently modulate power flow in response to changing weather, light, and demand conditions. They claim enormous improvements, like up to 6x better panel performance in shaded conditions, compared to legacy installations.

Optivolt further bets their technology tips both sides of the levelized cost of energy (LCOE) ratio, total system cost divided by the total energy a system produces over its life. If they’re right, they’ll fundamentally change solar’s unit economics. 

Solar cells within a panel are wired in series, so electricity flows through every cell in sequence, like bulbs on a string of Christmas lights. If cells get shaded (by a tree, dirt, bird poop, snow) they choke the current flowing through their section of the panel. The industry's usual fix is the bypass diode: panels are divided into a few zones, each with a diode that trips when shading drags its zone down, routing current around the whole zone so the rest of the panel keeps working. But the diode is only a safety valve. The energy the bypassed zone could still have produced is thrown away, not recovered. Because a single diode switches off a third of the panel at a time, a single shaded cell in a 108-cell panel can erase more than a third of its output. Plus, when shading is light enough that the diode doesn’t trip, the shaded cells simply drag their whole zone down anyway. Diodes also sometimes fail, and they don’t stop hotspots: roughly one panel in 10 develops a hotspot or diode failure over its life. Hotspots cook the cells around them and, on a roof or in a field, can start fires.

And shade is everywhere. Most American rooftops have some mix of shade, dust, or debris; every utility farm has soiling and interrow shading. The problems compound year over year as cells age at different rates.

The average solar farm underperforms projected performance by eight percent, which works out to $20 to $40 million in lost lifetime revenue per site, and up to $4 billion across a 10-gigawatt portfolio. To manage all these vulnerabilities, the industry has tacked on a series of improvised solutions that never resolve solar’s underlying design flaws, like drone flyovers with thermal cameras, manual fault isolation, and repair crews sent out with almost no data. This is how the industry “maintains” gigawatt-scale energy assets.

Plus, all energy optimization currently happens outside the panel. Solar installations come with power electronics, meaning components that manage, convert, and route the raw direct-current (DC) electricity panels generate. (This includes inverters, microinverters, optimizers, charge controllers, and, yes, even the bypass diode.) Together, they determine how much of what a panel makes actually reaches the grid. 

Unfortunately, panel manufacturers (mostly Chinese, commodity-focused, running on razor-thin margins) and power electronics manufacturers (higher-margin, IP-driven, usually American and Israeli) almost never collaborate. Enphase and SolarEdge built the category of module-level power electronics (MLPE) in the past 15 years, which made solar meaningfully better; they deserve credit. But today’s solar electronics are still just band-aids bolted onto a flawed design. 

A microinverter (the small box that converts and tunes the output of a single panel) can decide how much power to pull from that panel, but it can't move energy between the cells inside it. When a leaf covers three cells and drags a whole zone down to zero output, the microinverter just works with what’s left.

And electronics are expensive. Tesla published a white paper analyzing roughly 13,000 residential solar sites and found that for 80% of homes (those with moderate to good sun) module-level electronics delivered only one to two percent more energy than a plain string inverter, despite costing thousands of dollars more per install. You can imagine what that does to system-level LCOE.

No one tried to balance power at the level of cells themselves because the industry agreed it was impossible. 

Kalyanpur was 19, a computer engineering student at the University of Illinois, when he started working on solar energy. Every project just kept pushing him to solve what the rest of the industry had written off: redesigning how a solar panel is built from the ground up, with better power electronics that are embedded inside the panel itself. 

In 2017, Kalyanpur founded Optivolt and built what he calls the Power Balancer. It redistributes energy inside the solar panel at the substring level in real time, letting weaker cells borrow from stronger ones; instead of a shaded cell dragging down its neighbors, the strong cells carry it. Active redistribution replaces passive bypass, so bypass diodes vanish entirely. So do two of every three junction boxes, four of every six connectors, trunk cables, and combiner boxes. Folding the electronics into the panel makes the entire solar install simpler. Optivolt holds 13 patents on the architecture, with 25 more pending.

Convincing solar manufacturers to abandon their previous designs to try something completely new was not easy. Three barriers explain why everyone thought this was impossible:

Heat. Power electronics generate heat; solar panels degrade from it. Putting one inside the other was considered a non-starter. Optivolt’s answer was to make the electronics barely generate heat at all: the Power Balancer runs at 99.4% efficiency in typical conditions, so only about 0.6% of the energy passing through it is lost as heat, and the panel's thermal, materials, and mechanical design absorb that small amount gracefully.

Making power electronics generate almost no heat required aerospace-grade power-systems engineering: the kind done on satellite programs, where electronics survive decades of thermal cycling with no viability of repair. To do so, Optivolt recruited chief architect Linda Irish. She holds more than 40 electronics patents and previously designed power systems for a Jupiter satellite program and wireless-power systems at Qualcomm.

Regulation. No safety standard existed for module-integrated power electronics at UL, the private lab whose certifications electrical codes and insurers defer to; there was no easy way to certify the safety of Optivolt’s products. Optivolt invested four years in defining new safety test standards with UL, work funded in part by the US Department of Energy.

Cost. Integrated electronics had to replace the panel’s existing junction-box components, not add to the bill of materials. Optivolt’s supply-chain work brought the Power Balancer to a net simplification: fewer parts, lower total system cost.

Years of work later, Optivolt has a device that survives the same thermal cycling as the panel it lives in, backed by a 30-year warranty. The Power Balancer is built in Mexico and integrated into panels assembled in California. Kalyanpur, now 28, has been building toward this for nearly a third of his life.

Most solar innovations move just one side of the LCOE ratio. A cheaper panel lowers the numerator, the system's cost; a more efficient cell raises the denominator, its lifetime output. Either way, the gains are incremental.

The electric-vehicle market was stuck in a similar problem for years: electric cars cost too much and couldn’t go far enough. Finally, improved battery chemistry, vertical integration, and manufacturing scale pushed on cost and range at once. Kalyanpur hopes his technology unlocks a similar transformation. On the cost side, Optivolt’s Power Balancer replaces many existing hardware components rather than adding new ones, and no additional electronics are needed to retrofit in the field. (Optivolt panels are also far cooler when something blocks them. Hotspot testing has recorded shaded cells approaching 200 °C, hot enough to burn cells and fail encapsulation. This is why solar farms sometimes catch fire. Cooler panels can last longer and incur fewer maintenance costs.)

And as far as energy output, Kalyanpur says Optivolt delivers up to 50% more systems-level energy in real-world shaded conditions, and up to six times the yield in heavy shade. Utility farms see less shade than rooftop solar but still bleed energy to soiling (dust, debris, row shading). 

Put the two together and Optivolt projects a residential LCOE roughly 30% below a conventional system's (the exact figure varies from home to home). And for utility, their 250-megawatt simulation showed 13.8% more energy under heavy dust and soiling, worth $42 to $60 million in additional lifetime revenue per site. 

Kalyanpur is ready for solar to go mainstream.

Optivolt has shipped more than 20,000 off-grid units, secured US military and government contracts, and is launching its residential system in California in 2026. Their first utility-scale deployment is planned for later this year.

By figuring out how to embed power electronics directly into panels, Optivolt has also embedded sensors that monitor realtime panel output, temperature, and health. For utility solar, this visibility hopefully ends the need for drone inspections; at home, owners track these metrics through an app. Optivolt believes this information can proactively alert owners of maintenance issues, settle warranty claims against actual production curves, and help operators underwrite financing on verified output rather than projections.

Kalyanpur’s ambitions still manage to go further. He thinks fully off-grid data centers, run on solar and batteries, are only years away: cell-level balancing improves uptime enough to make them viable. Longer term, the team is looking at space-based solar (satellites that gather sunlight in orbit and beam the power down), where the same aerospace engineering behind the Power Balancer applies almost directly.

The technology to close the gap between solar’s performance and its potential is coming. It just involves mental resolve, tedious regulatory work, and teaming up with satellite engineers to solve some hard physics.

Sunlight is the most abundant energy source available to civilization yet we somehow capture almost none of what it offers. Solar isn’t intermittent the way people casually assume; it's predictable, cyclical, and, with enough capture and storage, continuous. Difficulties arise only between photons hitting solar cells and electrons reaching the grid. Fortunately, these are challenges ripe for new technology, not permanent fixtures of the hard laws of physics.