What Founders Should Make of 2026 SAFE Caps

For a U.S. technical founder raising pre-seed capital in 2026, the most useful reference point is not one market-wide “valuation.” It is the median post-money SAFE valuation cap for a comparable financing size. The reported medians rise sharply with the amount raised, so a cap should be evaluated alongside the capital required, the ownership it may represent, the milestone being financed, and a plausible next round.

These benchmarks are diagnostics, not prescribed terms. The objective is not to secure the highest headline cap. It is to choose a financing structure that gives the company enough capital to reach a meaningful proof point without creating unintended dilution or next-round friction.

2026 pre-seed SAFE cap benchmarks by raise size

The clearest pattern in the available 2026 data is that reported median caps rise with the amount raised. That makes a single “typical pre-seed cap” misleading.

Financing or SAFE-size cohort	Reported median post-money SAFE cap	Source status	Confidence
Below $250,000	$10 million	Secondary reporting of Carta data	Moderate
$250,000–$499,000	$12 million	Secondary reporting of Carta data	Moderate
$500,000–$999,000	$12.5 million	Secondary reporting of Carta data	Moderate
$1 million–$2.4 million	$18 million	Secondary reporting of Carta data	Moderate
At least $2.5 million	$35 million	Secondary reporting; directly supported for SAFEs larger than $2.5 million	Higher for the cap; moderate for cohort comparability

The complete bracket series comes from Finro’s secondary reporting of Carta’s Q2 2026 data. Every number in the table is a reported median post-money SAFE cap—not a recommended cap and not a priced-round valuation of the company.

There is an important definition issue. Finro describes its brackets by round or financing size, while Carta’s own discussion refers to individual SAFE size. Carta directly states that the median cap for SAFEs larger than $2.5 million reached $35 million in Q2 2026, up 40% year over year, and says caps increased across all SAFE-size groups over the preceding year. The public descriptions do not establish that “at least $2.5 million of financing” and “a SAFE larger than $2.5 million” identify precisely the same cohort (Carta).

Use the table as a diagnostic:

• Compare the proposed terms with the bracket nearest the actual raise.
• Check whether the source classifies an aggregate financing, a round, or one SAFE.
• Ask what company-specific evidence justifies being below or above the median.
• Calculate the ownership represented by the complete financing, not just one instrument.
• Test whether the company could credibly support the next financing implied by the cap.

A median describes the midpoint of an observed cohort. It does not determine what is correct or fair for a particular company.

A SAFE cap is a conversion term, not a priced valuation

A valuation cap is a contractual term that can affect the price at which a SAFE converts into equity when shares are issued later. It is not the same as investors buying shares today at a negotiated company valuation.

A post-money SAFE cap instead supplies a basis for calculating conversion under the SAFE’s terms.

Post-money SAFE caps are the relevant pre-seed benchmark because SAFEs dominate Carta’s observed market. In Q2 2026, 93% of observed pre-seed rounds used SAFEs, those SAFEs represented 95% of capital raised, and 91% of observed SAFEs were post-money. In the first half of 2026, 94% of post-money SAFEs had a cap: 73% were cap-only and 21% combined a cap with a discount, according to Carta’s analysis of 2026 SAFE terms.

A cap-only SAFE has a valuation cap but no separate conversion discount. A cap-plus-discount SAFE contains both. The governing document and the terms of the later financing determine how those provisions operate; the mere presence of both does not reveal the eventual conversion result.

The headline cap therefore cannot be evaluated in isolation. Review the governing documents for:

• the exact pre-money or post-money SAFE form;
• the valuation cap and any discount;
• most-favored-nation rights and possible elections;
• pro rata or participation rights;
• the treatment of other SAFEs and convertible notes;
• capitalization definitions and option-pool treatment; and
• conversion provisions for an equity financing, liquidity event, or dissolution.

Founders should model those provisions together rather than assume that two instruments with the same headline cap will produce the same ownership outcome. Finta’s fundraising guidance likewise recommends recording and modeling SAFEs and notes collectively, including caps, discounts, MFN rights, pro rata rights, and the distinction between pre-money and post-money structures (Finta).

Dilution examples: turn the cap into an ownership estimate

For a post-money SAFE that converts at its cap, a useful first-pass estimate is:

Approximate ownership attributable to the SAFE = investment ÷ post-money cap

This shortcut turns an abstract cap into an ownership estimate.

Investment	Illustrative post-money cap	Simplified calculation	Approximate ownership
$250,000	$10 million	$250,000 ÷ $10 million	2.5%
$500,000	$12.5 million	$500,000 ÷ $12.5 million	4.0%
$1 million	$18 million	$1 million ÷ $18 million	5.6%
$2.5 million	$35 million	$2.5 million ÷ $35 million	7.1%

These are arithmetic illustrations using the reported caps, not predictions of final cap-table ownership. In particular, exactly $250,000 is outside the reported below-$250,000 cohort. The first row therefore uses that adjacent cohort’s $10 million median as a boundary illustration; it does not claim that the source reports a $10 million median for an exact $250,000 financing.

The calculation also works in reverse:

Post-money cap = investment ÷ target ownership

If a founder wants a $750,000 SAFE to represent approximately 10% ownership when converting at the cap, the starting calculation is:

$750,000 ÷ 10% = $7.5 million post-money cap

That is a mathematical starting point, not a market conclusion. The founder would still need to compare the result with the relevant raise-size evidence, current traction, team strength, investor demand, and the economics of the next round.

The shortcut has important limits. Actual ownership can change because of:

• additional or stacked SAFEs;
• convertible notes;
• conversion discounts;
• MFN elections;
• pro rata participation;
• option-pool increases; and
• the precise capitalization and conversion provisions in each document.

A post-money SAFE can make ownership attributable to that instrument easier to estimate, but it does not freeze everyone’s final ownership.

Model the complete cap table before approving each instrument. At minimum, show how every outstanding SAFE and note converts under the same financing scenario rather than calculating one “representative” SAFE and assuming the rest behave identically.

Choose the raise first, then test the cap

Start with the change in the company’s risk profile that the capital must buy. Depending on the company, that may mean validating a painful customer problem, building a credible product, completing a difficult technical milestone, demonstrating repeatable usage, or reaching another proof point that makes the next financing plausible.

Then build the raise from expected cash needs and runway. Do not increase the amount merely to enter a higher cap bracket. A larger raise supplies more capital, but it can also create more conversion exposure, greater aggregate dilution, a more complex capitalization structure, and higher expectations for what the company should prove before raising again.

Once the amount is grounded in a milestone plan, test the cap through four connected inputs:

Input	Question to answer	Evidence to prepare
Milestone and runway	What must become demonstrably true before the next raise?	Product, hiring, customer, technical, and timing plan
Amount required	What does that plan cost with a realistic buffer?	Monthly cash forecast and use of proceeds
Ownership implied	What ownership may all outstanding instruments represent?	Modeled SAFE, note, and option-pool scenarios
Next financing	What price and evidence may the company need to support next?	Plausible round size, valuation range, and milestone case

A higher cap reduces the approximate ownership attributable to a post-money SAFE when it converts at that cap. But a higher cap is not automatically founder-friendly. If the company’s progress does not support the economics expected in the next financing, an aggressive cap can create negotiating friction. The actual conversion result will still depend on the SAFE and financing documents.

Run at least three cap-table scenarios:

1. The planned financing: All anticipated SAFEs close on the expected terms.
2. A constrained next round: The company raises less capital at a lower valuation.
3. A larger next round: The company raises more capital alongside an option-pool increase.

For each scenario, show ownership by founder, employee pool, current investor, SAFE or note holder, and new investor. Include all outstanding instruments and relevant rights.

The ask slide should then state the financing plainly:

• amount sought;
• post-money SAFE cap;
• use of proceeds;
• expected runway; and
• milestones the capital is intended to fund.

That presentation is more decision-useful than treating the cap as a stand-alone claim about what the company is worth.

Why the $100 million headline is not the market norm

Myth: A $100 million pre-seed cap is a normal 2026 market outcome. Reality: It was an upper-tail result among large SAFEs, not a typical cap across pre-seed financings.

For the large-financing cohort, the reported median was $35 million. Finro reports a $50 million 75th percentile and $100 million 90th percentile for rounds of at least $2.5 million across the six-quarter period from Q1 2025 through Q2 2026. Those percentile figures are not a Q2 2026-only distribution (Finro).

Carta’s Q2 2026 discussion says the 75th percentile for SAFEs larger than $2.5 million approached $60 million. The difference between $50 million and nearly $60 million may reflect the different periods being summarized, but that is not the only methodological limitation: one source describes a round or financing-size cohort, while the other describes individual SAFE size. The public reporting does not establish that those populations are identical.

The distance from the $35 million median to a $50 million or nearly $60 million 75th percentile—and then to a $100 million 90th percentile—shows substantial dispersion among large SAFEs. It does not establish a general negotiating range for a company raising far less.

Idea Farm’s third-party summary of Carta’s report also characterizes the $100 million figure as a high-end outcome and attributes elevated caps to enthusiasm around AI. The high-end characterization is consistent with the percentile data, but the supplied evidence does not demonstrate that AI enthusiasm caused those caps. Founders should not turn that narrative into an assumed sector premium.

The market context: similar dollars across fewer instruments

U.S.-based startups on Carta raised $3.19 billion through more than 11,500 pre-seed instruments in Q2 2026, compared with $3.22 billion through 14,825 instruments in Q2 2025. Aggregate dollars were nearly flat while instrument count declined by roughly 22%. Average instrument size reached $276,000, up 27% year over year, according to Carta’s Q2 2026 market report.

This indicates that similar aggregate capital was distributed across fewer instruments. It does not mean more than 11,500 distinct startups raised money, because an instrument is not the same as a unique company or a completed financing round. The figures also do not, by themselves, prove that investors became more selective or explain why instrument count fell.

Late-2025 figures provide context but should remain separate from the 2026 benchmark table. VC Lens’s secondary explanation of Carta’s Q4 2025 data reported median post-money SAFE caps ranging from $8 million for financings below $250,000 to about $30 million for financings of at least $2.5 million. Those are dated, secondary 2025 figures—not direct observations for 2026.

Repeated publication of Carta-derived numbers does not create independent confirmation. When several articles interpret the same underlying dataset, they remain multiple accounts of one source rather than separate market samples.

How to use the data—and where it stops

Treat these figures as benchmarks for U.S.-based startups and instruments observed on Carta, not as a census of the U.S. pre-seed market or a global standard. The cited public reports and summaries do not disclose complete sample sizes, inclusion rules, or cohort methodology for every benchmark. Their varying use of “financing size,” “round size,” and “SAFE size” also limits exact comparison.

AI concentration deserves the same care. AI companies captured 49% of Carta-observed pre-seed dollars in the first half of 2026, approximately matching their 50% share for full-year 2025 (Carta). That describes how capital was distributed by category. It does not reveal whether AI companies raised more frequently, raised larger amounts, received higher caps after controlling for financing size, or experienced some combination of those effects.

Do not transfer seed-stage AI valuation premiums onto pre-seed SAFE caps. “AI” is not a fixed multiplier, and a category label does not replace evidence of technical differentiation, customer demand, team capability, or a durable reason for the company to exist.

Before relying on a benchmark, verify:

• Period: Is it Q2 2026, H1 2026, a multi-quarter series, or older context?
• Geography: Does it cover U.S.-based companies, another market, or an unspecified population?
• Instrument: Is the figure for a SAFE, convertible note, or priced equity round?
• Structure: Is the SAFE post-money or pre-money?
• Cohort: Is the bracket based on one instrument, aggregate financing, or completed round size?
• Source status: Is the figure directly reported by the data owner or repeated by a secondary publisher?
• Statistic: Is it a median, percentile, average, or anecdotal high?
• Cap table: How do all outstanding instruments, rights, and option-pool assumptions interact?

The final decision rule has three steps: size the financing around a concrete milestone; calculate ownership under the complete set of SAFE terms; and compare the result with the correctly dated raise-size benchmark and a realistic next-round scenario. The goal is not the highest defensible headline cap. It is a financing structure the company can grow into.

Lunera focuses on the earliest moments of company building and technical founders working in areas such as developer tools, data infrastructure, applied AI, and foundational software. Founders who fit that focus may submit a concise note, deck, or product link directly without a warm introduction. Any company-specific SAFE decision should be reviewed against the governing documents and full cap table by qualified legal and financial advisers.