Heading back from a great trip to San Francisco. Something that invariably came up was the curious proliferation of 'trading neolabs' trying to cash in on the bubble in venture capital now.
While I've written on this topic before, I mostly think this is not a venture-type business.
Quant trading firms are some of the most profitable businesses out there, BUT a trading strategy is not a 'product'.
It does not benefit from the magic of sticky retention, viral word-of-mouth referrals, or annual recurring contracts.
Systematic strategies have practically no moat at all, other than non-competes and regulatory capture. They are easily ported firm-to-firm and die quickly if not constantly adapted. Like Sisyphus, the systematic PM is confined to pushing a rock up a hill.
Also, dare I say it, the TAM of systematic trading is not that high compared to most big software verticals. Even within the world of trading, fully systematic strategies account for a relatively small fraction of pnl. I recall digging into Citadel's commodities business (almost entirely discretionary) when I was interviewing there and the numbers were jaw-dropping to someone coming from HFT.
The most scarce skill quants have is that of fitting massive models on (relatively) small data. Unlike in eg weather prediction, market data cannot simply be 'generated' by eg numerical solutions of PDEs; it is capped by time.
I am sure there are problems these skills can be applied to. Cancer research just popped into my mind (limited # of cancer patients). Fellow ex-quants, please spend your time here! An industry with real TAM.