



Anti-dilution protection lowers the conversion price of an investor's preferred stock if the company later sells shares at a lower price, so each preferred share converts into more common stock. Full ratchet resets the price all the way down to the new, lower price, however small the down round. Weighted average blends the old and new prices according to how many cheap shares are sold. Broad-based weighted average is the market standard: it appeared in 98% of the deals in Wilson Sonsini's first-half 2026 data, against 1% for full ratchet.
Anti-dilution sits quietly in almost every priced term sheet and does nothing in an up round. In a down round it can decide how much of the company founders and employees still own. This guide explains how each type works, walks through one company's numbers under every version, and covers what founders should accept and what they should push back on.
Venture investors usually buy convertible preferred stock. Each preferred share converts into common stock at a ratio of the original purchase price divided by the conversion price. At the start, the conversion price equals the purchase price, so one preferred share converts into one common share.
Anti-dilution protection changes the conversion price, not the number of preferred shares:
This price-based protection is different from the standard adjustments for stock splits and dividends, which apply to all shareholders equally. It is also different from pro-rata rights, which only let an investor keep its percentage by putting in new money.
Anti-dilution applies to preferred stock in priced rounds. SAFEs and convertible notes don't carry it until they convert (see our guide to how SAFE notes work).
Under a full ratchet, the investor's conversion price drops to the price of the new shares, regardless of how many shares are sold at that price.
Full ratchet has almost disappeared from mainstream deals. It punishes founders and employees for market-wide repricing, and it makes the next round harder because new investors inherit a distorted cap table. Where it still appears, it is usually in rescue financings or deals where one investor has unusual leverage. Some term sheets use a time-limited ratchet, which switches to weighted average after a set date. The model documents of the National Venture Capital Association (NVCA) include this option.
Weighted average anti-dilution blends the old price with the new one, weighted by how big the cheap issuance is compared with the company's existing shares. The NVCA model Certificate of Incorporation uses this formula:
CP2 = CP1 × (A + B) ÷ (A + C)
Because B is smaller than C in a down round, the fraction is below 1 and the conversion price falls. The bigger A is compared with the new issuance, the smaller the adjustment.
The only difference is the definition of A:
Narrow-based formulas are now rare. Wilson Sonsini recorded narrow-based weighted average in 0% of its deals in 2024, 2025 and the first half of 2026.
Here is a hypothetical company before its next round, on a fully diluted basis:
The company then raises a $4,000,000 Series B at $1.00 per share, half the Series A price, issuing 4,000,000 new shares (C). At the old $2.00 price, that money would have bought 2,000,000 shares (B).
(a) No anti-dilution protection
(b) Broad-based weighted average
(c) Narrow-based weighted average (common only)
(d) Full ratchet
What it costs founders compared with no protection:
Employees lose too: the options' share falls from 7.14% with no protection to 6.90% under broad-based and 5.88% under full ratchet.
The small-bridge test. Suppose the company instead sells only $500,000 of shares at $1.00 (C = 500,000; B = 250,000):
That is why full ratchet is considered punitive: a small, cheap bridge can cost founders more than a large down round would under weighted average.
This example is simplified. It assumes one preferred series, no option pool top-up, and a new investor that does not price its round around the adjustment. In practice, new investors often insist on their target percentage after the adjustment, which pushes even more of the cost onto common stock.
Down rounds are also less frequent. In Cooley's data, down rounds fell from 32% of deals in Q1 2024, the highest since the report began in 2014, to 7.7% in Q3 2026 (Cooley GO data). Investors who want protection in a weak round now tend to use pay-to-play instead: Cooley recorded pay-to-play provisions in 11.1% of Q3 2026 deals, the highest level since 2014.
Every anti-dilution clause should come with a list of "exempted securities" that never trigger it. The NVCA model list includes:
Without these carve-outs, a routine option grant or a venture debt warrant priced below the last round could technically trigger an adjustment. See our guide to venture debt for how lender warrants work.
Waivers. The adjustment can usually be waived by a majority of the preferred holders. Founders should aim for a class-wide majority waiver, so that one holdout investor cannot block a rescue round.
Accept as market standard:
Push back on:
For investors, broad-based weighted average combined with pay-to-play protects capital without breaking the cap table. A crushed common class makes the company harder to run and to refinance, and later investors will often demand that a full ratchet be waived before they invest. For more on reading the full document, see navigating the term sheet.
In the UK, the BVCA model investment documents, updated in 2023, make broad-based weighted average the default and keep full ratchet and narrow-based formulas only as alternatives. UK anti-dilution is often delivered by issuing additional shares to the protected investor at nominal value, rather than by adjusting a conversion price. Founders raising money under SEIS or EIS should take specialist advice before granting investors any preferential rights, because those schemes restrict them.
Full ratchet resets the earlier investor's conversion price to the new, lower price, however small the down round. Weighted average blends the old and new prices according to how many cheap shares are issued compared with the existing share base, so a small down round produces only a small adjustment.
Broad-based weighted average. Wilson Sonsini recorded it in 98% of its deals in the first half of 2026 and in 100% in 2025.
Not directly in US-style charters. It lowers the conversion price, so the same preferred shares convert into more common shares. In the UK it is often done by issuing additional shares to the investor.
Issuing shares, options or convertible securities at a price below the investor's current conversion price, unless the issuance is on the exempted list (such as employee option grants or an IPO) or the required preferred holders waive the adjustment.
Only the definition of A, the existing share base. Broad-based counts common stock, outstanding options and all preferred as converted. Narrow-based counts fewer shares, which makes the adjustment larger. In our example, the new conversion price was $1.7143 under broad-based and $1.60 under narrow-based.
Rarely in a priced round. Founders can limit them through carve-outs, a pay-to-play clause, a time limit on any ratchet, or a majority waiver when a down round happens.
Anti-dilution protection only matters in a down round, but when it applies it can move several percentage points of ownership from founders and employees to earlier investors. Broad-based weighted average is the standard and a reasonable trade. Full ratchet can cost founders many times more, especially in a small bridge, and should be resisted or time-limited outside a genuine rescue. The details worth negotiating are the definition of A, the list of carve-outs, the waiver threshold and whether pay-to-play applies. Look up related terms in the GCN glossary.
Global Capital Network connects founders with angels, venture funds and family offices through our investor events and investor network. Get in touch to learn more.
This article is general information, not legal or investment advice. Speak to a qualified lawyer before signing a term sheet.