A government is on the cap table now
CuspAI raised $450M with the UK Sovereign AI Venture Fund alongside Kleiner Perkins and Bezos Expeditions. State money is a normal co-investor. Price what it asks for.
On 20 July 2026, CuspAI announced a $450 million Series B at a $2.6 billion valuation. Kleiner Perkins and NEA led it.
Look at who else was in the round. The UK Sovereign AI Venture Fund. Bezos Expeditions. AMD Ventures. Invest-NL, from the Netherlands.
Two governments and a chip maker, in the same round as two of the best known venture firms in the world.
This is not a grant story
Government money in startups used to mean a small grant, a long form and a report nobody read.
What is happening now is different in scale and in structure. State funds are taking equity, in competitive rounds, alongside tier one investors.
That should change how you think about your own options. Not because state money is better, but because it is available and most founders never look.
I have watched founders spend four months chasing a seed round. A national programme in their own country was open the whole time.
What state money actually buys you
Three things, none of which are the cheque size.
Time, because these processes run on their own calendar and are not affected by whether venture is having a nervous quarter.
Credibility with a specific audience, which matters if your customers are hospitals, utilities, banks or ministries.
And validation of the kind that survives a downturn. A sovereign fund does not mark you down because the market got jumpy.
Two sources, two shapes
Venture money
Fast, expensive, aligned on exit
- Weeks to close when it goes well
- Priced, dilutive, wants a big outcome
- Board seat and a strong view on speed
- Follows the market mood
State backed money
Slow, cheaper, aligned on presence
- Months, on a published timetable
- Sometimes non-dilutive, often matched
- Reporting obligations instead of a board seat
- Cares where the jobs and the research sit
The four things to check before you take it
State capital is not free money. It is money with a different set of conditions attached.
Read the conditions
- Where you must be Location commitments on headquarters, staff or research. Fine now, awkward if you later move to a customer market.
- What you must report Reporting is usually heavier than a venture investor asks for. Cost it in hours, not in principle.
- Who can buy you Some programmes bring approval rights on a foreign acquisition. Check the change of control terms early.
- What the clock does Application windows and disbursement schedules are fixed. Plan runway around their calendar, not yours.
The prompt
Most founders do not know what their own country offers. This is a research pass, not a decision.
You are mapping non-dilutive and state backed funding for a startup.
My company: [ONE LINE]
Sector: [X]. Country and region: [Y]. Stage: [Z].
Headcount: [N]. Doing R&D in: [COUNTRY].
Currently raising: [AMOUNT], for [PURPOSE].
Find and list the government, sovereign fund, development bank and
public research programmes that a company like mine could apply to.
For each one, give me:
NAME:
TYPE: (grant, matched funding, equity, loan, tax credit)
TYPICAL_SIZE:
DILUTIVE: (yes or no)
ELIGIBILITY: (the specific thing that could disqualify me)
STRINGS: (location, hiring, reporting, acquisition approval)
TIMELINE: (application window and time to money)
SOURCE: (the page you took this from)
Rules: only list programmes you can point to a source for. Write
UNVERIFIED next to anything you are not certain is still open. Do not
invent programme names. At the end, rank the top three by time to money,
and say which single one I should start this week.Verify every line it returns against the actual programme page. Funding schemes open and close constantly, and a confident wrong answer here wastes a month.
The honest trade
State money is slower and its paperwork is real. If you need cash in six weeks, this is not your route.
If you are twelve months out from a raise, it is often the cheapest capital you will ever take.
And it does something venture money cannot. It funds the eighteen months of work that has to happen before a venture investor finds you interesting.
The funding map prompt is on the resources page. Free, no email required.