Startup Banking: Where to Actually Put the Money After You Raise

    FDIC insurance stops at $250,000 per bank. Sweep networks stretch that to several million. Here is how founders should structure cash after a round closes.

    Financial Operations · 7 min read

    Jacob Sheldon ·

    The week a round closes, a founder gets a wire for several million dollars into an account insured up to $250,000. Nobody at the closing mentions this. It is not a crisis, and it is also not nothing, and March 2023 taught an entire generation of founders that concentration risk is a real thing that happens to real companies on a Thursday.

    The decision is not really which bank. It is how to structure cash across accounts so that no single failure, freeze or compromised login can reach all of it, while the operating account stays boring enough that payroll always runs.

    The short version

    • FDIC insurance is $250,000 per depositor, per insured bank, per ownership category. That number does not scale with your balance.
    • Fintech platforms stretch coverage by sweeping deposits across partner banks. Published figures were about $5M at Mercury and $6M at Brex as of early 2026.
    • Uninvested cash usually earns zero. Moving idle cash into a treasury product is the entire yield.
    • A money market fund is not FDIC insured. It holds government paper instead, which is a different risk, not the same one.
    • Split the money: operating account with three to six months of burn, treasury with the rest.

    Start with the insurance math, because it is the only hard number

    $250,000 per depositor, per insured bank, per ownership category. A $4 million seed round in one business checking account at one bank is insured for $250,000 of it. The other $3.75 million is an unsecured claim on the bank if the bank fails.

    That is the mechanic behind every sweep network you have seen advertised. Your platform holds your deposit across a set of program banks, each with its own $250,000 limit, and multiplies coverage by the number of banks. Mercury published up to $5 million in sweep coverage through partner banks including Choice Financial Group, Column N.A. and Evolve Bank & Trust. Brex Vault published up to $6 million. Both figures were current in early 2026 and both are the kind of thing that changes, so read the current disclosure rather than this paragraph.

    Two things founders misread about sweep coverage. It is pass-through coverage at the partner banks, so it depends on records being maintained correctly to identify you as the beneficial owner. And it is coverage against a bank failing, not against your account being frozen during an investigation, a dispute or a compliance review. Those are different failure modes and the second one is far more common.

    The structure that survives a bad week

    Most startups past a seed round land on three buckets, and the reasoning is the same at every size.

    Operating account: three to six months of burn. This is the account that pays vendors, receives revenue and funds payroll. It should be boring, at a platform with a reliable feed, and it should never hold the whole balance. Three to six months is enough that a treasury transfer delay never threatens payroll.

    Treasury: everything else. Cash beyond the operating buffer sits in a government money market fund or a Treasury ladder. This is where the yield is, and the yield is not decorative. At mid-3 percent published yields in early 2026, $2 million idle for a year is roughly $70,000. That is a junior hire, and it is the single easiest financial decision a post-raise founder makes.

    A second institution. Not a second account at the same platform, a second institution entirely. It does not need much in it. It needs to exist, with signatories set up and a small balance, so that if your primary is frozen on a Tuesday you have somewhere to receive a wire on Wednesday. Founders who lived through March 2023 all learned this the same way.

    If your payroll provider supports a dedicated funding account, use one. It makes payroll failures loud and early rather than silent and late.

    The FDIC and money market distinction, said plainly

    A money market fund is not a bank deposit. FDIC insurance does not apply to it, and any page that implies otherwise is being careless.

    What a government money market fund holds instead is short-dated US Treasury securities and repurchase agreements backed by them. The credit risk on that is about as low as financial risk gets, which is exactly why it became the standard home for startup treasury cash. But it is a securities position, and it can in principle break the buck, and the account it sits in is a brokerage account with different protections than a bank account.

    The practical implication is not "avoid it". It is "know which of your two risks you are holding". Deposits carry bank failure risk mitigated by FDIC insurance. Money market funds carry securities risk mitigated by holding government paper. Splitting across both is not indecision, it is the point.

    Choosing between the platforms

    Once the structure is settled, the platform choice is narrower than the marketing suggests.

    Consideration What to actually look for
    Sweep coverage The published limit, and which partner banks. Both change.
    Yield on idle cash Usually zero. The yield is on the treasury product, not the checking balance.
    Bundled card and spend Brex bundles corporate card and expense management. Mercury does not to the same degree.
    API and feeds Whether accounting tools connect natively and whether that connection stays up.
    Wire and ACH limits The number that bites first when you close a large vendor deal.
    Who holds the deposit Whether the platform is a bank or a program on top of partner banks.

    The short characterization that holds up: Mercury is the cleaner banking product with strong programmatic access, and Brex is the better fit if you would otherwise bolt on a separate corporate card and expense tool. Both are reasonable. Neither choice is worth three weeks of deliberation, and both are reversible in a way the structure decision above is not.

    The criterion nobody weighs enough

    Ask how the bank's data reaches your accounting system, and whether that connection stays up.

    This sounds like an IT question and it is actually the difference between books that are current and books that are three weeks stale. A stable direct feed means every transaction lands in your ledger the day after it clears, gets categorized, and a person reviews the exceptions. A flaky connection means someone downloads a CSV, uploads it, reconciles by hand, and stops doing it consistently by month four.

    One founder ran two accounts across two institutions for eighteen months. One had a native feed, one required a monthly export. At diligence, the exported account had four months of gaps and a $31,000 unreconciled difference that took a week to unpick. The bank was fine. The feed was the problem, and it had been the problem the whole time.

    Median connects Mercury, Brex, Ramp, Stripe and the major bank feeds directly, categorizes and posts activity every business day, and an accountant reviews what the automation flags. That works well when the feed is reliable and badly when it is not, which is why the feed belongs in the bank decision rather than after it.

    Frequently asked questions

    How much of my startup's cash is actually FDIC insured? $250,000 per depositor, per insured bank, per ownership category. Sweep networks stretch that across partner banks, which is where multi-million coverage figures come from.

    Should a startup keep all its cash in one account? No. Operating account with three to six months of burn, treasury with the rest, and a second institution that exists before you need it.

    Is a money market fund FDIC insured? No. It holds government securities rather than deposits. Low risk, but a different risk.

    What yield should a startup expect on idle cash in 2026? Published startup treasury yields sat in the mid-3 percent range in early 2026 and rates move. Uninvested balances typically earn zero, so the gap between doing nothing and doing something is the whole yield.

    Does the bank I choose affect my bookkeeping? Yes, more than founders expect. A stable feed keeps books current. A broken one means manual CSV work that stops happening by month four.

    Do this in the first week after a raise

    1. Write down your monthly burn. Multiply by four. That is your operating account target.
    2. Move everything above it into a government money market or Treasury product. Do it this week, not next quarter.
    3. Open an account at a second institution with signatories in place. It can hold $5,000. It needs to exist.
    4. Confirm every account has a working direct feed into your accounting system before you stop paying attention.

    See how Median keeps books current from those feeds, every business day rather than every month.

    Frequently asked questions

    $250,000 per depositor, per insured bank, per ownership category. That is the base number and it does not change because you are a company or because the balance is larger. Fintech banking platforms stretch it by sweeping your deposits across a network of partner banks, each carrying its own $250,000 limit, which is how coverage figures of several million dollars are produced. The coverage is real but it is pass-through coverage at the partner banks, not coverage of the platform itself.

    No, and the reason is operational as much as it is about insurance. A workable structure is an operating account holding roughly three to six months of burn, a treasury or money market position holding the rest, and a separate payroll account if your provider supports it. That way a compromised operating account, a frozen card, or a bank problem does not put the entire runway behind one login.

    No. A government money market fund holds Treasury securities and repurchase agreements, and it is not a bank deposit, so FDIC insurance does not apply. What it has instead is the credit quality of short-dated US government paper, which is why it is the standard place for startup treasury cash. Understand that difference before you move a round into one; the risk is small but it is a different risk, not the same one.

    Published startup treasury yields sat in the mid-3 percent range in early 2026, and rates move, so treat any specific number as stale the moment you read it. The more useful rule is that uninvested balances typically earn zero, so the difference between doing nothing and moving cash into a treasury product is the entire yield. On $2 million that is roughly $70,000 a year, which is a real hire.

    Yes, and it is the most underrated criterion in the decision. A bank with a stable direct feed and a real API delivers clean transaction data to your accounting system every day. A bank whose connection breaks monthly means someone is uploading CSVs and your books go stale between fixes. Ask any bank you are considering how its feed connects to accounting tools, and treat a broken feed as a recurring cost rather than an annoyance.

    The year-end close checklist

    Everything to reconcile, recognize and document before you file, as a printable PDF plus a week-by-week timeline you can reuse every year.

    We'll only email you about this. Unsubscribe anytime.

    Find out what is actually wrong with your books.

    Share accountant access to QuickBooks or Xero 48 hours ahead and we will spend twenty minutes going through what is broken, ranked, with what each one costs you. You keep the findings whether or not you work with us.

    Book the teardown