How Much Are Free Web3 and Blockchain Credits Worth?
Published Web3 credit amounts cluster around $5,000 per infrastructure vendor, and the largest cheque in the stack is not a Web3 program at all: it is the general cloud grant underneath it.
QuickNode's startup program runs up to $5,000 in RPC and data credits. 0x JumpStart runs up to $5,000 in swap API credits. Most other node, indexing, wallet and compliance vendors operate partner or accelerator tracks with no published number, negotiated per applicant.
That looks modest next to a six figure cloud grant, and it is sized about right. A Web3 credit covers one specific subscription through the months where that subscription is pure cost and your chain activity is close to zero.
AI Perks tracks the Web3/Blockchain category alongside $7.7M in credits across 194 companies. Eligibility depends on stage and funding, and the live terms sit there.
What Web3 Infrastructure Actually Buys You
You are not buying access to a blockchain. You are buying the right not to run one, plus the query layer that public chains conspicuously do not provide.
A node is free software on a public network. Running a production one is not, and that gap is what every vendor in this category is really selling.
Six layers sit here, and teams routinely budget for one and get surprised by five:
- RPC and node access - read state, submit transactions, subscribe to blocks. QuickNode, Alchemy, Infura, Chainstack and Ankr on EVM chains, Helius and Triton on Solana.
- Indexing and data - a chain is an append-only log, which is the worst possible shape for "show me this user's positions". The Graph, Goldsky, Dune and Covalent exist because you cannot query a chain directly.
- Wallets and key management - embedded wallets, account abstraction and MPC signing from Privy, Dynamic, Web3Auth and Turnkey, institutional custody from Fireblocks and Anchorage.
- Transaction and liquidity APIs - swap routing, gasless relayers, paymasters and bridges from 0x, LiFi, Gelato, Biconomy and Pimlico.
- Compliance and screening - Chainalysis, TRM Labs and Elliptic. Not optional the moment fiat touches your product.
- Storage - IPFS pinning and permanent storage from Pinata, Filebase, Filecoin and Arweave.
Audits sit outside all of this and have no credit program. Budget them as cash.
How Web3 Infrastructure Cost Behaves at Scale
The meter is not requests. It is weighted compute units, and a single badly shaped query can cost more than a thousand cheap ones.
Every major RPC provider prices this way. A call for the current block number is nearly free. A log query across a wide block range, or any read against archive state, costs orders of magnitude more. Your bill tracks query shape, not user count.
| Layer | What the meter counts | What quietly drains it |
|---|---|---|
| RPC and nodes | Weighted compute units per call | Archive state reads, wide log ranges, per-tab polling |
| Indexing | Queries served plus indexing time | Re-syncing from genesis after a schema change |
| Wallet infrastructure | Monthly active wallets | An airdrop that creates wallets which never return |
| Gas sponsorship | Real ETH or SOL, paid on-chain | Paymasters running during a mint or incentive campaign |
| Storage | GB pinned and gateway bandwidth | Serving media through a public gateway |
| Compliance | Addresses or transactions screened | Screening on every wallet connect instead of at cash-out |
Four things generate the invoice nobody forecast.
Polling. A frontend refreshing balances every few seconds, per open tab, is the single largest avoidable cost in this category. Websockets, webhooks and streaming endpoints do the same job for a fraction of the meter.
Archive access. Historical state is a separate, far more expensive tier at every provider. One analytics feature can move your whole account onto it.
Multichain. Every chain you add is another endpoint, another indexer and another allowance. Support quality varies enormously between a provider's flagship chain and its long tail.
Gas. Credits offset your subscription, never the on-chain fee. If you sponsor user transactions, that is real ETH or SOL leaving a real wallet, and no program in this category refunds it.
Web3 and Blockchain Credit Programs Compared
QuickNode and 0x publish $5,000 figures. Most other Web3 vendors run partner or accelerator tracks without a published amount, and the biggest numbers come from the cloud programs sitting underneath the whole stack.
| Provider | Layer it covers | Credit scale |
|---|---|---|
| QuickNode | RPC endpoints, token and NFT APIs, streams | Up to $5,000, published |
| 0x JumpStart | Swap, gasless and cross-chain trade APIs | Up to $5,000, published |
| Alchemy | EVM RPC, enhanced APIs, account abstraction | Startup program, amount varies by applicant |
| Infura (Consensys) | RPC across major EVM networks | Partner and accelerator tracks |
| Chainstack | Shared and dedicated nodes, multichain | Startup track, amounts not consistently published |
| Helius | Solana RPC, webhooks, enhanced transactions | Ecosystem and accelerator tracks |
| The Graph, Goldsky | Indexing and subgraph queries | Free query allowance, then usage-based |
| Privy, Dynamic, Web3Auth | Embedded wallets, account abstraction | Priced per monthly active wallet, startup tiers exist |
| Tenderly | Simulation, debugging and alerting | Startup track |
| Fireblocks, Chainalysis | Custody and compliance screening | Enterprise programs, negotiated per applicant |
| AWS, Google Cloud, Azure | Everything above has to run somewhere | Six figures at top tiers, the largest in the stack |
Where a figure is missing, the program is negotiated rather than published. Current values, eligibility and terms sit in the Web3/Blockchain and Cloud Infrastructure categories on AI Perks.
How to Choose a Provider and in What Order to Apply
Pick the chain first, because the chain narrows the vendor list to a handful. Then apply largest first: cloud, then RPC, then indexing, then wallets and compliance when a customer or a regulator forces it.
Never let a credit choose your chain. $5,000 is not a reason to ship on the wrong network. The chain determines your users, your liquidity and your fee model, and none of that is recoverable later.
Cloud program first. It is the largest single amount available, the hardest one to backfill later, and it covers hosting, storage and bandwidth for every other layer. Filter to Cloud Infrastructure on getaiperks.com.
RPC second, and be relaxed about it. This is the most portable layer in the stack. A provider is a standard JSON-RPC interface behind a URL, so switching is close to a config change. RPC is therefore the one layer where a credit can legitimately decide the vendor.
Indexing third, and be careful. This is the least portable layer. A subgraph or a custom indexer is vendor-shaped work measured in weeks, so choose on query patterns you have actually observed rather than on a credit.
Wallets, custody and compliance last. Buy them when a real user flow or a real regulator demands them. Paying for institutional custody speculatively is the fastest way to waste a Web3 budget.
Ecosystem grants run in parallel, not instead. They are token-denominated, milestone-gated and slow, so pursue them but do not budget against them.
What Founders Get Wrong About Web3 Credits
The two expensive mistakes are assuming credits cover gas, and self-hosting nodes to save a subscription a credit would have covered anyway.
- Assuming credits cover gas. They cover your vendor invoice. Gas, paymaster funding and bridge fees are on-chain costs paid in the network's own asset, and no credit program touches them.
- Self-hosting to save money. An archive node is terabytes, needs constant resync, and puts you on call for someone else's consensus bug. The subscription you avoided is smaller than the engineer you now need.
- Going multichain before product-market fit. Each additional chain multiplies vendors, endpoints, monitoring and support surface for a marginal user base.
- Letting the balance expire. Most Web3 credits are time-boxed and start burning down on approval, so claiming one before you have chain activity wastes most of it. AI Perks tracks terms so you can sequence approvals around real usage.
Frequently Asked Questions
Which Web3 providers offer free credits for startups?
QuickNode publishes up to $5,000 in RPC and data credits, and 0x JumpStart up to $5,000 in swap API credits. Alchemy, Chainstack, Helius, Tenderly and most wallet and compliance vendors run partner or accelerator tracks negotiated per applicant. Eligibility depends on stage and funding. Current programs are listed at getaiperks.com.
Do Web3 credits cover gas fees?
No. Credits offset your vendor subscription: RPC calls, indexed queries, wallet seats, pinned storage. Gas is paid on-chain in the network's own asset at the moment a transaction settles. If you sponsor user transactions through a paymaster or a relayer, that funding comes out of a wallet you top up yourself.
Should I run my own node instead of paying for RPC?
Only if node operation is your product. A production archive node runs into terabytes, needs continuous resync and monitoring, and makes your uptime depend on client releases you do not control. For nearly every application team the managed subscription is cheaper than the engineering time, especially while a startup credit is absorbing that subscription.
Are ecosystem grants better than infrastructure credits?
They are different instruments. A chain foundation grant is usually larger but token-denominated, milestone-gated and slow, and it carries price risk. An infrastructure credit is smaller, lands as a spendable balance quickly and reduces a bill you already have. Pursue grants in parallel, but do not build a budget on one.
Can I stack Web3 credits with AWS or Google Cloud credits?
Yes, and they cover genuinely separate bills. Cloud credits pay for where your backend, indexer and frontend run. Web3 credits pay for chain access, data and wallet infrastructure on top. Holding several at once is how teams fund the whole pre-revenue stretch. AI Perks tracks $7.7M across 194 companies at getaiperks.com.
How much does Web3 infrastructure cost before revenue?
Less than teams fear, until launch. Free tiers at most RPC and indexing vendors comfortably carry a prototype or a hackathon build. The step change arrives with production traffic, archive queries and a second chain, which is exactly the window a $5,000 startup credit is sized to cover.
Ship on the chain you would still choose at list price, then let someone else fund the infrastructure under it.