Node RPC vs. Dedicated Clusters: choosing the right infrastructure for your workload

Most onchain workloads run well on multi-tenant infrastructure. But as teams scale, questions come up: do we need dedicated nodes? Should we isolate certain workloads? What are the actual trade-offs?

This guide breaks down how Alchemy's two infrastructure models — Node RPC, our shared infrastructure solution, and Dedicated Clusters — work, when each one is the right fit, and how to think about the decision for your team.

Node RPC: the default for most workloads

Node RPC is Alchemy's production-grade, multi-tenant infrastructure optimized for elasticity and operational simplicity. It is a keyed private endpoint on a shared fleet. See how the tiers compare in public, private, and dedicated RPC endpoints. It's powered by Cortex, the same engine behind $1T+ in annual transaction volume for teams including Robinhood, Stripe, Coinbase, Circle, Chainlink, and Polymarket.

On Node RPC, scaling is near-instant. Traffic bursts are absorbed automatically across the node fleet. Failover is built in across multiple regions. Pricing is usage-based, so costs scale with your actual traffic rather than provisioned capacity.

For current shared RPC performance across providers, see Alchemy's RPC provider benchmarks.

Dedicated Clusters: endless customization, fully managed

Dedicated Clusters provide your own node infrastructure — provisioned, operated, and maintained by Alchemy, but configured to your exact requirements. Unlike single-node dedicated offerings, each cluster provisions a redundant group of nodes per chain, running on Cortex, the same engine as Node RPC. You get the same APIs and reliability guarantees in a single-tenant environment. For a deeper look at how single-tenant clusters work, see how dedicated blockchain infrastructure works.

Every cluster is built for zero downtime. Two or more nodes per chain per region enables rolling maintenance with no interruptions. Block-perfect consistency ensures every node returns the same view of chain state, eliminating errors from stale or conflicting reads. And real-time Grafana dashboards give you full visibility into node health, request patterns, and performance.

As your traffic grows, we expand your cluster alongside you — our automated snapshot and deploy capabilities get new capacity live faster than anyone in the industry. For unexpected spikes that exceed your contracted capacity, you can opt in to automatic fallback to Alchemy's shared fleet — no dropped requests, no manual intervention required.

Every cluster is configured to your exact requirements:

• Custom tracers and binaries deployed directly on your nodes for faster, more cost-effective simulation, tracing, and indexing
• Single-tenant isolation — SOC 2 Type II compliant, with no other customer's traffic, code, or data in your environment
• Regional deployment for low-latency proximity to your stack, chain infrastructure, or users
• Custom hardware configurations tailored to your traffic for the best possible performance
• Fixed monthly pricing — no per-request billing, no surprises

When Dedicated Clusters are the right choice

Node RPC handles the vast majority of workloads well. Dedicated Clusters are designed for teams with specific requirements that multi-tenant infrastructure can't satisfy. Here are four common scenarios.

Custom tracers or binaries. Workloads that depend on non-standard node software — custom EVM tracers, clients, or binaries — require the ability to push custom code to the node. Dedicated Clusters support this natively, making them the right fit for security and forensics teams that rely on custom tracers for indexing and event processing.

Regulatory or internal isolation requirements. Certain compliance frameworks and internal security policies require that no other customer's traffic, code, or data touches the environment. Dedicated Clusters provide single-tenant isolation with audit-ready controls and SOC 2 Type II compliance — a standard requirement for regulated financial institutions.

Regional deployment requirements. Node RPC already provides low-latency coverage across a wide set of regions. But if your workload requires deployment in a specific geography outside that footprint — co-located with your stack, sequencers, or validators — Dedicated Clusters can be deployed in the exact region you need. This is particularly relevant for trading firms, DeFi protocols, and high-frequency trading operations where every millisecond matters.

High-volume workloads where capacity-based pricing is more efficient. For very high-throughput, multi-chain workloads, a fixed monthly cost based on provisioned capacity can be more predictable than per-request billing.

If none of these requirements apply, Node RPC delivers the same best-in-class performance and reliability, with no additional setup needed.

Why not run nodes in-house?

Many teams we work with have considered — or are currently managing — their own node infrastructure. The experience tends to follow a familiar trajectory: significant investment in hiring, tooling, and monitoring, followed by ongoing operational load from client upgrades, network changes, and on-call rotations across every supported chain. For a deeper look at this trade-off, see our overview on the pros and cons of running your own node.

The risk compounds over time. A missed upgrade causes nodes to fall behind, requests begin failing, and customer-facing transactions break — users can't transact, frustration builds, and every minute of downtime means lost revenue and lost trust. Also, engineering resources that could be focused on product work are instead spent maintaining infrastructure that doesn't differentiate the business.

Running your own nodes vs. a managed cluster

Alchemy's managed cluster is Dedicated Clusters. Here's how the two compare:

This is the problem Dedicated Clusters are designed to solve. Alchemy handles deployment, upgrades, monitoring, and incident response end to end.

The hybrid approach

Teams that need Dedicated Clusters for certain workloads typically don't need it for all of them. The most common production setup is hybrid: Dedicated Clusters for the chains or workloads that require single-tenant control, and Node RPC for everything else.

Migration between the two is straightforward. Both use the same Alchemy APIs — routing a workload to Dedicated is as simple as pointing requests to a different endpoint URL. No code changes, no re-architecture required.

This approach gives teams single-tenant control where it matters, while preserving the elasticity and cost efficiency of Node RPC for the rest of their traffic.

Comparison

How to decide

The decision comes down to four triggers. Move a workload from shared (Node RPC) to dedicated when any of these is true:

• Custom tracers or binaries: your workload depends on non-standard node software you need to deploy.
• Regulatory or isolation mandate: a compliance framework or internal policy requires single-tenant compute.
• Region outside the Node RPC footprint: you need nodes co-located with your stack, sequencer, or validators in a geography shared infrastructure doesn't cover.
• Sustained high volume: capacity-based pricing is more efficient than per-request billing at your throughput.

If the answer to any of these is yes, Dedicated Clusters or a hybrid setup are worth evaluating. If the answer to all of them is no, Node RPC is the right fit.

In either case, the path forward is flexible. Teams can start on Node RPC and move specific workloads to Dedicated later without disrupting existing infrastructure.

Frequently asked questions

What is Alchemy Node RPC?

Node RPC is Alchemy's multi-tenant infrastructure, powered by Cortex. It's optimized for elasticity and operational simplicity: scaling is near-instant, failover is built in across multiple regions, and pricing is usage-based.

What are Alchemy Dedicated Clusters?

Dedicated Clusters give you your own node infrastructure, fully managed by Alchemy but configured to your exact requirements. They run in a single-tenant environment with support for custom tracers, regional deployment, custom hardware, and fixed monthly pricing.

What's the main difference between Node RPC and Dedicated Clusters?

Node RPC is shared infrastructure with usage-based pricing, built for most workloads. Dedicated Clusters offer single-tenant isolation with custom configurations for teams that need specific compliance, performance, or customization controls.

When should I choose Dedicated Clusters over Node RPC?

Dedicated Clusters are the right fit if you need custom tracers or binaries, have regulatory or internal isolation requirements, need deployment in specific regions, or run high-volume workloads where capacity-based pricing is more efficient.

Can I use both Node RPC and Dedicated Clusters together?

Yes. A common production setup is hybrid: Dedicated Clusters for workloads that require single-tenant control, Node RPC for everything else. Migration is as simple as changing the endpoint URL, with no code changes required.

Are Dedicated Clusters harder to maintain than Node RPC?

No. Both are fully managed by Alchemy. Deployment, upgrades, monitoring, and incident response are all handled end to end. Zero maintenance required from your side.

Do both Node RPC and Dedicated Clusters support the same APIs?

Yes. Both run on the same Alchemy APIs, so routing workloads between them requires no code changes or re-architecture.

Should I run my own nodes or use a managed cluster?

Running your own nodes means significant investment in hiring, tooling, and ongoing operational load: client upgrades, network changes, and on-call rotations across every supported chain. The cost that matters is rarely the servers; it's the engineering time, and the product work that time displaces. A managed Dedicated Cluster shifts deployment, upgrades, monitoring, and incident response to Alchemy for a fixed monthly fee, so your team ships product instead of maintaining infrastructure. See the comparison above for the full breakdown.

Get started

Over 70% of top onchain applications run on Alchemy. If you're evaluating infrastructure options — whether you're consolidating providers, replacing an in-house setup, or planning for scale — reach out to our team to find the right configuration for your workload.