Accelerator programs
Y Combinator, Techstars, AngelPad, On Deck — what each looks for, what they take, when they help, when bootstrapping is the better call. Plus the niche programs nobody tells you about.
Accelerators are one of the few startup decisions where the cost is permanent and the benefit is finite. You give up 6-7% forever for three months of curriculum, a Demo Day, a check, and a Slack channel. Sometimes that's the best deal in venture. Sometimes it's a tax on founders who didn't need it. Here's how we think about which programs are worth the equity.
Should you even apply to an accelerator?
The trade is the same across almost every program: equity for a bundle of capital, network, curriculum, and brand signal. The capital is small enough that it's not the point — $125K to $500K is six months of runway for a frugal team, not a rocket. The network and brand do the real work.
Apply when you're pre-traction and need the runway plus warm intros to the next round. Apply when you're attacking a market where the brand on your cap table opens doors a cold email won't (enterprise sales, hardware, regulated industries). Apply when you're a first-time founder who has never raised, never managed an investor, and would benefit from structured pressure to ship every week.
Don't apply when you already have product-market-fit signal — paying customers, real retention, organic growth — because at that point you're selling equity at a discount. Don't apply when you can self-fund six months. Don't apply when you'd happily take the network without the check; for that you want a fellowship like On Deck or South Park Commons.
Accelerators are most valuable to founders who are technically capable but commercially green. If you've shipped product and sold it before, you're paying for things you already have.
Y Combinator (YC)
YC is the default benchmark and it deserves to be. The current standard deal in 2026 is $500K total: $125K for 7% on a post-money SAFE, plus $375K on an MFN (most-favored-nation) SAFE that converts at the terms of your next priced round. Three-month batches run on the W26 / S26 / F26 cycle, ending in a Demo Day attended by most of the active early-stage venture market. The alumni network is the strongest in startups — Bookface, the internal forum, is the closest thing to a real-time customer-discovery and intro engine that exists.
The application is famously short. Founder fit and market thesis matter more than traction; YC has admitted plenty of pre-revenue, even pre-product teams when the founder story was sharp enough. Acceptance rate hovers around 1.5%. The W26 cycle had its application deadline in early September 2025; S26 closes in early March 2026.
YC is right when you want speed and the strongest network in startups, and you're willing to relocate to the Bay Area for the batch. It's wrong when you want hands-on operational support — YC's model is deliberately hands-off. Group partners advise weekly, will help you debug a hiring problem or a pricing model, but they will not roll up their sleeves and run your sales process. The expectation is that you can operate; they will accelerate.
The downside nobody mentions: a YC batch is now ~250+ companies. The brand still works, but per-company partner attention has thinned compared to the 2010-era batches of 60. You'll have to fight for that time.
Techstars
Techstars runs a 13-week program with a typical offer around $120K for ~6%, though deal terms have varied across programs in 2025-2026 and some city programs restructured their offers after the 2024 corporate reorganization. They run 50+ programs globally, many vertical (Sports, Music, AI, Sustainability) or geographic (LA, NYC, Boulder, Tel Aviv, London).
The mentor cohort is tighter than YC's and the model is more hands-on. A Techstars program will assign you 4-6 lead mentors who you'll meet weekly, plus a "mentor madness" period at the start where you'll do 60+ short meetings to find your matches. For founders who learn by being told what they're missing, that's high-leverage.
Techstars is right when you want a vertical-specific network — if you're building B2B AI ops tools, Techstars AI in NYC has shipped roughly 12 funded companies in 2025 and the partner network is genuinely deep on that vertical. It's also right when you're not based in SF and want a structured local program with global mentor reach.
It's wrong when you want maximum optionality on terms. The deal structure varies by program and the standard equity stake is meaningfully more expensive on a dollar-for-dollar basis than YC. Read the specific program's offer carefully before signing — and verify before applying, because terms have moved more than once in the last 24 months.
AngelPad
AngelPad is the boutique answer. ~$150K for 7%, twice a year, alternating between New York and San Francisco. Batches are tiny — 12 to 15 companies — and the partner-to-founder ratio is the best of any major accelerator. AngelPad ranked #1 by exit rate for several years for a reason: when you have 12 companies and 2 partners, you actually get the partners.
Right when you're an early enterprise / SaaS / B2B founder who'd benefit from intensive 1-on-1 feedback on positioning, pricing, and the first ten enterprise sales. Wrong when you want a large peer cohort to lean on — your batchmates won't give you the network you'd get out of a 250-company YC batch. Wrong when you want a high-visibility Demo Day; AngelPad's is intimate and curated rather than broadcast.
On Deck
On Deck is not a classic accelerator. It's a fellowship: cohorts of 100+ pre-product-market-fit founders run for 8 weeks, you pay tuition (~$7K to $15K depending on the program and your scholarship status), and they take no equity. On Deck Founders (ODF) is the flagship; there are sibling programs for Angels, Chiefs of Staff, Writers, and others that come and go.
The product is the network. You get matched into small peer groups, attend speaker sessions, and get warm intros to investors and operators. After the 2023 restructuring the staff is leaner and the cachet is somewhat diminished, but ODF cohorts in 2025 and 2026 have still produced founders who go on to raise serious rounds.
Right when you want a peer group and intros without giving up equity yet, you're still figuring out what to build, and you'll do an accelerator (or skip straight to a seed round) later. Wrong when you have a product and need capital now — On Deck doesn't write a check, you write one to them.
HF0 (Hacker Fellowship Zero)
HF0 is for solo founders specifically. ~$500K for variable equity, typically around 5%, 12 weeks residential in San Francisco. Selective, tiny batches (under 20 companies), and the model is monastic: you live with the other founders, meals are provided, and the only expectation is that you ship.
Right when you're a solo technical founder shipping fast and want the most intense focus environment available. Several HF0 alumni have raised seed rounds inside the fellowship without ever doing a formal Demo Day, because investors come to them. Wrong when you have co-founders (HF0 is built around the solo experience), or when you want a structured curriculum — HF0 is more "we cleared the runway, now go." If you need someone to tell you what to work on this week, this is not your program.
South Park Commons (SPC)
SPC is a community-and-fund hybrid. The Founders-in-Residence program is the entry point: you're admitted to the community, you get desk space and access to the SPC network, and there's no equity for the membership itself. The SPC Fund invests separately on standard SAFE terms if and when you're ready to take a check.
Right when you're between things, exploring ideas, and want a smart peer group while you figure out the next thing. SPC is unusually strong for technical founders coming out of senior roles at FAANG, OpenAI, Anthropic, or research labs — the conversations in the building are at that level. Wrong when you have conviction and need to ship now; SPC is structured for exploration, not execution velocity. If you already know what you're building and just need money and intros, take a SAFE from a fund and skip the residency.
500 Global (formerly 500 Startups)
500 Global runs programs across the world with deep emerging-market presence — LATAM, MENA, Southeast Asia especially. Program structures vary by region; the flagship Silicon Valley program differs in shape and check size from the regional accelerators.
Right when your market is outside the US, or when you want emerging-market investor relationships you won't find in a YC batch. 500 has been the connective tissue for a generation of LATAM and SEA founders the way YC has been for the Bay Area. Wrong when you're SF-focused — YC dominates that lane and you'll get more leverage there.
Antler / Entrepreneur First
These are founder-formation programs. You apply as an individual without a co-founder or a committed idea. The program helps you find a co-founder, validate an idea, and form a company; if your team clears the investment committee at the end of formation, they invest roughly $200K for around 10%.
Right when you don't have a co-founder yet but want to start. Wrong when you have a team and an idea you're committed to — the dilution is much higher than YC or Techstars, and you'd be paying for matchmaking you don't need.
Vertical and niche programs worth knowing
The list below dates fast — programs launch, close, change terms, and pivot focus every year. Always verify current programs and terms before applying.
- Cohere AI accelerator — for AI startups building on or around Cohere's models; privileged access to the foundation-model layer.
- Disney Accelerator — for media and entertainment startups; direct line into Disney business units that's otherwise impossible to get.
- Cleantech Open — long-running climate accelerator with regional chapters; non-equity in some tracks. (Newchip shut down in 2023 — don't confuse it with active programs.)
- Backstage Capital — fund and adjacent programming for underrepresented founders; check sizes and program shape have shifted, verify current status.
- Founder Institute — more pre-incubator than accelerator. Runs longer than 3 months, takes 3.5% for the program. Useful for very early founders who need structure before they're ready for a YC-tier program.
- The Combine — vertical SaaS; smaller cohorts, operator-heavy mentor bench.
- PearX — Pear VC's accelerator for tier-1 technical teams; smaller and more selective than YC.
- NYC HealthAI — healthcare AI accelerator with hospital and payer partnerships you can't easily get cold.
If your company sits cleanly inside one of these verticals, the niche program may give you more leverage than the brand-name generalist. If it doesn't, don't force the fit.
Application timing
Most programs run rolling deadlines or quarterly cycles. YC has hard cutoffs: roughly September for Winter, March for Summer, July for Fall, with offers extended within a few weeks of the deadline. Techstars deadlines vary by program; check the specific city/vertical you're applying to. AngelPad runs twice a year on a posted schedule.
Apply when you have a working MVP (or are within weeks of one), some early user signal (even 50 weekly active users from a beta is meaningful), and a specific six-month plan you'd execute if you got the capital. The plan matters: "we'd hire two engineers and try things" is a worse answer than "we'd run paid acquisition against this audience to validate this conversion rate, while shipping these three features."
Don't apply when you're still ideating without a prototype. The application asks what you've built; "nothing yet" is a hard pass at almost every serious program.
What they look for (the honest version)
Across most programs, the same four things show up:
- Founder market fit. Do you understand the market deeply, ideally because you've worked in it or been a user of the broken thing you're fixing?
- Velocity. How fast did you go from nothing to your current state? A team that built a working MVP in three weeks beats a team that took eight months.
- Willingness to ship. Have you actually built things — products, side projects, prior companies — or have you mostly talked and planned? Shipped artifacts beat decks.
- A specific thesis you can defend. Why this, why now, why you. If you can't answer all three in two sentences each, partners will assume you haven't thought hard enough.
Traction helps, especially at the more competitive programs, but it isn't required at the early stages of most accelerators. A thoughtful pre-product team with the right founder DNA will beat a mediocre team with $5K MRR.
For the application itself, see sub-skill 18 (deliverables — pitch deck for the application) — the deck format that works for an accelerator is materially different from the one that works for a Series A.
The alternatives — bootstrap, indie, grants
Accelerator equity isn't free. 6-7% of your company forever is a real number, and the math gets worse if you stack accelerator dilution with seed and Series A dilution. Sometimes the right answer is to skip the program entirely.
Bootstrap to first revenue. The Indie Hackers, IndiePage, and MicroAcquire (now Acquire.com) communities are full of founders who took zero outside capital and built profitable companies. If your product can charge $20-100/month and you can find 100 customers without paid ads, you don't need an accelerator.
Apply for non-dilutive grants. NSF SBIR is the big one in the US — Phase I is up to $275K, Phase II up to $1.7M, no equity taken. State programs, university tech-transfer offices, and vertical-specific grants (Mozilla Open Source Support for OSS, climate funds for clean tech, NIH for health) can stack to meaningful capital without dilution.
Take a small angel check. A founder you respect or a domain expert who knows your market might write a $25-100K check at a SAFE valuation that costs you 0-2% of your company. That's a tenth of the dilution of an accelerator for a third of the capital — but the network is one person, not 250.
Frame the decision honestly. The capital portion of an accelerator is rarely the main value. Ask whether the network, brand, and curriculum are worth 6-7% of your company forever. For some founders the answer is an obvious yes. For others — especially repeat founders, founders with strong existing networks, and founders with traction — the answer is a quiet no.
For the broader question of whether to raise at all, see idea to funding, and before you apply anywhere, run yourself through pre-build due diligence so the answers you give in the application are grounded.