Fernando
Llaver

The Grid Is Broken. Fernando's Bet Is That Software Fixes It.

There is a number worth knowing: $930 billion. That's the approximate value of renewable energy projects sitting in interconnection queues in the United States alone, waiting for access to a grid that has no room. Solar farms built, wind turbines spinning - nowhere to send the power. Not because the wires aren't there. Because no one has taught the wires how to share.

Fernando Llaver has spent more than two decades watching this problem get worse. He didn't inherit the energy industry's assumptions. He sat inside them, inside C-suites and operations rooms, long enough to see exactly where the logic broke down. The answer the industry kept reaching for - more infrastructure, more towers, more lines - was expensive, slow, and largely unnecessary. The grid had capacity. It just didn't know how to use it.

In 2021, he walked away from a comfortable position at the top of the energy hierarchy and co-founded Splight. The thesis was precise: treat the grid not as a dumb physical network but as a data problem. Apply machine learning. Watch the congestion clear.

Dynamic Congestion Management: The Grid's Hidden Lever

Splight's flagship product is called Dynamic Congestion Management, or DCM. It uses proprietary machine learning algorithms to analyze real-time data across transmission networks - modeling what the grid can actually handle at any given moment, versus what legacy static ratings say it can handle. The gap between those two numbers is where Splight lives.

The logic is similar to what airlines figured out with yield management, or what logistics companies did with routing optimization. The infrastructure was always capable of more. The constraint was always informational. Splight gives grid operators the intelligence to act on what their networks are actually doing - not what a manual from 1985 says they should be doing.

The platform operates across the U.S., Spain, Chile, and Portugal - navigating different regulatory environments, grid architectures, and data standards. That geographic spread isn't just a revenue story. It's evidence that the underlying machine learning approach generalizes. A grid is a grid, from Santiago to Sacramento.

From C-Suite to Co-Founder: The Long Game

Fernando didn't arrive at Splight through a pivot. He arrived through accumulation. Two decades in the energy industry - at the intersection of operations, platforms, and infrastructure - gave him a map of exactly which problems were structural and which were solvable. His time at Digital Energy Markets as an Orchestration Platform Owner sharpened his thinking about energy systems as data systems.

He studied at the University of Chicago Booth School of Business Executive Education, adding business architecture to his engineering instincts. Then, at some point that isn't marked on any press release but is visible in the decision itself, he decided the problem was big enough and the solution clear enough that staying on the inside of a large organization was the wrong place to be.

What the Record Shows

$28.1M and the Investors Who Bet on the Grid

The investor roster tells a story about where the grid problem sits in the global conversation. EDP Ventures is the VC arm of one of Europe's largest electricity companies - a strategic check that said the technology works on real infrastructure. Draper Cygnus and Draper B1 brought Silicon Valley network. UC Berkeley Foundation, which almost never writes startup checks, was in the seed round. For the most recent raise, Blue Bear Capital - a firm focused specifically on energy transition infrastructure - led.

Data center operators and utilities are now active in Splight's pipeline, which explains the timing and size of the 2025 round. AI infrastructure is consuming electricity at a pace that makes the grid problem urgent in a new, commercial way. Splight's order book was doubling or tripling annually before the AI data center wave arrived. The current demand environment is qualitatively different.

Fernando Llaver in His Own Words

The Specific Details That Explain the Big Picture

There is a particular kind of founder who is dangerous to incumbents: not the one who doesn't understand the industry, but the one who understands it completely and decides the assumptions are wrong anyway. Fernando Llaver is the second kind. He spent two decades inside energy - not writing think pieces about it, but operating inside its infrastructure, understanding its constraints, tracking where the waste accumulated.

The Unreasonable Group, which works with founders who are tackling global-scale problems, describes him as an "Unreasonable Entrepreneur" - someone willing to attempt what the reasonable industry consensus says isn't possible. In Fernando's case, the unreasonable bet was that machine learning could do in software what billions in physical infrastructure couldn't do in steel: give the grid room to breathe.

The IEA invitation to Paris is one of the quieter signals about where Splight sits in the global energy hierarchy. The International Energy Agency doesn't call startups for show. When policymakers from major economies are trying to understand how AI changes the grid, and they want a technical operator in the room, that's a different category of recognition than an award or a press release.

Fernando co-founded Splight with Thomas Vadora (CTO) and Carlos Caldart (Chief Data Scientist) - a team that balances business-side energy industry experience with deep machine learning research credentials. The company now runs on 69 people, spread across operations in four countries. For a company managing gigawatts of infrastructure, that's a lean number. It reflects the original thesis: this is a software problem, solved with software economics.

His aspiration isn't incremental. The goal is to make AI-driven grid intelligence the default operating mode for transmission networks globally - so that the congestion holding back clean energy isn't a project to eventually solve but a condition that's already been solved, quietly, in software, before anyone had to dig up a mile of ground.

A Few Details That Stick

• Splight's AI can give existing power lines 2-3x more firm transmission capacity - without installing a single new tower.
• When Llaver co-founded Splight in 2021, $930 billion worth of renewable energy projects were stuck in U.S. grid queues. His answer was software, not steel.
• AWS picked Splight from 800+ startups in 74 countries for its Clean Energy Accelerator - the energy-sector equivalent of making a World Cup squad from open tryouts.
• Splight operates in four countries across three continents with just 69 people - a ratio that only makes sense if the product is software.
• The IEA invited Fernando to Paris to help shape global energy and AI policy - a call that comes when your software is managing six gigawatts of real infrastructure.
• UC Berkeley Foundation, which almost never invests in startups, participated in Splight's seed round - a notable signal from an institution that takes long views.