How much to raise at seed: the 18-month runway math

How much to raise at seed comes down to one number: the cash it takes to hit a credible Series A in 18 months. Work backward from the Series A bar, divide by your real monthly burn, add a 4-month buffer for the next raise. For most 2026 software seeds that math lands between $2M and $5M.

Most seed-stage advice tells you to raise 18 months of runway. That's correct, but it's the wrong starting point. The right starting point is the Series A bar you have to clear, because that's the milestone the seed is buying you a shot at. Everything else, runway length, burn rate, team size, falls out of that one anchor.

This guide is the worked-backward math: Series A benchmarks, then the 18-month plan, then burn, then seed raise amount. Plus the two failure modes, raising too much and raising too little, that decide most outcomes.

How to size your seed round in 5 steps

Use this as the seed round size worksheet. Each step has one input and one output.

1. Define the Series A bar for your sector. B2B SaaS in 2026: roughly $1M–$2M ARR with strong net retention. Consumer: a defensible engagement metric at meaningful scale. AI infra: design partners plus a measurable performance edge. Talk to 5 Series A partners before you pick the number.
2. Estimate the team needed to hit it. Be honest about which 4–8 hires actually move the metric. Not the org chart you want at Series A; the smallest team that ships the proof.
3. Calculate fully loaded monthly burn. Salaries plus 25% for benefits and taxes, plus tools, infra, rent, legal. US engineers cost $25k–$35k/month all-in. Burn is almost always 80%+ payroll.
4. Multiply burn by 18 months. That's your operating need. Add a 10–15% contingency for the things you forgot.
5. Add a fundraising buffer. Series A processes take 4–6 months; you'll start with 8–10 months of runway left. Sanity check: at month 10, do you still have enough to walk away from a bad term sheet?

If steps 1 through 5 land you above $5M and you're not in AI infra, your team plan is too big for a seed. Cut headcount before raising more.

Work backward from the Series A bar, not forward from your wishlist

The wrong question is "how much can I raise." The right question is how much money to raise seed so you can credibly clear the next gate.

Series A graduation rates in 2025 split sharply by signal quality. Roughly 53% of "consensus" top-decile seed deals raised a Series A, versus 28% for non-consensus deals. The gap isn't about raising more cash. It's about hitting metrics that read as inevitable to the next investor.

So the math is: pick the metric, scope the team that can hit it, fund that team for 18 months. Anything else is over- or under-capitalized.

These are rough anchors, not promises. The point is the direction: bar first, raise second.

The 18-month runway math, in numbers

The 18 month runway target exists because the gap between rounds keeps getting longer. Startups raising a Series A in Q4 2024 waited a median of 774 days, roughly 2.1 years, between primary funding rounds. If you only fund 12 months, you're starting the next raise the day you close.

A worked example for a typical US B2B software seed:

Team at month 18:
  2 founders                  $20k/mo (below market, on purpose)
  4 engineers                 $120k/mo
  1 designer                  $25k/mo
  1 founding AE / GTM         $25k/mo
                              -------
  Payroll subtotal            $190k/mo
  Benefits + taxes (25%)      $47.5k/mo
  Tools, infra, rent, legal   $25k/mo
                              -------
  Fully loaded burn           ~$262k/mo

Operating need (18 mo):       ~$4.7M
Contingency (10%):            ~$0.5M
                              -------
Round size                    ~$5.2M

This doesn't mean every seed should be $5M. It means if you want this team for 18 months, this is the cheque. Smaller team, smaller cheque. The math doesn't care about your ambition; it cares about your headcount.

For context: the median US seed round was $3.8M in Q4 2025 at a median pre-money valuation of $16M. At those numbers, a $3.8M round on a $16M pre is roughly 19% dilution, which lines up with the ~20% ownership consensus seed founders typically give up.

Failure mode 1: raising too much (team bloat)

Don't raise the biggest cheque you can close. Raise the smallest cheque that funds the plan.

Excess capital at seed almost always converts to headcount. Headcount you can't reverse without firing people, breaking the team, and signaling distress to the next round. A 12-person seed-stage team burning $400k/month needs to clear a much higher Series A bar than the same company with 7 people burning $200k/month, because the next round has to fund the bigger entity at higher prices.

You also raise the bar on yourself. A $6M seed at a $25M post means the Series A has to mark up against $25M. A $3M seed at a $15M post leaves room for a clean 2x markup at Series A on real progress. Higher seed valuations compress the markup space and tighten the metric you have to hit.

The version of this failure that hurts most: raising $6M, hiring 10 people in 9 months, missing the metric, then trying to raise a Series A as a "seed-extension story" at flat valuation. That's the worst of both worlds.

Failure mode 2: raising too little (bridge round territory)

The opposite mistake is rationing yourself into a bridge. Bridge rounds accounted for roughly 40% of all seed-stage raising activity in 2024, almost always done at flat or down terms with existing investors carrying the company.

A bridge isn't a financing event; it's a confession. It tells the next investor you under-scoped the seed, missed a milestone, or both. Even a "clean" bridge at the same cap drags the cap-table conversation into your Series A pitch.

Most seeds that need a bridge needed a $1M–$1.5M bigger primary round. Pay the dilution upfront; bridges cost more in narrative damage than they ever save in equity.

If your honest 18-month plan needs $4M and you're sizing the round at $2.5M because that's "what you can close," either cut the team plan to fit or push the round until you can credibly raise the right number.

When to start the Series A process

Plan to start fundraising when you have 10 months of runway left, per Kruze Consulting's burn-rate guidance. That gives you 4–6 months to run the process and 4–5 months of post-close runway as a buffer.

✅ Good: Closed $3.5M seed in March, hire plan funded through August next year. Start Series A conversations in November, target close by April.

❌ Bad: Closed $3.5M seed in March, start Series A conversations the following October when you're at 6 months runway. Now every partner can smell the timeline and you're negotiating from weakness.

If you're sending more than 30 investor outreach emails a week during the Series A push, tools like Causo handle the personalization and follow-up cadence. For lower volume, a spreadsheet works.

FAQ

How much should you raise at seed? Raise enough to hit a credible Series A milestone with 18 months of runway. For most US software seeds in 2026 that lands between $2M and $5M, sized backward from your monthly burn at the team you actually need to ship the metric, not the team you'd like to have.

How long should seed runway be? Plan for 18 months of operating runway after the round closes, then start fundraising the Series A at the 10-month mark. That gives you 8 months to build the metrics, plus a 4–6 month process buffer before cash gets tight enough to weaken your negotiating position.

Is it better to raise more or less at seed? Less, until you have proof points. Raising more than you need at seed inflates burn, locks you into hiring you can't reverse, and raises the Series A bar you have to clear. Raising slightly under and topping up via a SAFE extension preserves optionality.

What's the average seed round size in 2026? The median US seed round was $3.8M as of Q4 2025 per the PitchBook–NVCA Venture Monitor, with the median seed pre-money valuation at $16M per Carta. AI-heavy rounds skew larger; non-AI software seeds tend to cluster $2M–$4M.