Google Cloud will discontinue its Blockchain Node Engine and Blockchain RPC products on December 15, 2026, and move existing customers to QuickNode, according to a report shared by Cointelegraph on October 6, 2026. The decision pulls one of the largest cloud vendors out of the business of running blockchain nodes directly and hands that work to a specialist provider.
Blockchain Node Engine launched in 2022 as Google Cloud's managed service for running dedicated Ethereum nodes, later expanding to other networks. The pitch was simple: instead of a developer provisioning, syncing, and maintaining their own node, Google would host it. Blockchain RPC extended that with endpoints applications could call to read chain data and broadcast transactions. Both are now on a clock.
The handover to a specialist
Rather than leaving customers to find their own replacement, Google is directing them to QuickNode, a company whose entire business is node infrastructure and RPC access across dozens of chains. For a large cloud provider, running blockchain nodes is a narrow, maintenance-heavy line of business that sits far from its core compute, storage, and AI products. Passing it to a firm built around that single function is a cleaner outcome than an abrupt cutoff.
The shift still forces work on the teams that depend on these endpoints. Any application pointing at a Google Cloud RPC URL has until December 15 to repoint to QuickNode or another provider, update authentication, and confirm the new endpoints behave identically under load. Migrations like this are rarely a one-line change. Rate limits, supported methods, and response formats can differ between providers, and payment platforms that settle on-chain need those details to match exactly before a cutover.
Infrastructure concentration is the real story
The retreat is a reminder of how much of crypto runs on a small number of node and RPC providers. Wallets, exchanges, and card programs that convert crypto to fiat all lean on these endpoints to check balances, read token prices, and push transactions to the network. When one provider exits, demand consolidates onto those that remain.
That concentration carries a cost that users rarely see directly. An RPC outage or rate-limit change at a major provider can slow or stall transaction broadcasting for every app wired to it. For a crypto card that tops up balances on-chain or settles stablecoin payments, reliable node access is not a backend nicety. It is the difference between a payment clearing and a payment hanging. Fewer independent providers means fewer fallbacks when one has a bad day.
Pricing is the second-order effect worth watching. Google Cloud competed on the strength of its brand and its bundled billing. With it gone, developers choosing between the remaining specialists have one fewer option at the enterprise tier. Whether that nudges RPC pricing up over time is the open question the December deadline sets in motion.
Two months to migrate off Google's endpoints
Teams running production workloads on Google Cloud's blockchain products have about two months to act. The practical sequence is to inventory every service that calls a Google RPC endpoint, stand up equivalent QuickNode endpoints in a staging environment, and run both in parallel long enough to confirm parity before flipping production traffic. Leaving the switch to the final week invites the exact kind of failure a migration is supposed to avoid.
For the broader market, the direct price impact is minimal. Bitcoin traded at $85,568 and Ether at $2,701 as of October 6, 2026, with the Fear and Greed Index at 67 (Greed), and an infrastructure vendor change of this kind does not move spot prices. The significance is structural, not immediate.
Overview
Google Cloud is sunsetting its Blockchain Node Engine and Blockchain RPC on December 15, 2026, and steering customers to QuickNode. The move is a measured exit, not a disruption, since it comes with a designated replacement. The takeaway for anyone building or operating on these endpoints: treat the deadline as a hard migration date, test for parity before cutting over, and note that crypto's node layer just got a little more concentrated.
