I'm afraid it doesn't even matter in this case. It can be anyone - with AI your ideas and expertise (which is what you're valued for - either as individual or a firm) are too easy to reproduce.
It's still years, and likely - decades for quantum computers to be used in practice.
There's hype in finance, too - but realistically, with proper developer tools, organisations can achieve over and above the required speeds using the traditional CPUs/GPUs - and without breaking the bank (no pun intended).
Even if one achieves ground-breaking results using an ultra-expensive chip, in no way it means that an average developer will achieve anything like that, even if provided access to cutting-edge quantum computers.
I think there's a lot hiding in the human-computer interface, and the true power of the existing CPU/GPUs is not fully utilised due to the inherent complexities of the programming languages. Needless to say that relatively simple hardware advancements such as vectorisation and multithreading, are rarely used to full extent in popular languages such as C++ or Python.
My bet is that the ROI of any Quantum Development will be negative for decades. So, unless one's planning horison is 10+ years, or you simply don't care for ROI, I'd choose something more mainstream such as CPU/GPU/TPU...
We're working on a specialised graph compiler for speeding up simulations and computing derivatives automatically (backpropagation/adjoint differentiation). It now supports C++, Python, and C#; AVX2 and AVX512 instruction sets, multithreading; Windows, and Linux.
Essentially, it allows model developers (such as quants in finance, engineers, and ML specialists) to code in Python without needing to think about the performance of repetitive calculations. As we all know, Python is one of the least efficient languages when it comes to complex calculations/simulations - and we help to resolve it. Long story short, with very few tweaks to the code certain types of calculations (such as pricing of derivatives, curve building, computing financial risks, or "small" NNs) can be accelerated by 100+ in Python and x20+ in C++/C#.
We're now looking to add support for Java (but it doesn't have Operator Overloading, so it's tricky), and some customers are asking to support GPU - which is a bit tricky because it's got a closed instruction set.
But I will share here when we're clear sharing :)