A few weeks ago, when @brevis_zk released the ProverNet whitepaper, I read through it. My first impression was: this isn’t one of those “beautifully written vision ZK architectures,” but rather about turning zero-knowledge proofs into a tradable resource, an open market, instead of each team running their own closed system.

Today, seeing the official launch of the ProverNet mainnet beta, the entire framework has finally taken shape.

Many people might only notice the word “launch,” but the real game-changers are actually three key details:

First, the proof task allocation mechanism has been changed to a continuous bidding system.

It’s not about evenly distributing tasks, nor about fixed queues, but letting provers compete in an open market—whoever bids reasonably, matches computing power, and responds quickly will get prioritized tasks. Essentially, this mechanism turns proving power into a real-time, liquid market resource.

Second, task settlements are directly made with USDC.

This point is crucial for provers:

No waiting for future token value, no relying on speculation, just real cash flow.

This means ProverNet isn’t an incentive experiment, but an economic closed loop operating from Day 1.

Third, the registration process and job matching mechanism for provers are already mature.

In the past, building ZK infrastructure involved not just computing costs, but also setup, maintenance, and fragmented tooling.

Now the entire process is abstracted into a system entry point, allowing direct participation without building your own infrastructure.

After the mainnet officially launches, another key piece will be added:

The staking and payment structure for BREV.

From that moment, ProverNet becomes a truly closed loop.

What surprised me even more is the barrier to entry.

Both GPU and CPU can participate—which means it’s not a certain type of hardware monopolizing the market, but truly an attempt to build an open proving power layer.
From the application side, its appeal is very direct:

You don’t need to build your own ZK proving system—just outsource tasks to ProverNet.

And this directly addresses a pain point the industry has long avoided—ZK is costly and unpredictable.

Especially in real production environments, with irregular loads, bursts of demand, and changing verification standards, self-built systems become expensive, bloated, and unscalable.

Brevis is now abstracting away this layer, essentially turning ZK from a specialized skill into callable infrastructure.

More importantly—the first migration isn’t a demo, but Ethereum block execution proofs.

These are proof tasks already running in production environments, with real workloads and multidimensional verification.

In other words, ProverNet has been stress-tested from day one, not just running experimental workloads.

This “direct connection to the real world” design is rare in the ZK space.
Don’t forget, Brevis has already generated over 250 million proof data points, spanning a huge range of scenarios.

Now ProverNet is distilling this experience into a reusable, competitive, and priceable underlying structure.

If you’re working on ZK, Rollup, off-chain computation, verifiable data, or zkML, this is the most worthwhile time to observe.

The beta version is still in the process of system formation, so you can not only see the design logic but also whether it can form a truly sustainable supply-demand market.

In a nutshell:

This isn’t “yet another ZK network launch,” but an experiment in commoditizing proving power.

Whether it will become industry infrastructure remains to be seen, but it’s already well worth watching closely.

Official materials:

Blog:

Whitepaper:

@KaitoAI #Yapping #MadewithMoss
@MossAI_Official #Starboard @Galxe @RiverdotInc @River4fun