How Alchemy's node infrastructure keeps DeFi fast at scale

Every time someone makes a trade on a DeFi app, swapping one token for another, the app fires a burst of requests to the blockchain in the background: price quotes, pool balances, gas estimates, wallet balances. Dozens of them, across several chains, before the trade goes through. The user never sees any of it. They just see a good price and a confirmed transaction, or they don't.

In Decentralized Finance (DeFi), the infrastructure is the product. When the data your app reads is slow, the user may trade at a different price than the one they saw. When a connection drops during a market swing, a transaction fails. When a record goes missing, someone's balance or history shows up wrong. The blockchain itself does not fix any of this for you. Your RPC layer does, or it doesn't.

We've spent years investing in the node and RPC infrastructure that closes the gap between what the blockchain knows and what your app can act on. That work matters for any onchain product, and DeFi is where it gets tested hardest, because in DeFi that gap is measured in money. Here's what it looks like in production, told through the teams running on it.

Speed, because every millisecond is priced in

In trading, the price moves while you wait. A few hundred milliseconds of lag can be the difference between the price a user saw and the price they actually get.

0x powers swaps for dozens of wallets and aggregators trading one token for another, routing billions of dollars in trades. 0x's goal is to find the best available price across many pools of liquidity at once. To do that, it fires dozens of RPC queries in the moment before each trade. If that data comes back slow, the price the user is quoted drifts, and they get a different deal than the one they intended to make.

0x tested every option and settled on one requirement: fast, consistent, accurate data every time. On our infrastructure they generate quotes in under a second and hold 99.9% uptime with one of the lowest failed trade revert rates in DeFi. For 0x, that reliability is what lets them promise their partners a dependable service. For the trader on the other end, it is the difference between a trade that clears at the price they expected and one that silently fails.

"Alchemy allows all of our integrators to have more responsive experiences and it allows us to offer a more reliable service — we're able to generate quotes in less than a second."

— Charles Reese, Engineering Manager, 0x Labs · Read the 0x case study

0x on Alchemy: $170B+ traded, 3.5M+ traders, 99.9% uptime, 2% revert rate, sub-second quotes.

The same pattern holds on Solana, where blocks come even faster. Solflare, one of the largest wallets in the Solana ecosystem, gives its 4M+ users a deep view of their own activity: balances over time, past trades, and profit and loss. All of that runs on historical blockchain data, which is one of the hardest things to serve quickly on Solana. On our Solana archive infrastructure, Solflare gets that data up to 20 times faster than other options, with responses under 200 milliseconds and no errors even while handling several thousand requests every second. For Solflare, that means shipping richer analytics without fighting its data layer. For its users, it means opening the app and seeing a full, accurate history load almost instantly instead of waiting or hitting gaps.

"Alchemy's Archival RPC has kept up with Solflare's throughput while keeping a consistent sub-200ms response time."

— Dušan K., Backend Engineer at Solflare · Read the Solflare case study

Solflare on Alchemy: 4M+ users, up to 20× faster archive queries, responses under 200ms, several thousand requests per second with zero errors.

Reliability, because a failed transaction is real money lost

Speed only counts if the system stays up when it matters most. For DeFi, the worst moment to fail is the busiest one: a big launch, a token distribution, or a wave of liquidations when the market moves fast.

Usual, a stablecoin protocol, found this out ten hours before a major token distribution, when the team recalculated demand and realized their existing setup would not hold. We scaled their capacity, tuned their infrastructure, and stayed in the room with live engineering coverage through the event. At the peak they handled 30,000 requests per second and served 38 million requests in the first hour, while more than 200,000 people tried to claim their tokens at the same moment, with effectively zero downtime and no failed claims. For Usual, that meant a signature launch went off without an outage becoming the story. For each of those 200,000 people, it meant they actually got what they came for instead of a spinning screen or a failed transaction. Our part was making sure the infrastructure underneath never became the thing that broke.

"Alchemy was able to give us a near perfect availability, serving massive amounts of requests per second (30k requests per second was our peak, and we still had buffer), without any hiccups or downtime. And all on short notice!"

— Adli Takkal-Bataille, cofounder at Usual · Read the Usual case study

Usual on Alchemy: 30,000 requests per second at peak, 38M requests in the first hour, 200,000+ simultaneous claims, zero downtime and no failed claims.

Ease of access, so your engineers can build what sets you apart

The best reason to run on us is what it frees your team to do. Every hour your engineers spend babysitting nodes, widening block history, or waking up to a pager during a traffic spike is an hour they are not spending on the product that sets you apart.

That is the trade each of these teams made with us. Usual scaled on ten hours' notice without adding infrastructure staff. 0x brought its data onto one provider instead of stitching several together. Solflare prototyped its own Solana archive system, saw that running it well would be a full-time job, and handed that job to us so its engineers could keep shipping features. In each case the differentiating work stayed with them, and the infrastructure heavy lifting came to us.

Who is this for?

DeFi isn't one workload. Whatever you're building, the failure you can't afford is the one that only shows up under load. Here is where the pressure lands hardest:

• Swap routers and aggregators. You fire dozens of RPC calls to price a single trade, so a slow read moves the price your user gets.
• Wallets and analytics apps. You show users their balances, history, and performance, so you live on fast reads of historical data.
• Stablecoins, RWAs, and token launches. Your riskiest day is your own launch or redemption window, so you need capacity that scales in hours, not quarters.
• Lending, perps, and liquidation engines. One dropped update during a market swing can mean a missed liquidation, so real-time reliability is the whole game.

Why we keep investing here

None of this is an accident, and none of it is a side project. Powering this kind of infrastructure is what Alchemy is; it is the core of who we are and what we have built the company around. All of it runs on Cortex, the engine underneath our RPC, trained on trillions of requests and years of running this exact workload. It routes each request along the fastest path, scales automatically when traffic surges, and reroutes around failures on its own. Simpler setups tend to slow down or turn requests away at the exact moment volume spikes; Cortex is built to do the opposite, absorbing the spike and keeping data accurate when it matters most. That is what lets us hold low response times and high uptime, measured continuously in our public RPC benchmarks, while powering $1T+ in annual transactions across the platform.

DeFi is the segment that stress-tests all of it at once: latency-sensitive, spiky, and unforgiving when money is on the line. The teams building the fastest onchain markets don't want to think about their infrastructure at all. Our job is to make sure they never have to.

If you're building in DeFi and your infrastructure is the thing slowing you down talk to our team and we'll stand up capacity with you.

Frequently asked questions

What is the best RPC provider for DeFi?

The right RPC layer for DeFi has to combine consistent low latency, block-perfect data accuracy, and reliability that holds during volatility spikes. Teams like 0x, Usual, and Solflare run latency-sensitive and high-traffic DeFi workloads on Alchemy for exactly these reasons.

How does Alchemy keep DeFi apps fast during traffic spikes?

Alchemy's RPC runs on Cortex, which handles smart routing, elastic scaling during volatility, and autonomous failover. Usual peaked at 30,000 requests per second during a token distribution with effectively zero downtime and no failed claims.

Does Alchemy support Solana archival data for DeFi?

Yes. Alchemy serves heavy Solana methods like getTransaction and getProgramAccounts on its Solana Archive Method, where Solflare sees up to 20x faster archive queries and sub-200ms response times. See how Alchemy built the fastest archival methods on Solana.

How many chains does Alchemy support?

Alchemy supports 100+ chains behind one interface, so adding a new network is a configuration change rather than a rebuild.