How Much Are Free Startup Credits in the UK Worth?
A UK-incorporated company can realistically assemble somewhere between $10,000 and $200,000 in credits across cloud, models and tooling, because almost every major program is open globally rather than gated to US entities.
That last part is what UK founders get wrong in both directions. Some assume the big programs are American and never apply at all. Others assume a UK Ltd is a handicap and flip to a Delaware C-corp earlier than they need to, paying real legal and accounting fees for a benefit that barely exists at the credits layer. Where incorporation genuinely matters is fundraising and customer contracts, not whether Google will give you cloud credits.
AI Perks tracks $7.7M in credits across 194 companies, and the overwhelming majority are available to a company registered in England, Scotland, Wales or Northern Ireland.
Which Providers Actually Give UK Startups Credits
The stack has three tiers: cloud platforms worth tens to hundreds of thousands, model providers in the low thousands up to around $25,000, and tooling in the hundreds to low thousands per product.
| Provider | Layer | Typical headline value | Open to a UK Ltd |
|---|---|---|---|
| Google for Startups Cloud Program | Cloud | Up to $200,000 for AI-first teams | Yes |
| Microsoft for Startups Founders Hub | Cloud | Up to $150,000 in Azure, tiered | Yes |
| AWS Activate | Cloud | $1,000 to $100,000 depending on route | Yes |
| Anthropic | Models | Up to $25,000 via partner routes | Yes |
| OpenAI | Models | Low thousands, usually via accelerator or investor routes | Yes |
| NVIDIA Inception | GPU and compute | No fixed figure, discounts plus partner cloud credits | Yes |
| Vercel, Supabase, Linear and similar | Tooling | Hundreds to low thousands each | Yes |
| UKRI AI Research Resource | National compute | Allocated GPU hours rather than a dollar balance | UK only |
Treat every figure in that table as a headline maximum, not an expectation. The top tiers are reached through specific routes rather than the self-serve form, and which route you qualify for depends on your stage, funding and who introduced you. Those conditions change often enough that they are worth checking against the current terms on getaiperks.com rather than a blog post from last year.
What Credits Are Actually Paying For
Credits do not make infrastructure free. They move the moment you start paying, which is a more useful thing than it sounds, because the shape of the bill changes completely between those two moments.
Early on, your cloud bill is mostly fixed cost. Managed databases, load balancers, NAT gateways, log retention and a staging environment run whether or not anyone uses your product. A seed-stage team with fifty users pays for capacity, not consumption. Credits at that stage are subsidising idle infrastructure, which feels wasteful but is exactly what stops a pre-revenue team from deleting their staging environment to save forty pounds a month.
Model bills behave the opposite way. Inference cost is almost perfectly linear in usage, with no economies of scale until you are large enough to negotiate committed spend. Every new user costs roughly what the last one did. So model credits are effectively a fixed number of users served, and their value rises as you grow, right up until they run out in a single heavy month.
Two line items catch UK teams specifically. Data egress is charged per gigabyte leaving the cloud, and credits usually cover it, which hides the cost until the credits end. And every one of these bills is denominated in dollars, so the sterling cost of whatever the credits do not cover moves with the exchange rate. Neither is a reason to avoid credits. Both are reasons to know your uncredited run rate before you need it.
The UK-Specific Layer Most Founders Miss
Beyond the global programs there is a thin but real UK-only layer: national compute allocations, innovation agency funding and bank-run incubator networks whose main value is the partner tier they unlock.
Three things worth knowing:
National compute is allocated, not sold. The UK operates public AI compute through UKRI, including the Isambard-AI system in Bristol and Dawn in Cambridge. Access comes as GPU hours awarded against a proposal rather than as a credit balance you can spend on anything. It suits research-heavy training workloads and suits a consumer app not at all.
Grants are cash, and cash is taxed differently from credits. Innovate UK and similar bodies award money, not vouchers. That is better in flexibility and worse in paperwork, and it interacts with your accounts in ways credits do not.
Accelerators and bank-run labs are routers. Their direct value is usually small. Their real value is that being inside one often moves you into a higher partner tier at the cloud and model providers, which is where the six-figure numbers live.
There is also an R&D tax angle that catches people out. UK R&D relief has, since the 2023 reforms, allowed certain cloud computing and data costs as qualifying expenditure. You cannot claim relief on money you did not spend, so heavy credit usage can reduce the qualifying cost base in the same year. That is almost always still a net win, but the arithmetic is worth running with your accountant rather than assuming.
How to Prioritise Credits
Cloud is the highest-priority layer, models sit above tooling, and a time-boxed grant is worth claiming only once there is a workload capable of consuming it.
Cloud carries the most weight. The cloud platform is the decision everything else sits on top of, and it is the slowest layer to settle. Picking the platform after you have already taken model credits locked to a different vendor is how teams end up running two bills.
Model credits matter most once your traffic shape is known. A grant you cannot spend inside its window is worth a fraction of its face value, so the value peaks when something is in production or close to it.
Tooling is a continuous layer, not an event. Analytics, error tracking, CRM and design tools are small individually and meaningful together, and they are the ones founders skip.
Everything deserves a re-check. Terms, tiers and partner routes change constantly, which is the entire reason a tracked list beats a bookmark folder. Start at getaiperks.com and filter by category.
Which routes are open to you at any moment depends on conditions the providers change without notice, and the current position for every program is listed on AI Perks.
What Founders Get Wrong About Credits
The most expensive mistake is not missing a program. It is treating credits as revenue and building an architecture that only works while someone else pays for it.
Four failure modes worth naming:
Architecting for the credit, not the bill. Teams provision generously while credits cover it, then discover their real run rate on the day the balance hits zero. Check your uncredited monthly cost at least once a quarter, even when you are not paying it.
Applying to one program. Approval is inconsistent and the criteria are opaque. Three approvals out of nine applications is a much better outcome than one out of one, and the applications are mostly the same information.
Burning the window. A grant claimed long before you have production traffic quietly loses most of its usable life, because the clock starts whether or not you are spending. Sequence by expiry, not by excitement.
Flipping to Delaware for credits. If the reason for the reincorporation is access to programs, check first whether the programs actually require it. Most do not. AI Perks lists which ones have a genuine entity requirement.
Frequently Asked Questions
Can UK startups get the same credits as US startups?
In almost all cases, yes. Cloud, model and tooling programs are run globally and accept UK-registered companies. A small number of grants and national compute schemes are UK-only, and a smaller number of US accelerator routes are not open to UK entities. The current position per program is tracked at getaiperks.com.
How much are free startup credits in the UK worth in total?
A realistic assembled stack runs from roughly $10,000 for a pre-product team to well over $150,000 for a funded, AI-heavy company with investor introductions. AI Perks tracks $7.7M in credits across 194 companies, and no single startup qualifies for all of it.
Do I need to incorporate in Delaware to get cloud credits?
No. The major cloud and model programs accept UK-incorporated companies directly. Reincorporation is a fundraising and contracting decision with real cost attached, and doing it purely to unlock credits is almost always the wrong trade. Check the actual entity requirements before assuming.
Do free credits affect my UK R&D tax claim?
Possibly. Relief is claimed on money actually spent, and since the 2023 reforms certain cloud and data costs qualify. Credits reduce what you spend, so they can reduce the qualifying base in that period. The net position is still normally favourable, but confirm the numbers with your accountant.
Which credits should a UK founder apply for first?
Cloud, because it takes longest and everything else sits on it. Then model credits once you know your traffic shape, then tooling on an ongoing basis. Applying out of order tends to strand grants in windows you cannot spend them in. Category filters are on getaiperks.com.
Are UK government grants better than provider credits?
They are different instruments. Grants give cash and flexibility with significant application effort and reporting. Provider credits are faster, larger in headline terms, and locked to one vendor. Most UK startups end up holding both, because they cover different parts of the same bill.
Build from London. Let Seattle and Mountain View cover the infrastructure bill.