Agents will further entrench the incumbents in quantitative trading
Quantitative hedge funds make software designed to produce uncorrelated return streams.
Coding agents change the cost structure of that software, making it cheaper to build and maintain what you need, from signal construction, portfolio optimization and simulation, to execution and order management.
Likewise, research agents — who can design signals, write data pipelines, build and test statistical models — will lower the cost of trying new ideas. The whole research process will become more scalable. Time-to-market for new signals will go down.
So overall the effect will be lower R&D costs.
Does this provide for a good opening for startup managers? I believe not. Even if it’s now an order of magnitude cheaper and faster to write the software and to do the research, defensibility in this business ultimately comes from scale. Larger funds get:
Cheaper and faster execution
Lower cost of capital/more leverage
More data, sometimes buying exclusivity
Preferential market access and better connectivity
Talent, and again paying for exclusivity in a way
Brand and trust signal
Agents don’t change anything about those advantages, but there are a couple of counterpoints worth considering.
First, if the cost of R&D drops by an order of magnitude, some capacity-constrained strategies become worth trading that previously weren't worth the headcount. That makes for some kind of opening for smaller managers, for whom capacity constraints are less of an issue. But the drop in cost applies to incumbents as well, and if they’re good at combining strategies at scale, they’re in a better position to harvest the margin increase on a larger asset base.
A second counterpoint is institutional inertia: Incumbents may fail to adopt the new technology and their competitive edge in engineering and research will erode over time. For sure, there’s always some amount of drag from existing code, processes, integrations, etc., and in my experience this is more significant than one would expect in a tech-forward sector like quant trading. But what I’ve also observed is that it takes an extremely long time for large, ossified players to disappear. Even if their tech stacks are known to be outdated and it’s gotten hard to attract talent, many of those funds still manage billions years later thanks to the above mentioned scale advantages.
Big picture, the one thing that’s for sure is that agents will make markets more efficient. The time from new data or new idea to trading live in production will come down meaningfully over time, so any new effects will be arb’d away more quickly.
