Canopy Q3 2026 in Review

At the end of Q2, I wrote that Q3 was about turning capital, technology and validation into mainnet. We did that. On September 7, Canopy mainnet went live, and within three weeks a team outside Canopy had taken an app from launch to its own independent network.

We are still the same small team. What changed this quarter is that the questions people ask us are no longer hypothetical. Does the upgrade process work? Does graduation work? Can someone who is not a blockchain engineer ship something real? We now have answers on the record.

Here is what happened, and what it means.

Mainnet is live

Testnet was where builders practiced. Mainnet is where apps have real users, real economics, and security from the network from the first block. By launch day, testnet had seen more than 1 million deployments, and whatever a builder had tested there carries over to mainnet unchanged, in whatever language they wrote it.

The work leading up to launch was not glamorous, and that was deliberate. Deploying a node now takes minutes. Every node ships with a built-in wallet and block explorer, so operators no longer bolt on third-party tools just to see what their own node is doing. The indexer became roughly 200 times faster with a database about 80 percent smaller, and hosted snapshots cut node sync from a full weekend to about an hour.

None of this is exciting to announce. All of it is what separates a network people can rely on from one they cannot.

Banyan: upgrades without downtime

On September 17 we shipped Banyan (v2.0), our first named protocol upgrade. It went out network-wide through Canopy Auto Update with zero downtime and no interruption to block production.

The reason this matters goes beyond one release and speaks to our thesis as a whole. Most onchain apps change as little as possible, as rarely as possible, because every change means redeploying and moving users and liquidity. On Canopy, every app inherits Auto Update, so a team can schedule a fix or a new feature without a migration. Banyan also improved compatibility with standard Ethereum tooling and made validators lighter to run.

Tools for builders and operators

The goal has not changed: describe what you want, and get a real, sovereign app. This quarter we closed more of the gap between the idea and the running network.

• Templates now cover five languages: Go, TypeScript, Python, Kotlin and C#. The network handles networking, consensus and state, so builders write only their application logic.
• A TypeScript SDK gives builders a fully typed, documented toolkit for working with Canopy.
• Canopy KMS, launched August 28, keeps our highest-value signing keys (treasury disbursements, bridge transfers and chain creation) inside a dedicated signing enclave, with velocity controls and anomaly detection on top.
• Node operator releases added an offline transaction signer, simpler configuration and manual consensus round recovery.

Apps started graduating

The milestone I am proudest of this quarter did not come from us. On September 22, Canoliq, a liquid staking protocol, became the first third-party app to graduate on the Canopy launchpad. It crossed from launch to an independent network producing its own blocks, with its own explorer and wallet support, in about twelve minutes.

Builders kept showing up on testnet too. By late July they had launched more than 793,000 apps, up from 620,000 at the end of Q2. Among them were liquid staking, lending and onchain games, and several were built by people without traditional engineering backgrounds, working with AI. Block Rocket, a community-built multiplayer game running entirely onchain, drew 2,632 players in its first round and 4,515 in its second. Other community builders shipped a reputation scoring app and an open version of a fact-checking algorithm that normally runs on private servers.

Partners who extend what an app can do

In July we announced Canopy Confidential Apps with Fhenix. Using Fhenix's CoFHE coprocessor, data is encrypted in the Canopy wallet, computed on while still encrypted, and returned as signed, verifiable results before it settles. That makes things like sealed-bid auctions and payroll possible onchain without exposing the underlying data. Until now, building that meant hiring a cryptographer. With Canopy, it will be a plugin.

In September we began working with Rayls, the EVM Layer 1 connecting institutions to blockchain liquidity, further entering the enterprise space. We are selective about partners. The test is simple: does this let a builder on Canopy ship something they could not ship before?

What we learned

A launch quarter brings a lot of attention, and not all of it lasts. The airdrop and the run-up to mainnet brought many people to Canopy. Some came for points. The ones who stay are the ones who build, and that is who we are focused on now.

The clearest lesson of the quarter is that proof beats promises. The graduation mechanism, the upgrade process and the key management system were all documented long before Q3. What changed this quarter is that each one was used for real, and the results are public. That is the standard we want to hold ourselves to: ship it, test it where the stakes are low, then show the result.

The road ahead

Q3 took Canopy from testnet to mainnet. Q4 is about the apps. Confidential Apps with Fhenix are scheduled to roll out, more teams will follow Canoliq through graduation, and we will spend more of our time showing the builders and what they ship than talking about ourselves.

We plan to stay small, move fast, and keep letting the product do the talking.