Startup friendly banks, by the numbers: the SVB successor scores 8.0 for startups and 4.5 overall, and the reason is a price list nobody publishes

I went looking for startup friendly banks the way a first finance hire usually does, which is by asking who banked the last company. Then I tried something slower, and more boring. Bank Index, a public directory that scores banks and account providers from 0 to 10 on what a business needs, has a line on every hand-checked card called Startups and VC-backed companies. I read that line on 8 cards, and then the overall score beside it.

The two numbers disagree, often by a lot. The bank that took over Silicon Valley Bank's business scores 8.0 for startups and 4.5 overall. I had assumed a bank that good at one thing would at least be average at the rest. I was wrong, and the reason is worth a founder's ten minutes before the next wire lands.

The startup line, top to bottom

Mercury leads with 10. Its card quotes the company's own annual letter: one in three US startups use it, and it onboarded 2.4 times more AI companies in 2025 than in 2024, alongside venture debt and an investor network it calls Mercury Raise.

Then come 2 banks at 8.0. The SVB division of First Citizens Bank is described on its card as “dedicated to technology, life-science and venture-backed companies, with 39 offices in 15 states”, with venture debt and Global Fund Banking carried over from the original Silicon Valley Bank. CIBC runs CIBC Innovation Banking, with venture debt and growth capital and over 11 billion dollars in funds managed across North America, the UK and parts of Europe.

A larger group sits just below, at 7.5. Western Alliance Bank has an innovation banking practice under the Bridge Bank brand. Customers Bank has a Tech and Venture group offering term loans, ARR credit facilities and capital call lines for venture funds. JPMorgan Chase has its Innovation Economy unit, with 550 or more bankers advising 11,000 or more companies from pre-seed. HSBC has HSBC Innovation Banking, the former SVB UK, serving venture-backed companies in the UK, Sweden and Denmark. Capital One gets its 7.5 through Brex, which its card says it acquired for 5.15 billion dollars in April 2026.

What struck me, reading these reasons in a row, is how little they have in common. Mercury's is about volume and self-service. The First Citizens reason is about a division, its offices and fund banking. Capital One's is about a company it bought. The same number stands for three different kinds of help, and a founder reading only the number would never see it.

That is a respectable list of venture banking desks. Now look at the overall scores next to it.

Why the SVB successor scores 4.5

The First Citizens card explains its own number in one line: the average of 13 scored needs, all 13 backed by filings or the bank's own pages. So I did the average by hand. Strength 7.0, startups 8.0, cross-border 5.0, customer satisfaction 5.0, business cards 4.5, fintech partners 4.0, online businesses 4.0, trade 4.0, fees 3.5, interest 3.5, onboarding 3.5, non-resident owners 3.5, high-risk sectors 3.0. That adds to 58.5, and 58.5 divided by 13 is 4.5.

The low lines are not insults. They are sentences about what the bank does not publish. On fees the card says “No public business account fee schedule was found on the main site.” On onboarding it says “No self-service online application or stated opening timeline for a general business account is published.”

I find that oddly familiar, and it annoyed me slightly, because I wanted the number and the card was right not to invent one. It is how relationship banks have always worked: a banker, a term sheet, a price you learn in a meeting. The index scores what a company can check from outside, and a price you only learn in a meeting scores as a price nobody can check.

The card also records what makes it a real bank in the old sense. Its strength line gives investment-grade ratings of BBB+ at S&P and Baa2 at Moody's and a common equity tier 1 ratio of 11.15 per cent at year-end 2025. The summary says the bank bought SVB's deposits and loans from the FDIC in March 2023, and did not buy the UK subsidiary, which went to HSBC separately.

The same pattern, four more times

Western Alliance scores 7.5 for startups and 4.7 overall. Its strength line is solid: common equity tier 1 of 11.0 per cent, total capital of 14.1 per cent of risk-weighted assets and 98.7 billion dollars of assets at 30 June 2026. It also notes nonaccrual loans rising to 0.92 per cent of funded loans, which is the sort of detail a treasury policy should want to know about.

Customers Bank scores 7.5 and 4.8 overall. Its strength line scores 6.0, and the reason is two facts side by side: KBRA rates the bank BBB+ with a positive outlook set in December 2025, and an August 2024 Federal Reserve written agreement “still carries no termination date.”

I stopped on that line for longer than I expected. A written agreement is a supervisory action made public, and the card puts it next to a positive rating outlook without trying to reconcile the two. I like that choice a lot. A sales deck would have picked one of them and dropped the other. For a finance hire the pair is more useful than either on its own: the rating says what the market thinks, and the agreement says what the supervisor is still waiting for.

Capital One also sits at 7.5 and 4.8 overall, with strength at 6.5 on a Baa1 rating and a common equity tier 1 ratio of 13.7 per cent at 30 June 2026. CIBC does better overall at 5.4, with strength at 8.0 on ratings of AA, Aa2, A+ and AA+, though the same line notes 42 million dollars of US record-keeping penalties paid in 2024.

JPMorgan Chase and HSBC are the exceptions. They score 7.5 for startups and 6.1 and 6.8 overall, because they score well on nearly everything else, with strength at 10 and 9.5.

So the honest summary of venture banking in the index is this. The specialists are exactly what the name says. They lend against ARR and they bank funds, and, on the First Citizens card at least, they publish almost nothing a stranger can price.

A short digression about the word friendly

An aside, because the phrase in every search box is startup friendly, and friendly has nothing to do with any of these scores. The index does not measure warmth. It measures whether a product exists for venture-backed companies, and whether the terms are public.

Mercury is friendly in the sense that you can open an account on a Sunday. First Citizens is friendly in the sense that it offers venture debt and fund banking to companies that have raised from investors. Those are different favours, and a company at seed often needs the first long before it needs the second. Back to the numbers.

I have some sympathy for the word, though. Founders reach for it because the alternative is to list what they need, and that list is long and dull: an account that opens online, a card with sensible limits, someone who answers when a wire is stuck, and later a lender who understands recurring revenue. The index splits that list into separate lines so that it can be read one line at a time, which is less satisfying than a single word and a great deal more useful.

The line that decides where the round sits

Every card also scores financial strength, and for a company holding a round this is the line that matters after SVB. Mercury leads the startup line and scores 4.5 for strength, because it is not a chartered bank and carries no agency rating. First Citizens scores 7.0, HSBC 9.5 and JPMorgan a full 10.

I suspect most seed companies only read the strength line after a scare. It is the dullest line on the card and the one that matters most on the worst day of the year, and I would rather a finance hire read it on an ordinary Tuesday, when nothing is on fire and the answer can still change where the money goes.

Read together, the two lines draw the choice most seed companies actually face. The easiest account to open and the strongest balance sheet are rarely the same institution, which is one reason to split the money. Our piece on how much cash to keep in one bank covers the concentration side, and the one on checking a bank's financial health covers the filings behind the strength line.

If you use the directory for this, the front page of Bank Index lets you pick the needs that matter and recalculates the ranking for them. The site says that when you select needs, the score becomes the plain average of the ones you picked, so I ran the arithmetic on these 8 cards. Startups and strength together put JPMorgan first at 8.75, HSBC at 8.5, CIBC at 8.0 and First Citizens at 7.5, with Mercury down at 7.25. Startups and onboarding flip it: Mercury leads at 9.25, and First Citizens falls to 5.75.

It surprised me how completely the order turns over. Nothing about any bank changed between those two runs, only the question, and the question is the part a founder controls. That seems obvious written down. It did not feel obvious while I was staring at a single overall number and trying to decide whether a bank I had heard of for decades was suddenly a bad one.

Open the cards themselves after that. The overall is an average of 13 needs, and a seed company has perhaps 3 of them. I would read the startup line, the strength line and the fees line, and ignore the rest until the company has a reason to care about trade finance.

What I still cannot tell you

I cannot tell you what First Citizens charges a Series A company for a business account, because the card could not find it and I have not found it either. The SVB division may well price differently from the retail bank. That is a guess, and nothing public settles it.

I also cannot tell you whether the Capital One and Brex line will hold its 7.5 once the deal settles in, and it is a strange thing to be unsure about, because an acquisition like that either keeps what made the product good for startups or quietly turns it into a feature of a much larger bank's card business. The card dates the acquisition to April 2026. I suspect the score will move within a year, one way or the other, and I would bet on the product changing before the score does.

I keep thinking about the gap between 8.0 and 4.5. It is the distance between a banker who understands your cap table and a price list you are not allowed to see. It is the trade Silicon Valley Bank offered for decades, and the one March 2023 taught founders to price.