Largely, though I receive 1 or 2 job specs every week for start ups with the keywords 'hft' and 'low latency'. Admittedly there's going to be duplication there if you read them closely.
I think it's a bit of a myth that (ignoring FPGAs) that writing a low-latency software trading system is a time/cost expensive process. Anecdata = I worked at two firms where we did a rewrite from scratch with teams of 5-6 people and traded in the market within 3 months. I'd argue a senior dev that's been around the block a few times could achieve similar when you remove corporate politics, and bikeshedding over design.
The big firm part is paying for multiple quants at $200k++ to come up with strategies and historic market data access for trading models. Small firms are getting backing as long as the co-founders are 'ex-CxO from MegaCorp'.