in 2026, a clean down round is no longer a death sentence. half the series b cohort that closed between january and april raised below their last mark. nobody flinched. the cap table moved, the company kept building, the next round will be priced off arr not off vintage.
the outcome founders should actually fear is the flat round with structure. same headline valuation as the last raise — and a 2x participating preference, a ratchet, or a pay-to-play buried in the side letter. on the announcement it looks like a win. on the exit waterfall it pays the new investor before common sees a dollar.
founders sign these because the headline number protects the story. the cost shows up two years later, at the moment the founder thinks they're getting wealthy.
the math behind the warning
start with two companies, same last round, same exit. one takes a clean 25% down round. the other takes a flat round with 2x participating preference. the difference at exit is not subtle.
company a raises $40M at $160M pre, 25% off the last $200M. clean terms — 1x non-participating preference, single trigger acceleration, standard. company b raises $50M flat at the $200M mark of the last round. inside the deal — 2x participating preference, full ratchet, pay-to-play on the next round. headline valuation: $200M either way. the spread shows up on the day someone writes the check that ends the company's life as an independent entity.
at a $400M acquisition, company a's new investor takes their $40M back at 1x, common gets the remaining $360M to split with the rest of the preference stack. at company b, the $50M comes off at 2x — that's $100M — and the new investor also participates pro-rata in the remaining $300M as if it were common. on a 25% post-money stake, that's another $75M. the flat-round investor walks with $175M off a $400M exit. common collects $90M less than the down-round version. the headline that protected the story cost the team a house each.
why founders take the flat round anyway
three reasons, in order of how often we see them.
one. the lead pitched the valuation as the win. "we held the line at $200M when half the market is down 30%." it sounds true. the founder doesn't realize the lead is buying optionality on the downside — the participating preference and ratchet are how the lead gets paid if the line doesn't hold. when the lead frames the cap as a victory, the lead is naming the lever the founder is allowed to look at.
two. the term sheet looks normal because the founder hasn't seen a participating preference up close. 1x non-participating is the default at series a. a participating term shows up at series b and reads like a small variation. the founder reads it and thinks fairness. it is not fairness. it is double-counting, and a 2x version is double-counting twice.
three. the founder compares the term sheet to the previous round, not to clean down rounds happening that quarter. brex did a 50% cut in 2024. ramp closed at 28% below their peak in 2025. neither company collapsed. the comparison set for "fair outcome" is the down-round-with-clean-terms group, not the imagined disaster scenario the lead is talking the founder out of.
the named pattern
the most expensive version we have watched in person, details changed: a series b ai company in late 2024, $80M raised at a $400M flat valuation. structure — 2x participating preference, 1.5x cumulative dividend, full ratchet. the founder felt great on announcement day. andreessen had passed, sequoia had passed, and a top-tier multi-stage fund had agreed to hold the cap.
eighteen months later the company exited to a strategic at $520M. on paper, a 30% return. on the waterfall, the series b fund collected $160M from the $80M check — 2x preference plus participation, plus accrued dividend. founders and early team split $90M instead of the $220M the same exit would have paid with clean terms. the announcement quote — "we maintained our valuation through a difficult market" — was true. it cost the team $130M.
what to do when the flat-with-structure offer arrives
three rules. apply them before the cap question.
rule one — strip the structure before negotiating the valuation. ask the lead directly: 1x non-participating, no ratchet, no pay-to-play. if those three terms are not standard, the cap is a misdirection. solve for the waterfall first, then negotiate the headline.
rule two — model the exit, not the cap. at a 2x return, 3x return, 5x return, what does common get under each scenario? a 2x participating preference looks small until you draw the line at a 2x exit, which is the most likely outcome statistically. common gets crushed at exactly the multiple most companies actually hit.
rule three — accept the down round when it is the cleaner deal. a 25-35% down round with 1x non-participating preference is almost always better than a flat round with any structure. the cap moves, the cap table stays clean, the next round prices off the new mark with no overhang. the team has 0% of their net worth gated behind a waterfall trap.
how zift handles this
zift doesn't price your round, but it tells you exactly how many weeks of runway you have when the term sheet arrives. every monday morning, the briefing surfaces cash on hand, current burn, and the date you run out — so when a lead pushes a flat-with-structure offer at the last minute, you know whether you're negotiating from twelve weeks of cash or four. that delta is the difference between walking and signing.
if you're a finance lead at a series a or b team modeling the post-money cap table against multiple exit scenarios, zift handles the runway side — carta handles the waterfall math itself.
the cap is the headline. the structure is the price. founders who survive the next exit aren't the ones who held the line on valuation — they're the ones who read the side letter.