David Fiszel.
"Somewhere in Ohio, there is almost certainly a manufacturing company that makes radiators," says David Fiszel, the founder and Chief Investment Officer of Honeycomb Asset Management, "but doesn't know they're about to become a growth company thanks to orbital data centers. If you believe in data centers in space, you need a million radiators over the next ten years for these data centers."
This is the picks and shovels layer of the space economy, and it is a window into one of the most distinctive investment minds operating in the hedge fund world today. While the rest of the market debates which rocket company will win the launch wars or which satellite operator will dominate low-Earth-orbit broadband, Fiszel is three steps ahead, asking a more fundamental question: what must exist at an industrial scale before any of this becomes real?
The Architecture of the Thesis
Fiszel's thesis is straightforward in design, even if its implications are still widely underappreciated.
Computing is expensive. A significant portion of that expense is thermal. The GPUs that power modern AI training and inference workloads generate enormous heat, and managing that heat on Earth requires energy-intensive cooling systems that add substantial cost to every data center operator's balance sheet. Space offers a different thermal equation. The vacuum of space eliminates convective cooling entirely, which means waste heat can only be shed one way: by radiating it out into a background that sits at roughly 2.7 kelvin, just a few degrees above absolute zero. That constraint is precisely what makes radiators, not chillers or cooling towers, the critical hardware of orbital computing.
"Space is about three kelvin, super cold," Fiszel explains. "And that could be a cheaper way to compute."
Combined with laser inter-satellite link technology that is already proven in commercial satellite constellations, abundant uninterrupted solar power, and a rapidly maturing reusable launch industry that is driving the cost of putting payloads into orbit down toward commodity pricing, the economics of orbital computing begin to look, if not inevitable, at least plausible within a ten-year window.
"Maybe in 2026 this seems ridiculous," Fiszel says. "But by 2029, people might have a complete flip on that perspective."
The Supply Chain No One is Mapping
Every orbital data center requires rockets to get there. Rockets require propellant, and the single largest component of that propellant by mass, across virtually every heavy-lift launch vehicle flying today, is liquid oxygen, the oxidizer that must be produced, stored, and delivered at scale. As launch frequency increases, which it must if orbital infrastructure is to become economically viable, the demand for liquid oxygen will scale with it in a near-linear relationship.
Fiszel is looking at the largest suppliers of liquid oxygen. Not because he expects to corner the market in industrial gases, but because he believes the market has not yet priced those companies as the growth businesses they are about to become. The fundamentals of the business have not changed. The demand curve is about to.
Radiators tell a similar story. Thermal management is not just a problem for terrestrial data centers. It is a problem for any computing system, orbital or otherwise, that has to manage the heat generated by high-density processing. In orbit, where radiation is the only mechanism for rejecting heat, the problem is existential. The solutions to that problem, at the scale required for orbital infrastructure deployment, will need to be manufactured somewhere, by someone, using supply chains that aren't yet available at the required volume.
"I'm looking quite heavily into which companies, in order to have data centers in space, you need a lot of rocket launches," Fiszel says. "And so what's needed for that?"
The Creative Edge
What Fiszel describes as his core competitive advantage is not a proprietary model or a superior data feed. It is, in his own framing, creativity. The ability to see a connection that is not yet obvious, to hold an idea that seems premature, and to translate that idea into a specific, actionable investment thesis before the market has done the same.
"I'm a creative, and that is definitely my superpower," he says. "I think about things in a way that I think is different. I'm also an optimist, so I tend to be a believer in things that people don't necessarily see."
The orbital data center thesis and the supply chain thinking that flows from it are perhaps the clearest current expressions of that. It is a thesis that requires you to believe several things simultaneously: that the thermal engineering problems will be solved, that launch costs will continue to fall, that laser-linked transmission will scale to commercial data center workloads, and that the industrial companies sitting at the base of that supply chain are about to be rerated by a market that has not yet connected the dots.
Each of those beliefs is individually defensible. Together, they point to a very specific set of investment opportunities that most of Fiszel's peers are not currently considering.
There is a reason this layer is not the headline story in most coverage of the space economy. Radiators are not exciting. Liquid oxygen is not exciting. The picks and shovels layer of any technological revolution rarely is, at least not until the trajectory is far enough along that the infrastructure underpinning it becomes impossible to ignore.
