Want to save on startup taxes? Start before you file

The short version:

• Most founders meet their tax bill in spring, as if it were weather. By the time you file, most of the decisions that materially affect your tax bill have already been made. Tax is decided during the year and at company formation. Filing just records the decisions you already made. 
• The biggest savings hide in windows that close early: the 83(b) election (30 days from getting your stock), your entity and its elections, and the R&D tax credit, which can hand even a pre-profit startup up to $500,000 a year back against its payroll taxes.
• Founders miss these not from laziness but from blindness. If your books are a shoebox you open in March, you can't see in September that a decision is on the table while you can still make it.

For a lot of founders, the tax bill arrives like weather. Sometime in spring a number shows up, larger than hoped, non-negotiable, apparently the product of forces beyond anyone's control. You pay it, wince, and resolve to think about taxes earlier next year. Then you don't.

Here's what the spring version hides: by the time you file, your tax bill is mostly already set. Taxes aren't weather. They're the sum of decisions, about your entity, your timing, your paperwork, your credits, that you made, or quietly failed to make, over the preceding twelve months and at the moment you started the company. Your accountant in April is often documenting decisions that were made months earlier, rather than creating new tax-saving opportunities. 

Which is a more hopeful idea than it first sounds, because it means the bill was never handed down by fate. It was decided. And decisions can be made better, if you make them while you still can.

Tax season is for reporting, not deciding

The mental model worth replacing is the one where taxes are a thing that happens to you at filing time. In reality, filing is the last and least strategic step. It documents a year that has already happened.

Almost every meaningful lever sits upstream of it. The choice of entity and its elections, the timing of income and expenses, the credits you're eligible for, the elections with their own hard deadlines, these all had to be handled during the tax year, or at the very start of the company, to count. By the time your return is due, the person who "saved on taxes" already did the saving months ago. Filing just writes it down.

That's the whole reframe, and it changes what you do with it. If tax is decided in real time and merely reported in retrospect, then the founders who pay less aren't the ones who found a cleverer accountant in April. They're the ones who could see a decision coming while there was still time to make it.

The windows that close before you file

The clearest way to feel the point is to look at where the real money is, for early-stage companies specifically. In almost every case, the deadline is earlier than founders expect, sometimes years earlier.

The 83(b) election, 30 days, no exceptions

When you receive founder stock or exercise options early, you usually have a 30-day window to file an 83(b) election with the IRS. Do it, and you're taxed on the value of your equity now, while it's near zero. Miss it, and you can end up owing ordinary income tax on your shares as they vest and appreciate, potentially an enormous, avoidable bill. In practice, missing the deadline is usually irreversible.. It's decided the month you get your stock, and the clock is unforgiving. 

Your entity and its elections. 

Whether you're an LLC, an S-corp, or a C-corp shapes your entire tax picture, and the elections that move between them are calendar-bound. A profitable LLC that would save on self-employment tax by electing S-corp treatment generally must elect S-corp status by March 15 (for calendar-year taxpayers), although limited late-election relief may sometimes be available. A startup planning to raise venture money will almost certainly need to be a Delaware C-corp, and doing that conversion sooner is cleaner, and unlocks other benefits that depend on timing. What you can't do is reach back after year-end and re-elect as if you'd chosen differently in January. 

The R&D tax credit, even with no profit. 

This is the one founders most often leave on the table. Qualified early-stage companies can apply the R&D credit against their payroll taxes, up to $500,000 a year, which means a pre-revenue startup burning cash on engineers can convert that research spend into real cash back, with no income-tax liability required. Recent laws also restored the immediate deduction of domestic R&D costs, removing a painful multi-year drag. But the credit only exists if you tracked qualifying expenses through the year and elect it on a timely-filed return. No records, no credit, and the paperwork is a during-the-year discipline, not an April scramble.

QSBS, the founder's long game. 

For C-corps, Section 1202 (Qualified Small Business Stock) can let founders and early investors exclude a large share of the gain when the company is eventually sold, recently expanded, with more generous caps and asset thresholds and partial benefits kicking in a few years before the full holding period. It can be one of the most valuable tax positions a founder ever holds. And it's set at the beginning: by whether you're a C-corp, and when the stock was issued, years before any exit is in view. A decision made, or missed, at incorporation.

Timing and deadlines, generally.

Estimated quarterly taxes cost you whether you overpay (you've lent the government money interest-free) or underpay (penalties). Accelerating or deferring an expense across December 31 can change the year it lands in. And for the growing number of founders operating across borders, foreign-related filings, the kind triggered by a US company with foreign owners, or a US founder with an overseas entity, carry steep penalties for being late, entirely separate from whether you owe any tax at all.

None of these are exotic. They're just early. Every one of them is decided before the return that reports it is ever filed.

Why founders miss them: they can't see their own numbers

The reason these windows close unnoticed usually isn't carelessness. It's a lack of visibility.

If your books are a shoebox of receipts you hand over in March, you literally cannot see, in September, that your income has crossed the line where an S-corp analysis pays off, or that this quarter's engineering spend is credit-eligible, or that your estimated payment is drifting. The decisions require knowing your position in something close to real time, and most founders simply don't. Not because they're not smart, but because the information isn't in front of them when the window is open.

This is the real argument for keeping your books current, and it's a very different argument from "avoid catch-up fees." Clean, monthly books aren't a compliance chore you do to keep the tax authority happy. They're the dashboard that tells you a decision is available while you can still act on it. The savings don't come from the bookkeeping itself. They come from the choices that bookkeeping lets you see in time to make.

What this means in practice

Treat tax as a year-round activity with a few decisive moments, not a spring event. Concretely, that means a handful of habits.

• Get the setup right early, because the setup is the expensive-to-reverse part: the entity, the 83(b) election, a clean cap table. These are cheap to do correctly at the start and painful or impossible to fix later.
• Keep your books current enough to answer one question at any time: what's my position right now? That single capability is what makes every other decision possible.
• Have the strategic conversations mid-year, not at filing, the S-corp analysis, the R&D eligibility review, the estimated-payment check, while you can still change the outcome.
• Know your hard deadlines, especially the unforgiving ones: the 83(b) 30-day window, quarterly estimates, entity elections, and any cross-border filings that apply to you.

One necessary caveat: this is general information, not tax advice. The rules change, several of the ones described here changed in 2025, and the right move depends entirely on your specifics. Run any of these decisions past a qualified tax professional before acting.

How this looks in practice

We spend our days on exactly this problem at Inkle, where we handle accounting, bookkeeping, tax, and compliance for US startups, many of them run by founders operating across borders, with a US entity on one side of the world and a team or a home country on the other. The pattern we see most is not founders making bad tax decisions. It's founders discovering, after the window closed, that a decision had been available at all.

Cross-border founders get a double dose of it, because they're juggling deadlines in two systems at once, each with its own elections and its own penalties for lateness. The fix is unglamorous and it's the same one for everybody: keep the books current and keep a tax view open year-round, so the levers show up while they can still be pulled. Done that way, tax stops being a springtime surprise and becomes what it always was underneath: a series of decisions you get to make on purpose.

The takeaway

You can't out-clever a number that's already fixed. By April, most of yours is. What you can do is decide well in the months when the decisions are still open, choose the entity, file the election, track the credit, set the money aside, so filing becomes a formality rather than a reckoning. The founders who pay less tax aren't luckier or better-represented at the deadline. They just refused to treat their tax bill as something that happens to them. It's a decision. Make it while you still can.

Frequently asked questions

How can a startup reduce its tax bill? 

Mostly through decisions made during the year and at formation, not at filing: choosing the right entity, filing an 83(b) election on time, claiming the R&D tax credit, timing income and expenses, and setting aside estimated taxes. By the time you file, most levers are locked, so the savings come from acting early, which requires knowing your numbers as you go.

Can a startup with no profit claim the R&D tax credit? 

Yes. Qualified early-stage companies can apply the R&D credit against their payroll taxes, up to $500,000 a year, so a pre-revenue startup with real engineering spend can get cash value even with zero income-tax liability. You have to track qualifying research expenses through the year and elect the credit on a timely-filed return, so clean records matter.

What is an 83(b) election and why does it matter? 

It's an election you can file within 30 days of receiving founder stock (or exercising options early) that lets you be taxed on the equity's value now, while it's low, rather than as it vests and appreciates. Missing the 30-day window can create a large, avoidable tax bill later, and there's no way to fix it after the fact. It's one of the most consequential deadlines in a startup's first month.

Should my startup be an LLC, S-corp, or C-corp? 

It depends on your plans. Many bootstrapped or profitable service businesses save on self-employment tax with an S-corp election. Startups raising venture capital almost always need to be a Delaware C-corp, which also opens the door to QSBS benefits. The elections have calendar deadlines and are hard to reverse, so decide early and with a professional rather than at tax time.

What is QSBS (Section 1202)? 

Qualified Small Business Stock is a provision that can let founders and early investors in a C-corp exclude a large share of their gain when the company is sold, provided the stock is held long enough and other conditions are met (the caps and thresholds were expanded in 2025). Crucially, eligibility is determined when the stock is issued and the company is formed, years before any exit, which is why it's a decision to get right at incorporation.

Does monthly bookkeeping actually save on taxes? 

Not directly, but it's what makes the tax-saving decisions possible. Current books let you see, during the year, that an election is worth making, that spend is credit-eligible, or that an estimate is off, while you can still act. Books you only assemble at filing time can record what happened but can't change it.