Back to swell.vcSWELL VC · THE FUND III THESIS
AI for critical infrastructure
Fund III invests at pre-seed and seed, building a concentrated portfolio across space and defense, enterprise infrastructure, industrial systems, and biology. Our bet is that as AI advances, value accrues to the companies that build and operate these systems. We back the founders doing that work.
Satellites, mines, power plants, factory floors, laboratories and the supply chains between them run the economy and defend the country. Software already helps operate them, but much of the work still depends on people interpreting complex signals and responding to conditions that rarely repeat exactly.
AI expands what can be understood and automated in these environments. Models can interpret sensor streams, help design proteins and turn fragmented inventory records into decisions. Founders can build products that were previously impossible, changing how these industries operate.
Building these products creates assets that access to the same models does not give a competitor: satellites already operating in orbit, sensors underground, laboratories generating experimental data, and software integrated into customer operations. Better AI expands what founders can do with those assets. Operating them generates data competitors cannot get from the outside, and the founders who turn that data into their next product extend the advantage before anyone else can.
WHERE WE INVEST
Satellite platforms and defense technology built to operate under mission conditions.
Loft Orbital · Array Labs · Deepnight · Modern Intelligence
Energy, robotics and manufacturing companies developing new ways to expand resources, produce energy and automate physical work.
Parametric · NanduX · Exodys Energy
Companies combining foundation models with wet labs, where experiments test what the models predict and generate data that improves the next round of designs.
RyboDyn · Onava
Software that runs, coordinates or secures consequential operations, with deep integration into the customer’s business.
Crisp · Credal
Fund III’s focus excludes consumer apps, general marketplaces and undifferentiated SaaS.
THE TEST
Better models expand what customers can do. The company’s advantage comes from what it has built and what it learns by operating it.
FROM FUND I TO FUND III
AI lets satellites interpret what they see in orbit and alert people within seconds, so they can respond while it still matters.
That’s what Loft Orbital is enabling at scale through the newly announced $1B program with Marlan Space. Loft was our first investment out of our first fund. We backed the team in 2018, before they had a product or revenue.
Better models expand what customers can do with Loft’s satellites. Reproducing that service takes spacecraft and years of mission operations, even when the model is available to everyone.
Read about the program ↗THE SAME THINKING, IN DIFFERENT BUSINESSES
INDUSTRIAL SYSTEMS · FUND II
Parametric is building robots for warehouse operations where the variety of products and tasks has made traditional automation uneconomical, starting with box assembly and fulfillment.
Better AI expands the work those robots can perform. Delivering it reliably takes hardware designed for the task and experience inside customer facilities, with each deployment generating data the team can use to improve the next one.
INDUSTRIAL SYSTEMS · FUND III
NanduX combines muon sensors with AI to help locate copper and guide drilling decisions. The measurements come from sensors deployed underground, giving its software information that existing datasets cannot supply.
The company has announced an agreement with Carpala Minerals to acquire a 25% stake in the Placeres copper project in Chile, where it plans to guide the first drilling program. It is turning its exploration technology into ownership in the assets it helps discover.
BIOLOGY · FUND III
Onava connects protein design with laboratory tests of biological function. More capable design models can improve what the team tests, while the experiments generate evidence that does not exist until the work is done.
A competing team needs the experimental methods and biological expertise to produce that evidence and use it in the next design cycle.
ENTERPRISE INFRASTRUCTURE · FUND I
Crisp connects and standardizes retail data so brands and retailers can manage inventory and replenishment. Its AI agents use that foundation to help teams identify problems and act on them.
The foundation is the point: a competing model does not arrive with Crisp’s retail integrations or its understanding of how products move through customer operations.
A PERSPECTIVE ON HOW ADVANTAGES EVOLVE
A head start built on data does not hold on its own. Difficult-to-build systems buy time. Operating them generates the knowledge that will one day make them easier to reproduce. The founders must use that time to turn what they learn into a capability they own. Explore how that choice plays out over successive operating cycles.
A conceptual illustration, not a prediction.
WHY SWELL
We backed Loft before product or revenue. At that stage, the investment rests on understanding what the founders know, what they can build, and why customers will need it. It also rests on who they are. The founders we back have something to prove, and they cannot leave the problem alone until it is solved.
Rusty and Jay evaluate every company together. We invest with concentration, and each investment must have the potential to return the fund.
WHERE WE BEGIN
At pre-seed and seed, we are underwriting the people and the opportunity while the advantages are still taking shape. We back founders with rare technical depth who are obsessed with solving the customer’s problem.
If you’re building one of these companies, connect with Swell ↗