The H1 2026 Down Rounds and Recaps Report

Down rounds fell to under 14% of new fundings on Carta in Q4 2025, while PitchBook recorded 15.9% YTD and SVB reported roughly half of US VC-backed tech companies sat below their last private mark in 2025. The math has shifted: pricing down is now a market condition, not a verdict on your company.

• How common are down rounds in H1 2026
• Down round vs flat round 2026 vs up round
• The AI/ML skew distorting 2026 down round data
• Down round terms that actually hurt
• Pay-to-play, structured deals, and ratchets
• Startup recapitalization: when a recap beats a clean down round
• Extension rounds and tender offers as down round alternatives
• The H1 2026 founder playbook for walking through a down round
• What this means for your raise
• FAQ

H1 2026 finally killed the old story on down rounds. The headline number is not 17% from Carta Q3 2025 or 15.9% from PitchBook YTD. It is the ~50% of US VC-backed tech companies whose 2025 valuation sat below their last private mark, per SVB's State of the Markets H1 2026. Half the ecosystem is at or below water. Treating your priced down round in 2026 the way founders treated it in 2021 is the single most expensive cognitive mistake a Seed or Series A founder can make this year.

The rest of this report is the operational guide. How often down rounds actually happen versus extensions and recaps. Which terms incoming investors will push when they sense leverage. How recaps and tender offers work as alternatives. And the H1 2026 playbook for closing a flat or down round and keeping the team intact afterward.

How common are down rounds in H1 2026

Down rounds peaked in mid-2025 and have fallen for two consecutive quarters. The aggregate share is now meaningfully lower than founder anxiety suggests.

The cleanest snapshot of where the market sits in H1 2026:

The contradiction between Carta and PitchBook is not noise. Carta measures priced rounds on its platform, which skews toward US-incorporated startups using its cap-table software. PitchBook captures a broader universe of venture activity including later-stage and structured deals. Reading both: priced down rounds are getting less common quarter over quarter, even as the share of companies sitting below water has expanded to about half the ecosystem.

The gap matters because it explains the lived experience. Most founders are not closing down rounds. Most founders are running extensions, bridges, and tender offers to avoid pricing at all. The 50% figure includes everyone who would price down if they had to. The 14% figure is everyone who actually did.

Carta Q2 2025 also documents the median 696-day wait between rounds, the longest gap in recent history, and only $46.9B funded across H1 2025 versus $97B over the same period in 2024. The cash isn't gone. The pricing event is just being deferred.

Down round vs flat round 2026 vs up round

A down round prices the company below its last round. A flat round prices it at the same valuation. Anything else is an up round. The distinction matters because legal protections trigger on the price delta, not on founder mood.

The breakdown that controls your cap-table math:

• Down round 2026: new pre-money below the prior post-money. Anti-dilution provisions for existing preferred holders fire on this event, and any pay-to-play clause activates against non-participating preferred.
• Flat round 2026: new pre-money equal to prior post-money. Anti-dilution typically does not trigger, but founders still face the optical signal of stalled valuation growth.
• Up round: new pre-money above prior post-money. Existing preferred is diluted but at a higher price; no protection triggers.

Wilson Sonsini's down rounds FAQ notes that flat and down rounds became markedly less common during the 2021 boom and consistently more common from 2022 onward. The 2021 baseline (sub-5% per AngelList) was the anomaly, not the current rate.

A useful frame: in H1 2026, an up round at the seed stage is still entirely available if your metrics support it. Carta Q3 2025 shows median pre-money valuations of $16M at primary seed (up 14% year-on-year) and $49.3M at Series A. The market is bifurcated, not closed. Strong metrics still price up. Soft metrics price down or don't price at all.

The AI/ML skew distorting 2026 down round data

AI/ML startups account for 29.3% of all 2025 down rounds and at the same time post a median Series B step-up of 2.1x versus 1.4x for everyone else. Both statements come from the same PitchBook-NVCA Q2 2025 data set. They are not contradictory; they are the entire story.

What the data is telling you: a small number of AI/ML companies are pricing dramatically up at Series B and pulling the category averages with them. The rest, especially companies that raised at 2021 to 2022 AI hype-cycle valuations and have not since shown commercial pull, are pricing down. The headline "AI is up" hides a sharp internal split.

PitchBook-NVCA Q1 2026 confirms the concentration: just 0.05% of deals captured a disproportionate share of Q1 2026 deal value. The mega-rounds at the top are pulling the median up while the median founder underneath is staring at flat or down.

If you are operating in or adjacent to AI/ML, this matters for two reasons. First, you cannot use category-average step-ups as a benchmark when pitching your round; the average is meaningless. Second, incoming lead investors know the bimodal distribution and will price your round against the bottom half unless you can show you are in the top tail. Pre-empt the question with comparable metrics from companies above your stage rather than category-level averages.

Down round terms that actually hurt

Price is the smallest part of what hurts in a down round. The non-price terms are where the dilution actually compounds, and where founders who only negotiate valuation lose the most.

The four terms to read line by line in any down round term sheet, per Cooley GO's down round financings primer:

• Liquidation preference greater than 1x: a 2x or 3x preference means new investors are paid out 2x or 3x their investment before common holders see a cent. On a modest exit, this can wipe out founder and employee proceeds entirely.
• Participating preferred stock: after the preference is paid, participating preferred also shares pro rata in the remaining proceeds with common. The combination of 2x preference plus participation is the most punishing structure new investors push.
• Accruing dividends: typically 6-8% per year, compounded, paid out on liquidation or conversion. Over a four-year hold this adds 25 to 35% to the effective preference.
• Warrant coverage: a free additional equity grant to new investors, usually expressed as a percentage of their investment. 25 to 50% warrant coverage is common in distressed rounds.

These four terms travel together. A "flat round at the same valuation" with 2x participating preferred and 10% accruing dividends is a worse outcome than a clean down round at 30% lower valuation. Founders who only fight on price lose the war on terms.

For a sector-by-sector breakdown of which terms appear most often in 2026 sheets, read term sheet red flags in 2026.

Pay-to-play, structured deals, and ratchets

Pay-to-play forces existing preferred shareholders to participate pro rata in the new round or lose their preferred rights. Non-participants get converted to common, surrendering their liquidation preference, their anti-dilution protection, and any board or information rights tied to preferred status.

The mechanic exists for a single reason: the new lead investor wants the old cap table cleaned up. If half your prior preferred holders are inactive funds that will not write into the new round, pay-to-play converts them to common in one move. The founder benefit is real (a cleaner cap table makes the next priced round easier), but the cost is that you have just publicly burned your relationship with every legacy investor who could not or would not participate.

Anti-dilution protection comes in two main flavors, per Cooley GO:

• Full ratchet: existing preferred is repriced to the new lower price on all of its shares. Maximally punishing for common holders. Rare outside true distressed deals.
• Broad-based weighted-average: a formula that adjusts conversion price using total outstanding capitalization. Less drastic. The standard.

What to negotiate: ask existing preferred holders for an anti-dilution waiver in exchange for participation in the new round on favorable terms. Many will take the trade because the math on full anti-dilution plus a falling cap table is worse for them than a modest waiver with new money attached.

Half the US VC-backed tech ecosystem sits below its last private valuation in H1 2026. A down round stopped being a death signal in 2025, and stopped being remarkable in 2026.

Structured deals are the umbrella term for any of the above terms layered together. PitchBook tracks them separately because pure pricing analysis understates how much value is being transferred to new investors via non-price mechanics. If your incoming term sheet has more than two of (preference > 1x, participating, accruing dividends, full ratchet, warrant coverage, pay-to-play), it is structured, not standard, and the headline valuation tells you almost nothing about real outcomes.

Startup recapitalization: when a recap beats a clean down round

A startup recapitalization restructures the existing cap table rather than just adding new money on top of it. Common moves: collapse preferred classes into one, convert old preferred to common, refresh the option pool, and price the new round at a lower valuation.

Why a recap beats a clean down round in certain situations:

• Inactive prior investors: if your Series A lead's fund is winding down and cannot participate, a recap converts their preferred to common in one motion, freeing the cap table for a new lead.
• Layered preferences: by Series C, the stack of liquidation preferences from prior rounds can total 2 to 4x of invested capital. A recap collapses the stack into a single new preferred class.
• Underwater option pool: a recap is the natural moment to refresh the pool and reprice old grants, which a clean down round forces you to do separately.

Morgan Lewis's 2024 playbook frames the call this way: extend runway before pricing, communicate transparently with the team early, evaluate recapitalization versus new-money alternatives, review anti-dilution and preferred terms line-by-line, and plan option-grant refresh or option-exchange programs to retain talent through underwater option vintages.

The downside of a recap is that existing investors take a real haircut, and the negotiation is harder than a simple priced round. Expect 8 to 12 weeks of lawyer time, not 4 to 6. If you can avoid a recap by raising an extension or bridge, do so. If your cap table is genuinely broken, recap is the right tool.

Extension rounds and tender offers as down round alternatives

Before you take a priced down round in H1 2026, exhaust the three structural alternatives: extension, bridge, and secondary tender. SVB's H1 2026 report explicitly notes "more companies turning to extension rounds to meet runway shortfalls" as the structural pattern of this market.

The alternative menu, in order of founder preference:

• Extension round: new money raised on the same terms as the prior priced round, usually from existing investors plus one or two new participants. No revaluation, no anti-dilution trigger. The price you pay is dilution at the old (often optimistic) cap, which is rough math but better than triggering a full down round. See bridge rounds and extensions for the full structural breakdown.
• Convertible bridge: new money raised on a SAFE or convertible note that converts at the next priced round, usually with a discount and a cap. Defers the pricing question by 6 to 18 months. Useful when you have a clear milestone (revenue threshold, product launch) that will support a clean up-round later.
• Secondary tender offer: not new primary capital but liquidity for existing common and option holders. Carta documented 396 tender offers in 2025, with ~20% at Series E or later. The structure lets you retain employees through underwater options without raising new equity at all.

The right choice depends on what's broken. If you need cash to fund 12 more months of operation and your existing investors will support, extension. If you need to keep your top engineers from leaving but cash is fine, tender. If both, layer extension and tender into the same closing. For founders thinking about how long they actually need to extend, read post-seed runway management.

The H1 2026 founder playbook for walking through a down round

Walk in with 9 months of runway minimum, a transparent team narrative ready, and a refreshed option pool plan already drafted. Founders who start the process at 4 months of runway lose every negotiation that follows because they have no walk-away option.

The H1 2026 sequence, in order:

1. Extend runway before pricing: cut burn, defer non-critical hires, and stretch cash to at least 9 months before you start preliminary investor conversations. Term sheets favor sellers, not buyers, only when the seller can credibly say no.
2. Talk to the team early, in person: communicate the situation to senior team members before any paper goes out. They will hear it from the recruiting market within 30 days of the round announcement otherwise. Carta's framing calls this out as the single most under-prioritized founder action in down-round scenarios.
3. Review the prior cap table line by line: surface every legacy preferred right, every anti-dilution clause, every pro-rata commitment. The new lead will do this themselves; you want to know the answer first.
4. Evaluate recap versus new-money first: decide whether the structural problem is cap-table cleanup (recap) or pure cash (new money plus extension). The path determines who you call.
5. Refresh the option pool pre-money: every priced down round will require an option-pool top-up; do it pre-money before negotiation, not post-money where it costs you double.
6. Plan the option exchange: for engineers with grants from the prior round whose options are now underwater, design an exchange program (cancel old underwater grants, issue new grants at the new fair-market value) in parallel with the round. Without it, the top 10% of your engineering team will leave within 6 months of closing.

For the day-by-day mechanics of walking through a Series A down round process, the linked guide breaks the workstreams into investor diligence, employee communication, and legal review in parallel.

One operational note. The investor outreach in a down round is meaningfully harder than in an up round; partners screen against the headline first and the math second. Tools like Causo can help target investors who actively price into soft markets rather than burning warm intros on funds that will pass on the optics.

What this means for your raise

The single most useful frame from H1 2026: the market is bifurcated, not down. Strong metrics still price up at seed and Series A, with Carta Q3 2025 median pre-money at $16M for seed and $49.3M for Series A. Soft metrics now have a real menu of options before a true down round: extension, bridge, tender, recap, structured deal, or a clean priced down round at lower valuation. Picking the right tool from that menu is the H1 2026 founder skill that did not exist in the 2021 playbook.

Note also Y Combinator's published position: a large share of each batch already raised at higher valuations than the YC investment implies, and YC explicitly does not treat its standard investment as a down round. The valuation-only frame is a 2021 artifact. The 2026 frame is outcome improvement. Apply the same logic to your own round.

FAQ

What is a down round and how is it different from a flat round? A down round prices the company below its last priced round. A flat round prices it at the same valuation. Both materially differ from an up round, which prices above. The distinction matters for anti-dilution math: full ratchet and weighted-average protections only trigger on a true down round, not a flat one.

How common are down rounds in 2026, and are they still considered a stigma? Carta logged under 14% of all new fundings as down rounds in Q4 2025, the lowest in nearly three years, while PitchBook recorded 15.9% YTD. But SVB's H1 2026 report shows roughly half of US VC-backed tech companies sit below their last private mark. The stigma is gone; it is now a baseline market state.

What does a startup recap or recapitalization look like in practice? A recap restructures the existing cap table rather than just adding new money on top. Common moves: collapsing preferred classes into one, converting old preferred to common, refreshing the option pool, and pricing the new round at a lower valuation. Existing investors usually take a haircut so new money will commit and so the option pool retains enough room to retain talent.

What is pay-to-play in a venture down round and how does it work? Pay-to-play forces existing preferred shareholders to participate pro rata in the new round or lose their preferred-stock rights. Non-participants get converted to common stock, surrendering liquidation preference and anti-dilution protection. It is the lever incoming investors use to clean up an inactive cap table without negotiating with every prior backer one by one.

How do you survive a down round without losing the team or top performers? Three moves: refresh the option pool before pricing so new grants land at the new fair-market value, communicate the round narrative to the team in person before paper goes out, and design an option exchange or top-up program for engineers whose old options went underwater. Skipping any of these costs you at least one key hire in the 90 days after closing.

Related on the hub

• The H1 2026 AI startup funding report — Related fundraising basics guide.
• The H1 2026 Pre-Seed Funding Report — Related fundraising basics guide.
• The H1 2026 State of Seed Fundraising Report — Related fundraising basics guide.