This post was originally published on https://inflection.fund/writings/hard-futures
We are renaming our SVRGN bog to Hard Futures. The content is re-orgnised to optimise for exploration and search.
Capital flows downhill. It pools where the TAM is clear and the next round is already circled. In 2026, 90% of all venture dollars are deployed against AI and 50% of LP $ are raised by the largest 10 venture firms. When consensus grows, capital concentrates and prices inflate. Returns compress. The crowd has never been larger or more expensive to stand in. This dynamic opens outsized opportunities for those operating outside of it.
We work uphill. We invest into hard futures. We look for exposure to companies that contribute to bringing a hard future to reality:
Hard technology. We back businesses that fuse bits and atoms: compute in the stratosphere, silicon that reprograms itself, defense for assets in orbit, software led optical factories, space laser ranging. Durable moats form where code has to move matter. That is where economies of scale meet data flywheels and where durable businesses compound.
Hard markets. We go where consensus capital won’t — unfashionable geographies (like Europe), contested domains, the earliest stages, sectors traditionally resistent to innovation. Difficult terrain thins the crowd. A thin crowd leaves low prices and asymmetric, high convex opportunity. Low entry valuations are earned by being there before anyone else was looking.
Hard to underwrite. When a founder describes bodyguard satellites, self-reprogramming silicon or an interactive map of the subsurface, there is no comparable or playbook. We read that difficulty as the signal. The harder a thing is to underwrite, the fewer people are bidding, and the further the price drifts from the future it implies. That gap is the alpha.
Then there is the second half of the name.
Future as distribution - the spread of outcomes. Consensus capital prices the expected case and crowds the center of the distribution curve. We position for the tails - the low probability, high magnitude futures the crowd assigns zero odds and therefore gives away (almost) for free. That asymmetry is convexity that compounds across our portfolios.
Future as a time horizon. We back capabilities that do not yet exist at scale and that are at the cusp of transitioning from R&D to commercialization. The diffusion gap between military and commercial deployment has compressed from centuries to merely a decade, thereby fitting inside a single fund’s life of 10-15 years.
We underwrite hard problems and route productive capital towards the futures that the market has mis-priced. The difficulty and time horizons keep the crowd out and prices low.