HEV Partners
We invest in
Artificial intelligence.
The physical world.
And where they converge.
We're two siblings who went deep on opposite sides of the line, one in energy and hardware, the other in machine learning and software. Somewhere in the last ten years the two stopped being separate subjects. Now we back founders from seed to pre-IPO, with a thesis that fits on a whiteboard:
01
Humanity is entering its next technological era on infrastructure built for the last one.
02
How we work and live will be re-engineered around AI and autonomy.
03
How much we can do depends on power, materials, and machines.
Most investors live on one axis. Software funds along the top, hardware and energy funds along the bottom. We invest across the whole chart, and we can do the technical work ourselves on the green diagonal too, where a model and a machine are the same product.
01
What we back
We are technical on both sides, so we can underwrite any cell in this table, including the middle one.
Mostly a model
Where they meet
Mostly atoms
Power & grid
Forecasting, dispatch and trading software for power markets.
Grid equipment that thinks. Solid-state transformers, smart inverters, software-defined power.
Generation, storage and transmission hardware.
Compute & AI infra
Frontier models. AI clouds and the software stack on top of them.
Power and cooling for compute. Energy-first data centers, on-site generation.
Chips, transformers, turbines, cooling hardware.
Robotics & autonomy
Simulation, perception, world models.
Physical AI. Robots that learn on the job, in factories and in the field.
Actuators, sensors, motors, batteries.
Industrials & manufacturing
Design, planning and control software for factories and supply chains.
Autonomous production. Plants and processes with a model in the loop.
Advanced materials, machine tools, new ways to make old things.
02
The partners
Two siblings who went deep on opposite frontiers, then started investing together.
We grew up at a dinner table where the conversation centered around problem-solving. A math puzzle one of us had brought home and dared the table to solve. A piece of technology taken apart until everyone understood how it was made. And a standing rule that you had not really learned something until you could explain it to someone with no background at all, and they got it.
We started early and went in opposite directions. Evan wanted to be an inventor, then found math, then computer science, then machine learning, which turned into AI. Hilary started in the humanities and, somewhere in middle school, became fixated on energy as the thing underneath everything else: change the energy system and the change ripples through all of civilization.
We started investing together as undergraduates at MIT, first in public markets, betting on which companies our two fields would make winners. Then earlier, and earlier. Now we invest to keep learning, and back the companies that create with purpose.
Energy · Hardware · Atoms
Hilary Vogelbaum
Partner
Early
Organic photovoltaics research at Case Western; first published paper. Energy policy and power-market work through school.
MIT
SB, Materials Science & Engineering, concentration in energy and political economy. Research at the MIT Energy Initiative on the techno-economics of energy technologies.
Corporate
Shell's corporate seed team (microgrids and power tech). Iberdrola corporate strategy and utility-scale renewables, including 24/7 clean-power products for corporates.
VC
First hire on the S2G Ventures energy fund. $300M of flexible capital across power, mobility, industrials and the built environment.
HBS
MBA, Harvard Business School. Climate & Energy Club.
In plain English
Flexible capital
A fund that can write an equity check or a loan, whichever the company actually needs. Rare in energy, where the two usually live in different buildings.
Software · Machine intelligence · Bits
Evan Vogelbaum
Partner
Early
Math and CS competitions at Phillips Exeter; RSI 2018.
MIT
SB and MEng in Computer Science. Published first author papers at ICML. Co-led AI@MIT and director at SBC@MIT.
Quant
High-frequency equities signals research at Hudson River Trading, the world's 2nd largest proprietary trading firm.
AI
Machine learning lead at Reducto (a16z-backed, Series B).
In plain English
Faster loops, better answers
Software performance, trading, and machine learning all come down to one thing: how quickly you can go from signal to decision to result. Evan's work has been shrinking that loop.
03
Selected investments
Where each company sits on the diagonal.
Mostly a model
Mostly atoms
Frontier Models
Growth
AI Neo-cloud
Growth (Exited)
Energy-first compute
Growth
Solid-state transformers
Early-stage
Fully reusable rockets
Growth
Also invested across frontier models and AI search, data for robotics, AI infrastructure and developer tools, enterprise AI, and cybersecurity.
04
On the whiteboard right now
Our current worldview
Short essays on the theses behind our investments. Do you agree? Want to challenge our thinking? Write to us.
Hilary · the physical side
The grid is going medium-scale
Centralization is so last decade. Fully distributed takes forever. The future of the grid will be measured in megawatts.
Evan · the AI side
Optimization eats the world
Software ate the world by encoding processes. AI eats it by optimizing them. The first wave wrote down how things work. This wave asks how they should work, then rewrites them every night.
Building in AI, in the physical world, or in the space between?
Write to whichever of us is closer to your problem. We read everything together anyway.
HEV Partners
Hilary Vogelbaum & Evan Vogelbaum · © 2026