How to raise a US seed round in 2026
Last month a founder told us he'd been fundraising for six months and had 40 first meetings but no term sheet. The deck was fine. The metrics were fine. The problem was he'd started in October and was still running the same process in March. Momentum died, and once momentum dies, the round dies with it.
The US seed market split in half this year. AI/ML companies are raising $4.6M on average while the rest of the market sits at $3.1M. Total US venture funding hit $19B in March across 515 deals, with early-stage rounds taking a third of the capital but most of the deal count, according to AlleyWatch's March report. If you're raising outside AI, the bar is higher than it was two years ago and the timelines are longer.
This is the playbook we run with founders raising US seed rounds: what readiness actually looks like, which funds are writing checks, what the deck and model need to show, and how to compress a six-month slog into 12 weeks.
Know if you're ready before you start
Three things tell us a founder is ready to raise, and two of them have nothing to do with traction.
Traction the market rewards right now. What counts as traction depends on what you're building. B2B SaaS companies need $30K-$100K MRR growing 10%+ month-over-month, or named enterprise design partners with clear intent. AI infrastructure companies can go out pre-revenue if the team is exceptional and the technical moat is real—demo plus design partners plus a credible build plan is enough. Applied AI products need usage data, weekly actives, retention curves, and a path to revenue even if you're not charging yet. Consumer or D2C companies need $200K-$800K monthly revenue with retention that holds. Marketplaces need two-sided liquidity in at least one geography and visibility into take rate.
A round size you can defend in one sentence. Investors ask "why this amount?" in the first 10 minutes. The answer that works is milestone-based: "$3M gives us 18 months of runway, gets us to $X ARR by month 14, and puts us in a Series A conversation by mid-2027." If you can't say that sentence cleanly, you're not ready.
Founder time. A serious US seed raise takes 12-18 weeks of full-time founder attention. Stretched raises lose momentum, and momentum closes rounds more than anything else.
The numbers, without the spin
From AlleyWatch's March data: 317 seed deals raised $1.43B, with a $2M median and $4.5M average. Series A deals numbered 108 and raised $4.8B, with a $23.9M median and $44.4M average. NYC took 20.7% of national capital, with San Diego at 11.3%, Austin at 10.4%, Palo Alto at 10.5%, and San Francisco at 10.1%. The Bay Area is no longer dominant by share. AI/ML seed rounds averaged $4.6M, materially above the cross-sector median.
If you're raising AI infrastructure or applied AI, $4-6M at a $20-30M post-money is achievable. If you're raising non-AI consumer or B2B SaaS, $2-3.5M at $12-18M post is closer to the median. Price your round accordingly.
Which funds are actually deploying
Peony's investor brief tracks who's writing checks. Here's who we see closing deals.
First Round Capital writes $3.5M typical checks and stays sector-agnostic across B2B and B2C. Initialized Capital leads with $1-4M and tilts heavily toward AI and infrastructure. Founder Collective averages $3M checks across sectors and has 24 unicorns from sub-$100M funds. Floodgate writes $150K-$1M+ using their "Pattern Breakers" framework. Felicis is multi-stage but still writes seed checks.
On the specialist side, Boldstart Ventures focuses on inception-stage enterprise with their $250M Fund VII. Uncork Capital writes $1.5M-$3M for B2B SaaS, dev tools, and infrastructure. Pear VC does $1-6M pre-seed and seed in dev tools, AI, fintech, and climate. Lerer Hippeau raised a $200M Fund IX and anchors the NYC ecosystem.
Multi-stage funds reach down selectively. Sequoia does seed through their Scout program but it's rare. Andreessen Horowitz writes seed checks in specific theses—crypto, bio, AI—but not broadly. Lightspeed does selective US seed.
Accelerators are a separate path. Y Combinator's standard deal is $500K across four batches per year with roughly 1% acceptance. Techstars is $120K standard over 13 weeks. 500 Global writes $150K standard checks with a global portfolio. AI Grant, founded by Daniel Gross and Nat Friedman, writes $250K standard checks for AI-only companies.
What the deck needs to show
The structure is standard—see our data-driven pitch deck breakdown—but four sections need US-specific calibration.
Why now. A US why-now slide should reference structural unlocks specific to this moment: AI inference cost collapse, the open-source model wave, regulatory shifts like FTC enforcement or SEC rule changes, or platform shifts like Apple's ATT or browser cookie deprecation. "AI is changing everything" reads as lazy.
Market sizing. US VCs trust bottom-up TAM and dismiss top-down. Number of customers times ACV times penetration assumption. A "$1.2T market" slide ends the meeting.
Competition. US VCs are more sensitive to incumbents-with-AI than they were two years ago. Address this explicitly: how does Microsoft, Google, OpenAI, or Anthropic change your competitive picture? "We're a thin wrapper on GPT" is fatal. "The incumbent has 800 things to ship and we have one" works if you can defend it.
Capital efficiency. The 2024-era "$15M seed at $80M post" mindset is dead outside AI infrastructure. Lead with capital efficiency. What does each $1M raised buy in headcount, runway, and milestone delivery?
The financial model that survives diligence
We wrote a long breakdown of the financial model investors actually check. The US-specific additions: 409A valuation and option strike price modeling matter even at seed. US investors expect a 409A post-close and want to see option pool grant dynamics modeled. Enterprise SaaS founders should model 6-9 month average sales cycles by default. Call out burn multiple explicitly—net new burn divided by net new ARR. Under 1.5x is good; over 2x raises questions.
Outreach mechanics that convert
We covered the warm intro versus cold email data in depth. The short version: cold email response rate to top US seed VCs runs 1-5%. Warm intro response rate from a portfolio founder runs 50-80%. Twitter DMs work for solo GPs and partners who post actively, if the message is specific.
The sequence that works: map 30-50 target funds matched to your stage, sector, and geography. Build the warm intro graph for each fund. Reach out to portfolio founders first, not the VCs directly. Batch your outreach—move 12-18 funds into the funnel within a single two-week window. Follow up on day 4, day 10, and day 21 if you haven't heard back.
Timeline from first meeting to wire
Weeks 1-4 are investor readiness. Lock the deck, model, and memo. Run mock sessions. Weeks 5-8 are angel and operator-level conversations. Lock 1-2 strong angel checks to set the price. Weeks 8-12 are the first-meeting cycle with target VCs. Aim for 12-18 first meetings in this window. Weeks 10-14 are partner meetings and diligence with the lead candidates. Weeks 12-18 are term sheet, negotiation, closing documents, and wire.
Five mistakes we see constantly
Pricing at 2024 valuations. The market reset. If you're not in AI infrastructure, your valuation should reflect the new median.
Underestimating the AI thesis filter. Many US VCs will only fund AI-native or AI-defensible companies. If your pitch doesn't address AI, you're out before the second slide.
Chasing the Bay Area when other geographies fit better. NYC, Austin, and Miami are increasingly viable. Match your geography to your customer base and the funds that know your sector.
Skipping operator angels. The norm is 4-8 operator angels alongside the institutional lead. They close the round and open doors for the next one.
Treating YC as binary. Plenty of seed-fundable companies skip YC and run directly with First Round, Initialized, or Founder Collective. YC is one path, not the only path.
Legal baseline before you sign
We wrote about term sheet red flags separately. The US baseline: Delaware C-Corp is the default. File 83(b) elections within 30 days of any restricted stock grant—missing this costs millions at exit. Standard founder vesting is four years with a one-year cliff. Keep a clean cap table on Carta. Use Stripe Atlas or a startup-specialist firm like Cooley, Gunderson Dettmer, or Wilson Sonsini.
Getting started
Be honest about the three readiness signals. If you're short on two, build instead of pitching. Lock the deck and model before you do anything else. Cut a 30-50 fund target list calibrated to your sector and stage. Build the warm intro graph for the top 20 funds. Block 8-10 hours a week for outreach mechanics, and protect that time like you protect product work.
If you want help compressing this from 18 weeks to 12, Phase 2 of our engagement is built for that. Book a call if you want to talk through your specific raise.
