Not to single you out JohnFen, but I'm seeing this take in many places, and it is very wrong. The risk here is being attributed to the wrong place. A startup checking account should be low risk, it is just plain cash as it comes in from sales and/or funding round investments. No derivatives, no funky illiquid stuff, just cash. The activities of the startup _may_ be risky, and it can totally fail because of that...by running out of cash, that should be safely and easily available in a deposit account. A startups cash can hit zero, but they expect to be able to access it so that it can hit zero!
So the argument that by the mere fact SVB banked startups as customers was more risky than big banks is not quite precisely right, as big banks have business accounts and plenty of startup customers too. But as other comments have mentioned things were often made quite hard for them there, and SVB was the better product to startups...as just a literal checking account for day to day operations.
The giant risk management failure was in how SVB managed the deposits and indeed the mismatch of the assets it held against them. We all know the long dated UST issue now. Their deposits ballooned so quickly over the last 18mo, the mistake was to put them into such long locked up bonds immediately. A better approach could have been to assume that money could "come out as fast as it came in", and to have held shorter duration securities, shorter maturity bonds, money market funds, repo market commercial paper etc. Then as the steady state pattern emerges post this influx of deposits, then make a better risk management based decision of what proportion to now put in longer dated assets, medium dated assets and shorter dated ones.
And above this specific risk, clearly in hindsight, there was an overall systematic risk in having a non-diversified customer base who all consume the same information sources, highly networked and correlated with each other in their behaviour.