You have two accelerator offers in your inbox and a spreadsheet full of numbers pulled from blog posts. One post says Techstars takes 6 to 9 percent for $120k. Another says Neo invests $600k on a $10M floor. A third lists IndieBio’s biotech deal. Here is the problem: all three of those numbers are wrong in 2026. Techstars moved to $220k in Fall 2025. Neo replaced its accelerator with a $750k residency in February 2026. IndieBio no longer exists as a brand.
Accelerator terms change constantly and almost nobody dates their data. The comparison content you find ranks well, reads confidently, and describes deals that stopped existing years ago. If you are choosing between offers, or deciding whether to apply at all, you are likely negotiating against stale numbers.
So we verified the terms ourselves: 33 programs, checked against each program’s own published pages, with a source and an as-of date on every row. Where a program does not publish a number, the table says so, because that silence is information too.
What the top accelerators actually take in 2026
The condensed table below covers the headline terms. The full database adds sources, as-of dates, program length and format, follow-on rights, and regional variations for every row.
| Program | Check | Equity | Instrument | Fees | Worth knowing |
|---|---|---|---|---|---|
| Y Combinator | $500k ($125k + $375k) | 7% on the $125k | Post-money SAFE + uncapped MFN SAFE | None | Pro-rata in later rounds; US, Canada, Cayman, Singapore entities only |
| Techstars | $220k ($20k + $200k) | Min 5% common, plus SAFE conversion | CEA + uncapped MFN SAFE | None | New deal since Fall 2025; APAC programs still $120k total |
| a16z speedrun | Up to $1M ($500k now + $500k next round) | 10% on the upfront $500k | SAFE | None | No board seat; requires employee option pool |
| South Park Commons | $1M ($400k now + $600k guaranteed follow-on) | 7% on the $400k | SAFE | None | $600k participation in your next external round is guaranteed |
| Neo Residency | $750k + $450k compute credits | No fixed %, uncapped | Uncapped SAFE | None published | Replaced Neo Accelerator in Feb 2026; converts at next round price |
| Entrepreneur First | Up to $250k ($125k + optional $125k) | 8% on the first $125k | Post-money SAFE + uncapped MFN SAFE | None | Second tranche requires Delaware C-corp + SF relocation |
| 500 Global (Flagship) | $150k | 6% | Not published | Not published; reportedly $37.5k deducted from the check | Net cash may be ~$112.5k; verify before signing |
| Antler (US) | “$500K-$1M” commitment, structure not published | Not published | Not published | None to join | Last published deal (2023: $250k at $2.75M post) is likely obsolete |
| SOSV (HAX / SOSV SF / SOSV NY) | Up to ~$550k average pre-seed | 10%+ ownership target | Post-money SAFE, no cap, target ownership stated | None stated | HAX first check is $250k: $150k cash + $100k in-kind |
| PearX | $250k to $2M by stage | Not published | Not published | None | Requires SF Bay Area relocation; 9 months free office |
| gener8tor | $100k | 7.5% | SAFE | None stated | Terms non-negotiable; gBETA is their free no-equity track |
| ERA (NYC) | $150k | 6% | Post-money SAFE | None published | Two 4-month NYC cohorts per year |
| Berkeley SkyDeck | $210k | 7.5% | Post-money SAFE | $7.5k program fee | Half of fund profits go back to UC Berkeley |
| AngelPad | $120k + credits | ~7% (5% common + ~2%) | Common stock + investment | None stated | No documented cohort since 2019; verify it is actually running |
| Alchemist | ~$30k net average | ~5% typical, flexible | Common equity grant | Tuition netted against investment | Enterprise focus; 6 months, remote-friendly |
| AI Grant | $250k + $600k credits | No fixed % | Uncapped no-discount MFN SAFE | None | Converts at your next round’s terms |
| HF0 | Not published | Not published | Not published | Not published | Secondary sources conflict ($1M for 5% vs $500k + 3%); ask directly |
| Betaworks Camp | Up to $500k | Not published | Not published | None; travel on you | Thematic cohorts; NYC residency required at start and end |
| Sequoia Arc | Not published | Not published | Not published | Not published | Terms are explicitly company-specific by design |
| Founder Institute | No cash invested | 2.5% warrant | Warrant | You pay: $799 to $1,199, more for teams | Equity is shared with local leaders and mentors |
| Startupbootcamp | EUR 25k expense coverage | 8% per 2024 pages; not restated since | Direct shareholding | None stated | Ask for the current percentage in writing |
| Station F Founders Program | No standard investment | 0% automatic | Conditional EUR 50k co-investment right | Desk fee from EUR 259/month | The old 1% stake was dropped in the Feb 2024 terms |
| LAUNCH | $125k | 7% | Not published on current pages | None published | 14 weeks, heavy fundraising focus |
| 43North | $1M to 5 winners/year | 5% | Warrant, non-negotiable | None | Requires Buffalo relocation; 2026 is the final competition |
| MassChallenge | Zero-equity cash prizes | 0% | None | None | Non-dilutive; eligibility caps on prior funding and revenue |
| Plug and Play | No standard check | 0% for the program | Separate optional VC investment | None | Program acceptance and investment are separate decisions |
| NVIDIA Inception | No investment | 0% | None | None | Membership program: credits, GPU pricing, investor exposure |
| StartX | No funding stated | 0% | None | None | Stanford founder community |
| Dreamit | $500k-$1.5M pre-Series A | “Small amount”, % not published | Not published | Not published | No longer a classic cohort accelerator; securetech focus |
| Afore FIR | “At least $100k”, flexible | Not published | Not published | None | 10 weeks in SF, 5 to 8 teams |
| Boomtown | Not published | Not published | Not published | Not published | Appears pivoted to corporate innovation; no open cohort found |
| APX (Berlin) | No program since 2021 | n/a | n/a | n/a | Now an earliest-stage VC, still investing |
| Village Global | No program as of 2026 | n/a | n/a | n/a | Now a pre-seed/seed VC writing $500k-$3M lead checks |
Condensed view. Every figure comes from the program’s own published pages except where marked; figures the program does not disclose are listed as not published rather than filled with guesses.
Get the full database
The complete version adds a source link and an as-of date for every row, plus program length and format, follow-on and pro-rata rights, regional term variations, and the notes on which widely-cited numbers are now obsolete. It is a live spreadsheet, so it gets corrections as programs update their deals.
What a point of equity costs, program by program
One way to compare transparent programs: divide the cash wired at acceptance by the fixed equity percentage. This ignores uncapped SAFE tranches (which convert into additional dilution later at your next round’s price), so treat it as a comparison of headline deals, not final ownership math.
| Program | Cash at acceptance | Fixed equity | Cash per point |
|---|---|---|---|
| gener8tor | $100k | 7.5% | $13.3k |
| Entrepreneur First | $125k (first tranche) | 8% | $15.6k |
| ERA | $150k | 6% | $25.0k |
| 500 Global | $150k (before reported fee deduction) | 6% | $25.0k |
| Berkeley SkyDeck | $210k (before $7.5k fee) | 7.5% | $28.0k |
| Techstars | $220k | 5% minimum | $44.0k |
| a16z speedrun | $500k upfront | 10% | $50.0k |
| South Park Commons | $400k upfront | 7% | $57.1k |
| Y Combinator | $500k | 7% | $71.4k |
Read this the way an investor would. The programs paying the most per point (YC, SPC, a16z) are pricing in their acceptance rate and network effect. The programs paying the least are pricing in geography or stage. Neither end is wrong, but you should know which trade you are making. A $100k check for 7.5% prices your company at $1.33M post. A $500k check for 7% prices it at $7.1M. Same word, accelerator, five-times difference in implied valuation.
The fee traps
Four patterns in the data deserve a highlighted warning:
You pay them. Founder Institute charges founders $799 to $1,199 up front (teams pay 1.5x to 2x) and still takes a 2.5% warrant, shared among local leaders, mentors, and headquarters.
The check is smaller than the headline. 500 Global’s flagship page says $150k for 6%. Multiple secondary sources report a $37.5k program fee deducted from that check, and 500’s own regional pages publish a comparable fee structure. If you get an offer, ask what lands in your bank account.
The investment nets against tuition. Alchemist’s structure results in average net proceeds around $30k. That is not hidden, but it surprises founders who read “accelerator investment” and expect six figures.
The program fee rides alongside the SAFE. Berkeley SkyDeck invests $210k for 7.5% and charges a $7.5k program fee per startup.
What programs do not publish is also data
The transparency spread is wide. YC, Techstars, EF, a16z speedrun, and South Park Commons publish complete deals: amount, instrument, percentage, conditions. PearX and Betaworks publish the check but not the equity. HF0 publishes essentially nothing, and the secondary sources contradict each other. Sequoia Arc states outright that every deal is company-specific.
Unpublished terms are not automatically bad terms; they usually mean the program negotiates per company, which can work in your favor if you have leverage. But it moves work onto you. If the terms are not on the website, get them in writing before you invest interview time, and compare them against the transparent programs above as your baseline.
This is the same discipline we push founders toward in Elite Founders: treat every partnership offer, accelerator or otherwise, as a priced transaction with a baseline you can defend, not as a prize you won.
The zombie list
A surprising share of the accelerators still ranking in comparison articles are not operating as accelerators anymore. As of July 2026: AngelPad’s site describes an active program but no cohort has been documented since 2019. Boomtown appears to have pivoted to corporate innovation services. APX dropped its accelerator in 2021 and operates as an early-stage VC. Dreamit ended fixed cohorts and focuses on securetech investing. Village Global’s Network Catalyst is gone, replaced by conventional pre-seed investing. And 43North, which pays $1M for 5% if you relocate to Buffalo, has announced 2026 as its final competition.
Before you spend a weekend on any application, confirm the program ran a cohort in the last twelve months.
Three mistakes founders make with accelerator terms
1. Comparing checks instead of implied valuations. $500k for 10% and $125k for 7% are not “bigger and smaller checks”, they are different prices for your company. Do the division before you decide anything.
2. Ignoring the uncapped SAFE tranche. The modern standard deal (YC, Techstars, EF) is a fixed slice plus an uncapped MFN SAFE. That second tranche converts at your next round’s price, which means the true dilution depends on a valuation that does not exist yet. Model it at your realistic next-round number, not your hopeful one.
3. Trusting undated data. Techstars changed its deal in 2025. Neo changed in February 2026. Station F dropped its equity stake in 2024. Founder Institute cut its warrant from 4% to 2.5%. If the article you are reading does not say when its numbers were checked, assume they are wrong.
Where this fits in your funding path
An accelerator is one instrument among several, and the table makes the real question visible: what are you buying for that equity? If the answer is capital alone, some of these deals are expensive capital. If the answer is distribution, network, or a forcing function for a specific stage transition, the same deal can be cheap. That evaluation depends on where your company actually is, which is a diagnosis worth doing before any application, and it is exactly the kind of question we work through at our Founders Meetings, live sessions where founders pressure-test decisions like this one.
Download the full database above, put your offers next to the baseline, and negotiate from data instead of blog posts.
