To avoid burning hundreds of thousands of dollars on a startup, definitely do one thing: validate the idea first.
The main trap
A startup solving someone's problem doesn't mean the whole market needs it. This is exactly where many founders go wrong: they fall in love with their idea and assume everyone wants it.
So before starting, bring the project's full technical spec, ideology, and what solutions it provides into a presentation. Then consult a strong marketer, business consultant, or analyst.
The best solution
The best path is to talk with a marketer, explain the project fully, and ask them to analyze the market for you. With an outside, unbiased view, they assess the idea's fit to the market.
If that's not possible — do cust-dev
If that's not possible, do cust-dev yourself: hold dedicated interviews with at least 20 people you consider your target audience. Find out how they feel the problem, how they solve it now, and whether they'd pay for your solution.
In any case, it's better to validate first than to take hundreds of thousands from a fund or the state budget and then fail. Even if the money is your own, this significantly reduces the risks.