United States / Blog / The state of US startup funding, 2026 — what 34 verified programmes actually say
United States · noteThe state of US startup funding, 2026 — what 34 verified programmes actually say
What we counted, and how
This piece is built from the 34 programmes in our US funding directory as of 15 August 2026 — federal grants, SBA lending, state programmes, tax incentives, accelerators and corporate credit schemes. Every record went through a two-stage check: a researcher documented each programme from primary sources, then a second, adversarial pass re-fetched every official page with instructions to treat the record as wrong until the page proved otherwise.
That method matters more than the count. Most funding roundups are assembled from other roundups, which is how a programme that closed in 2024 keeps appearing in 2026 listicles. Seven of our 34 records carry a needs-verification flag precisely because an official page would not confirm a detail — we would rather print the doubt than the guess.
The headline: two-thirds of US programmes are actually open
Of the 34 programmes, 22 accept applications right now — either open rounds or rolling intake. That 65% openness rate is the highest of the three markets we track, and it is not close: our Australian directory runs at 7 of 20 open on any given day (35%), and the UK at 15 of 32 (47%).
The difference is structural. US federal money is dominated by rolling and multi-cycle intake — SBIR alone runs across 11 agencies with staggered solicitations — while Australian flagship programmes are overwhelmingly round-based, real but shut between windows. For a founder, the practical reading is simple: in the US, the constraint is rarely "is anything open" and usually "which of the open things fits".
- United States: 22 of 34 programmes open (65%)
- United Kingdom: 15 of 32 open (47%)
- Australia: 7 of 20 open (35%)
Non-dilutive money dominates the count
Thirteen of the 34 programmes are grants, and 22 carry no equity cost at all once you include tax incentives, credit programmes and partner credits. The centre of gravity is the SBIR/STTR system — reauthorised in April 2026 through fiscal 2031 — where a Phase I award now runs to $323,090 and a Phase II to $2,153,927 under the government-wide ceilings, with agencies like NSF publishing their own variants.
The equity-free layer is also where the openness concentrates: 8 of the 13 grants accept applications today, and the SBA lending programmes — up to $5 million under 7(a), $5.5 million under 504 — are rolling by design. If your mental model of startup funding is "venture or nothing", the verified list says otherwise.
What equity actually costs, where it is published
The accelerator layer is unusually transparent about price. Y Combinator publishes its standard deal: $125,000 for 7% on a post-money SAFE plus $375,000 on an uncapped MFN SAFE. Techstars publishes 5% common stock plus a $200,000 uncapped MFN SAFE. 500 Global publishes $150,000 for 6%, subject to terms.
Contrast that with several residency-style programmes whose current cheque sizes are not stated on their own pages — our records flag those explicitly rather than repeating secondary-source figures. When a programme does not publish its terms, the honest listing is "not disclosed", and the honest founder response is to ask before applying.
The seven flags, and why they stay listed
Seven US records carry a needs-verification flag. The reasons are mundane and instructive: a federal agency site that bot-blocks automated checks, a state programme whose award range appears only in third-party trackers, a corporate contest between cycles. We keep flagged programmes listed with the doubt printed on the record, because silently dropping them would misrepresent the landscape and silently trusting them would misrepresent the facts.
The flags are also a maintenance signal: each one is a re-verification task with a date attached. A directory that cannot show you its doubts is showing you its guesses.
What this means for a funding calendar
The data suggests a simple sequencing discipline. The rolling layer — SBIR agency solicitations, SBA lending, corporate credits — can be worked at any time and should not be crowded into a panic month. The round-based layer needs a calendar: three US programmes in our set run on cycles, and missing a window means months of waiting, not weeks.
And underneath all of it sits the boring prerequisite the programmes themselves state: financials. Most applications in the grant layer ask for statements, spend plans or both. Books that close monthly turn an application into an export; books that do not turn it into archaeology. That part, at least, is entirely in a founder’s control.
Reading about it is optional. The books aren’t.
A named accountant, software underneath, licensed partners where the law wants them.