Brevis_ZK

TL;DR
Brevis Vera is now open for anyone to use end-to-end at vera.brevis.network. Capture an image on any C2PA-enabled camera or phone, bring it into Vera, apply edits that get cryptographically attested as legitimate, and download a proof file that anyone in the world can verify in their own browser. Vera introduces zero-knowledge proofs as a new primitive in the C2PA stack, designed to preserve the privacy of both the source media and the editing process across the full chain of custody.
From Concept Demo to Working Tool
In March, we introduced Brevis Vera (https://blog.brevis.network/2026/03/09/brevis-vera-proving-whats-real-in-the-age-of-ai/) with an interactive demo that walked visitors through the idea: hardware-backed capture, zero-knowledge proofs of every edit, browser-side verification. The concept landed, but the tool was only a guided illustration at that point.
Today, the full version is open!
Anyone with a C2PA-enabled device can capture a real image, sign it at the source, bring it into Vera, apply real edits that get cryptographically attested, and produce a proof that any third party can verify on the spot. The same site that previously walked you through the concept now lets you run the workflow on your own photos.
What the Working Tool Does
The Editor accepts a C2PA-signed source image and surfaces the capture metadata: issuer, public key, signature, image hash. From there you apply a growing set of supported transformations: crop, exposure, color adjustments, brightness, denoise, and other edits working photographers reach for first. Each edit you make becomes part of the provable edit path.
When you're ready to publish, Vera generates a single proof file in the background. The proof captures the original signature, every transformation applied, and the chain that ties one to the other. You walk away with two files: the edited image as a PNG, and the proof as a .bvproof file.
The Verifier is a separate flow. Anyone can upload an edited image plus its proof file and within seconds get a cryptographic confirmation that the image traces back to a real capture device, that only the disclosed transformations were applied, and that nothing was hidden or added along the way. Verification happens entirely in the browser.
Where ZK Adds to the C2PA Stack
C2PA support is already showing up in the editing tools photographers use most. Editor tools like Adobe Photoshop sign in-app edits either through a key embedded locally in the application or through their own cloud servers. Both work, and both have known tradeoffs. The local-signing path exposes the signing key to reverse engineering, and the cloud-signing path involves uploading sensitive data to a third-party server, potentially including the original media and the full edit history.
Vera's ZK approach removes the extractable signing key problem entirely. The privacy benefit, where source media and editorial workflow stay private through the signing process, is realized when ZK proof generation is embedded inside local editor tools. We've built Vera as a complementary primitive that existing editing tools could adopt as the C2PA ecosystem matures, rather than as a competitive replacement.
Why It Matters Now That It Works
The March announcement made the case for the broader category shift: away from detecting fakes after the fact, and toward proving provenance from the source. With the tool open to everyone and the ZK primitive starting to surface as a real upgrade path for the C2PA stack, that shift becomes practical.
A photojournalist filing from the field can publish images that come with cryptographic proof of where they were captured and how they were processed, in a way that an editor, a lawyer, or a reader can verify independently. A stock photo platform can require provable provenance on submissions and run the check automatically. A platform fact-checking team can verify a viral image without trusting any of the parties that handled it on the way in. Once a proof exists, none of these workflows require Brevis or any other party to be in the loop. The proof speaks for itself, and anyone can verify it in their own browser.
Get Started
Try it: vera.brevis.network.
If your camera or phone hasn't been set up for C2PA yet, we put together a short guide covering the most common devices and how to switch on the feature: C2PA Setup Guide (https://blog.brevis.network/2026/05/07/c2pa-setup-guide-turning-on-content-credentials-on-your-camera-or-phone/).
Or try Vera right now with this sample image (https://drive.google.com/file/d/1CApx0zJiaDjxXO4ySqklHUu6iMi4lN99/view?usp=sharing).
The reference implementation is open source on GitHub.
What's Next
Over the coming weeks, you'll start seeing Vera in the hands of professional photographers working across nature, wildlife, portrait, and editorial photography. They've been part of a private pilot, and as the launch week unfolds they'll be posting their own captures, edits, and proofs to their own audiences with their own take on what verifiable provenance changes about working with photographs in 2026. We're also lining up a series of podcasts and conversations with the same group, starting this week.
That photographer wave is a demonstration of what becomes possible when zero-knowledge proofs join the C2PA standard. Brevis is in the business of building the best ZK technology in the world, not the next photo editor. Our real goal with Vera is to establish ZK as a primitive in the C2PA stack and to see it adopted by the editing tools, camera manufacturers, and platforms already participating in the ecosystem. The security and privacy gains introduced by our ZK technology merit a reimagining of how the industry operates.
If you're a creator, publisher, or platform interested in building on Vera, reach out through the partner form (https://form.typeform.com/to/lduiliob?typeform-source=brevis.network).

TL;DR: Pico Prism 2.0 proves Ethereum mainnet blocks at an average of 6.1 seconds on the network’s current 60M gas limit, with 99.9% of blocks finalizing within the 12-second slot. The full setup runs on 16 RTX 5090 GPUs across two machines at roughly $100K in total hardware cost. Tested against Pico Prism 1.0 on the same 60M gas baseline, the new system delivers ~5.3x more proving efficiency per block.
In February, we previewed Pico Prism’s transition to a 16-GPU dual-machine setup, with early results on the same 45M gas benchmark blocks Pico Prism 1.0 had tested.
That transformation is now complete.
Pico Prism 2.0 is officially live, fully optimized, and benchmarked directly on the production 60M gas blocks Ethereum runs today.
The 2.0 release is a full-stack rebuild across the zkVM ISA, the distributed proving architecture, the emulator, and the GPU proving backend. The result is a system that proves larger blocks on a quarter of the hardware Pico Prism 1.0 used, faster on average, and lands squarely on the Ethereum Foundation’s headline real-time proving targets.
The Headline Results
Pico Prism 2.0 was benchmarked on 1000 consecutive Ethereum mainnet blocks starting at block 24,000,000, on the network’s current 60M gas limit.
MetricResultAverage proving time6.1sBlocks proven within 12s 99.9%Hardware16 RTX 5090 GPUs across 2 machinesTotal hardware cost~$100KBlock gas limit60M (current Ethereum mainnet)
For a fair head-to-head, Pico Prism 1.0’s 64-GPU configuration was re-tested on the same 60M gas blocks. The 1.0 system averages 8.1 seconds per proof. Pico Prism 2.0 hits 6.1 seconds on a quarter of the hardware, which works out to a ~5.3x improvement in compute work per block:
Pico Prism 1.0: 8.1s × 64 GPUs ÷ 60M gas = 8.64 GPU-seconds per million gas
Pico Prism 2.0: 6.1s × 16 GPUs ÷ 60M gas = 1.63 GPU-seconds per million gas
→ ~5.3× efficiency
Chief among the Ethereum Foundation’s real-time proving targets are sub-10-second average proving latency and on-prem hardware capex under $100K. Pico Prism 2.0 smashes both, running on consumer GPUs any team can buy off the shelf.
Benchmarks are fully reproducible. Binaries are available at https://github.com/brevis-network/pico-ethproofs.
Inside Pico Prism 2.0
Four upgrades land together in the 2.0 release. Each is significant on its own, and together they produce the 5x lift.
1. From RISC-V 32IM to RISC-V 64IM
Pico’s zkVM execution environment has moved to RISC-V 64IM, replacing the earlier 32-bit ISA. The 64-bit instruction set matches how real programs are written, giving Pico a richer execution environment and shorter execution traces on most workloads. The system trades a slightly more elaborate chip set for fewer cycles per program, and on real blocks, fewer cycles is what counts.
RISC-V 64IM is fully functional in Pico Prism 2.0. Formal verification of the new ISA implementation is underway.
2. A New Two-Machine Architecture
Pico Prism 2.0 runs on two machines, each with 8 RTX 5090 GPUs, connected over a 100 Gbps interconnect. At the center sits a global scheduler that operates as a shared task board for the proof pipeline. Both machines pull work dynamically from the scheduler rather than receiving statically partitioned assignments.
The architecture is built around three principles:
Global scheduling. Unfinished tasks live in a shared pool. Either machine can claim them as it frees up, which keeps GPUs busy instead of waiting on upstream work to complete.
Data locality. Each machine independently runs the same emulation and produces consistent local records. The scheduler only needs to dispatch task indices rather than heavy intermediate artifacts. Where local tasks are available, machines prefer them, keeping cross-machine traffic minimal.
Maximum parallelism. Combine and RISC-V chunk tasks are pulled from the proof tree dynamically, with autonomous load balancing across all 16 GPUs.
The result is a proving pipeline that behaves as a distributed work queue rather than a fixed sequence.
3. Ahead-of-Time Emulation
Pico Prism 1.0’s emulator interpreted programs at runtime, decoding and dispatching every instruction on the fly. The 2.0 emulator runs natively compiled Rust generated directly from ELF binaries, removing per-instruction decode and dispatch overhead entirely.
Frontend efficiency matters more than it sounds, because real-time proving is a balanced pipeline. If emulation can’t feed work to the GPUs fast enough, the GPUs sit waiting. AOT compilation removes a meaningful share of that frontend cost and keeps the proving stack continuously fed.
4. A Complete CUDA Rewrite
Pico’s GPU backend has been rewritten from the ground up, with deep optimization across the components that sit on the critical path of every proof. FRI commitment now uses adaptive LDE batch NTT, FRI opening uses Montgomery batch inversion, and quotient computation runs through a JIT compiler with an optimized constraint IR.
The rewritten stack delivers immediate speed gains and provides a cleaner, more extensible foundation for future GPU architectures and proving system advances.
Looking Ahead
The race to real-time Ethereum proving has been the defining challenge of the zkVM space for the past two years. In December 2025, the Ethereum Foundation declared the performance race effectively won and shifted focus to soundness foundations for L1 zkEVM integration through 2026.
Pico Prism 2.0 is the production system for the performance side. Going forward, work continues on the soundness side. Brevis is actively contributing alongside the EF’s security roadmap to ensure Pico Prism meets the 128-bit provable security target set for L1 zkEVM integration, with formal verification of the new RISC-V 64IM ISA already underway as part of that work.
In March 2026, Brevis was selected as one of four prover teams in the Ethereum Foundation’s On-Prem Proving Initiative through Ethproofs. The pilot funds the cohort to prove 1-in-10 Ethereum L1 blocks on self-owned hardware under real-world conditions, testing whether ZK proving can scale as decentralized infrastructure rather than depending on a handful of cloud providers. The program kicks off in May 2026 and is the closest thing yet to a dress rehearsal for live L1 zkEVM integration.
Each step on that path takes Pico Prism closer to ZK becoming part of Ethereum’s core infrastructure.
About Brevis
Brevis is a verifiable computing platform powered by zero-knowledge proofs, serving as the infinite compute layer for Web3. Applications can offload expensive computations off-chain while proving every result on-chain. The Brevis stack includes Pico zkVM for general-purpose computation, the ZK Data Coprocessor for trustless access to historical blockchain data, Pico Prism for real-time Ethereum block proving (99.8% coverage on 16 GPUs, hitting the Ethereum Foundation’s $100K hardware target), Vera for ZK-proven media authenticity, and ProverNet, the decentralized marketplace for ZK proof generation now running on mainnet. To date, Brevis has generated 340M+ proofs across 50+ protocols on 8+ blockchains.

TL;DR: Brevis ($BREV ) is partnering with Camp Network, the IP-focused Layer 1 blockchain, to bring zero-knowledge verification to intellectual property licensing, royalty distribution, and ownership attestation. The collaboration addresses three core challenges that AI has created for the IP economy: proving ownership without exposing sensitive information, computing complex royalty splits at scale, and enabling confidential licensing for high-value creative assets. Camp provides the IP infrastructure, including registration, licensing, remix attribution, and royalty mechanisms. Brevis provides the ZK trust layer for data authenticity, off-chain computation, and privacy.
AI Broke Traditional IP Protection
AI has fundamentally changed how creative work gets produced. Images, characters, music, stories, and entire franchises can now be remixed or replicated in seconds. The systems that protect creators haven’t kept pace.
Three problems stand out:
IP abuse and unlicensed derivatives. AI models freely generate outputs in the style of artists or popular franchises without attribution, licensing, or compensation.Broken incentive tracking. Even when creators want to license their work, there’s no practical way to track who used what, which derivative generated revenue, or how royalties should flow through complex chains of AI-assisted co-creation. Running that kind of multi-hop attribution on-chain is too slow and too expensive.Privacy risk for IP owners. IP owners who want to participate in licensing often can’t because the process requires exposing contract details, corporate information, or metadata about unreleased works.
Camp Network is built to solve these problems. As a purpose-built L1 for intellectual property, Camp provides the infrastructure for on-chain IP registration through its Origin framework, remix licensing, automated royalty splits, and a 200+ project ecosystem of creators, AI developers, and licensing platforms. IP assets are tokenized as ERC-721 NFTs with traceable provenance, usage rights, and attribution baked in at the protocol level.
What Camp needs is a trust layer that can handle verification, computation, and privacy without compromising the data it’s protecting. That’s where Brevis comes in.
Proving Ownership Without Revealing Anything
Most valuable IP starts off-chain. Art, characters, brand assets, storylines, datasets. To bring these assets into Camp’s on-chain registry trustlessly, there needs to be a way to verify their authenticity and the claimant’s ownership without requiring the creator to upload sensitive documents.
Brevis zkTLS solves this by proving that data originates from a specific authenticated source without exposing the data itself. Combined with Camp’s Origin registry, this creates a verification path where creators can establish ownership cryptographically.
Consider an animation studio that wants to license a character into an AI model. Normally, this would require uploading copyright documents, contract metadata, and studio information to prove legitimacy. With zkTLS, the studio instead generates a proof that the asset originates from the official copyright holder’s server and that the claimant has legitimate ownership. No files, no images, no legal documents are exposed. The AI model gets a green light for licensed usage, and the studio’s confidential information stays confidential.
Camp’s AI verification tool handles the initial authenticity assessment, while Brevis generates the ZK proof and posts it to Camp’s blockchain, which then mints the IP license. The creator’s ownership is now on-chain and verifiable, without any sensitive data ever leaving their control.
Scaling Royalties for AI Co-Creation
AI-generated content has complicated ancestry. A remix of a remix, multiple upstream creators, model-assisted transformations, multi-party contributions. Tracking who deserves what share of revenue through these derivation chains is computationally intensive work that doesn’t belong on-chain.
Brevis handles this by offloading the entire computation, including royalty splits, attribution graphs, and usage metrics, then generating a ZK proof that the results are correct. Camp verifies the proof on-chain and executes the distributions for a fraction of what native on-chain computation would cost.
Take a scenario where a user remixes a Camp-registered artwork using one of Camp’s ecosystem tools. The remix has twelve upstream creators who each contributed at different stages. Brevis computes the entire derivation graph off-chain, calculates each creator’s revenue share based on Camp’s royalty rules, and returns a zero-knowledge proof alongside the final payout amounts. Camp verifies the proof and executes instant, gas-efficient distributions.
Creators get paid, the computation stays affordable, and every payout is provably correct.
This is what makes AI-native licensing markets viable. Without scalable, verifiable royalty computation, the economics of multi-party creative collaboration fall apart. The attribution chains are too deep and the parties too numerous for on-chain execution, but the financial stakes demand trustless verification.
Confidential Licensing for High-Value IP
Some IP transactions involve parties that need confidentiality as a prerequisite for participation. Studios, investors, and DAOs may want to license premium franchises or verify catalog value, but they can’t do so if the process requires revealing the franchise name, contract terms, valuation details, or corporate identity.
With Brevis zkTLS integrated into Camp’s framework, IP owners can prove selective facts about their holdings without exposing the underlying information. A studio could generate proofs like:
“Our IP catalog valuation exceeds $100M”“We possess full licensing rights for commercial use”“Verified documents back this claim”
All without naming the franchise, showing revenue breakdowns, or revealing corporate details.
This opens up an entirely new category of IP transactions. Investor DAOs can evaluate licensing opportunities based on cryptographic attestations rather than NDAs. Studios can explore partnerships without competitive exposure. AI developers can access legally compliant assets with verified provenance without the IP owner sacrificing any confidential information in the process.
Camp’s IP Layer, Brevis’s Trust Layer
Camp handles registration, licensing, remix attribution, and royalty mechanisms. Brevis handles authenticity proofs, off-chain computation, and privacy verification. Together, we form the infrastructure for programmable IP in an era where AI has made traditional protection models obsolete.
The combined stack means IP can be licensed, verified, and monetized without any party giving up more information than necessary. Ownership proofs, royalty computations, and confidential valuations all run through the same ZK infrastructure, keeping the heavy lifting off-chain while the on-chain verification costs stay negligible.
The collaboration is actively in development across all three areas. More details will follow as the integration progresses.
About Brevis
Brevis is a verifiable computing platform powered by zero-knowledge proofs, serving as the infinite compute layer for Web3. Applications can offload expensive computations off-chain while proving every result on-chain. The Brevis stack includes Pico zkVM for general-purpose computation, the ZK Data Coprocessor for trustless access to historical blockchain data, Pico Prism for real-time Ethereum block proving (99.6% coverage, 6.9s average), and ProverNet, a decentralized marketplace for ZK proof generation. To date, Brevis has generated hundreds of millions of proofs across 40+ protocols on 6 blockchains.

TL;DR
Brevis ($BREV) Vera is an end-to-end media authenticity attestation system that lets anyone verify a published image or video originated from a real device and was only edited in provable, legitimate ways. By combining hardware-backed capture signatures (C2PA) with zero-knowledge proofs of the editing process powered by Brevis Pico zkVM, Vera preserves authenticity from capture through every edit to final publication. Now live.
The Trust Problem
Every day, millions of images and videos are shared online with no way to verify whether they’re real. Deepfakes have gotten good enough that even trained eyes can’t reliably tell the difference, and the tools to create them are becoming more accessible by the month. The default reaction to any striking image online has shifted from curiosity to suspicion.
The obvious response has been to build better detectors. Train AI models to spot AI-generated content. But this approach has a fundamental flaw: it’s a moving target. Every improvement in detection is matched by an improvement in generation. The two sides are locked in a cycle that never resolves, and the detectors are consistently a step behind.
If you can’t reliably detect fakes after the fact, the only solution is to prove authenticity from the source.
Introducing Brevis Vera
Brevis Vera takes a very different approach. Rather than analyzing whether media looks real, it lets media prove where it came from and what happened to it.
Vera is an end-to-end attestation system that verifies a published image or video originated from a real-world capture event on a real device, and that every edit applied to it along the way was legitimate and provable.
How It Works
Starting at the Source
Brevis Vera builds on the C2PA (Coalition for Content Provenance and Authenticity) standard, which a growing number of device manufacturers now support. C2PA allows a device to cryptographically sign media at the moment of capture, binding the content to the hardware and producing tamper-evident provenance metadata.
This answers the first question: was this captured by a real camera on a real device?
But this is only the starting point, because raw media is rarely what is ultimately published in the real world.
The Editing Gap
Journalists crop images, creators blur faces, editors redact private information and adjust exposure and color. Subtitles and annotations are added, and eventually everything is compressed for faster mobile loading.
These edits are legitimate and necessary. But the moment you modify a signed image, the original hardware signature no longer applies. Even a simple crop breaks the cryptographic binding between the signed file and the published version. Authenticity and editing are in direct tension, and until now there hasn’t been a way to reconcile them.
ZK Proofs of the Edit Path
This is the core innovation behind Brevis Vera.
Vera integrates with open-source editing libraries and uses Brevis Pico zkVM to generate a zero-knowledge proof of the entire editing process. When an editor modifies media with supported software, Vera takes the original C2PA-signed metadata and raw media as input, executes the transformations, and generates a proof that mathematically attests to three things:
the output derives from the signed original, only permitted transformations were applied, and no hidden or malicious edits were introduced
The proof is generated locally and can be verified independently by anyone, without exposing the raw content or the editorial workflow.
What This Enables
Brevis Vera preserves cryptographic proof of real-world origin through the entire editing process. It maintains the privacy of both the original raw content and the editorial workflow. Verification happens without centralized intermediaries, and the entire system is open-source.
For the first time, published media can carry verifiable proof that it came from reality and was transformed only in legitimate, provable ways.
Live Now
Brevis Vera is live today. The first release integrates with an open-sourced image editing library and supports a wide range of common transformations. We are currently in discussions with popular consumer-facing image and video editing applications to bring Vera directly into widely used creative tools. We are also open-sourcing the Vera reference implementation on GitHub (github.com/brevis-network/brevis-vera).
Want to see it in action? Try our interactive conceptual demo (http://vera.brevis.network/) to experience how Vera works firsthand.
If you’re interested in trying the full version or collaborating on Brevis Vera, reach out through partner form (https://form.typeform.com/to/lduiliob?typeform-source=brevis.network).
About Brevis
Brevis is a verifiable computing platform powered by zero-knowledge proofs, serving as the infinite compute layer for Web3. Applications can offload expensive computations off-chain while proving every result on-chain. The Brevis stack includes Pico zkVM for general-purpose computation, the ZK Data Coprocessor for trustless access to historical blockchain data, Pico Prism for real-time Ethereum block proving (99.6% coverage, 6.9s average), and ProverNet, a decentralized marketplace for ZK proof generation. To date, Brevis has generated hundreds of millions of proofs across 40+ protocols on 6 blockchains.

TL;DR: Matrixdock is partnering with Brevis ($BREV ) to launch an Incentra campaign that adds $1,000 in XAUm rewards on top of existing CAKE emissions for the XAUM/USDT liquidity pool on PancakeSwap (BNB Chain). LPs earn dual rewards, CAKE from PancakeSwap’s farm plus XAUm distributed through Incentra based on fees generated. The campaign launched February 26, 2026 and rewards are verified with zero-knowledge proofs so every distribution is transparent and provable.
Tokenized Gold Meets DeFi Liquidity
Matrixdock is one of Asia’s leading RWA tokenization platforms under Matrixport Group. Their flagship product, XAUm, is a gold-backed token where each unit represents one troy ounce of 99.99% pure LBMA-accredited physical gold, securely vaulted with partners like Brink’s across Singapore and Hong Kong. The token is live across Ethereum, BNB Chain, Solana, Sui and Plume, with over $84M in total asset value and physical redemption available in major Asian wealth centers.
Tokenized gold has seen significant traction as both institutional and retail investors look for on-chain exposure to traditional safe-haven assets. But minting a gold token is only half the equation. For XAUm to function as a productive DeFi asset rather than a static store of value, it needs deep, reliable liquidity on decentralized exchanges.
That’s where this campaign comes in. Matrixdock is using Brevis’s Incentra platform to add an extra layer of rewards for liquidity providers in the XAUM/USDT pool on PancakeSwap, creating a dual-reward structure that makes providing liquidity to tokenized gold meaningfully more attractive.
Campaign Details
Eligible pool: XAUM/USDT on PancakeSwap v3 (BNB Chain, 0.05% fee tier)
Reward pool: $1,000 in XAUm tokens plus CAKE emissions
Reward calculation: Based on fees generated by each LP’s position
Campaign start: February 26, 2026
Duration: 30 Days
Claim chain: BNB Chain
Incentra link: https://incentra.brevis.network/campaign/?pool_id=0x497e224d7008fe47349035ddd98bedb773e1f4c5&type=3&chainId=56
LPs in this pool earn from two sources simultaneously. PancakeSwap’s existing CAKE farm rewards continue as normal, while Incentra distributes additional XAUm rewards proportional to the fees each LP generates. The more actively a position contributes to the pool’s trading volume, the larger its share of the XAUm reward pool.
How It Works
Incentra calculates reward eligibility by reading on-chain LP activity, specifically the fees each position generates over the campaign period. These calculations happen off-chain where they’re computationally cheap, and Brevis generates a zero-knowledge proof verifying that the results are correct and match the actual on-chain data.
This means reward distributions are fully transparent and verifiable. LPs can confirm that their share was calculated fairly without trusting Matrixdock, Brevis, or any third party to report numbers honestly. The proof speaks for itself.
How to Participate
Add liquidity to the 0.05% Fee Tier XAUM/USDT pool on PancakeSwap v3 (BNB Chain)Earn CAKE farm rewards as usualIncentra automatically tracks your LP fees over the campaign periodClaim your XAUm rewards through the Incentra dashboard once the campaign concludes
Gold, Liquidity, and Verifiable Rewards
RWA tokens like XAUm bridge traditional assets into DeFi, but their utility depends on having liquid, active markets. This campaign aligns incentives for both sides: Matrixdock gets deeper liquidity for its flagship gold token on one of BNB Chain’s largest DEXs, and LPs get a compelling reason to provide that liquidity with dual rewards they can actually verify.
For Brevis, this represents another expansion of Incentra into the RWA sector, where transparent and provable reward distribution is especially important given the institutional-grade standards these assets operate under.
About Brevis
Brevis is a verifiable computing platform powered by zero-knowledge proofs, serving as the infinite compute layer for Web3. Applications can offload expensive computations off-chain while proving every result on-chain. The Brevis stack includes Pico zkVM for general-purpose computation, the ZK Data Coprocessor for trustless access to historical blockchain data, Pico Prism for real-time Ethereum block proving (99.6% coverage, 6.9s average), and ProverNet, a decentralized marketplace for ZK proof generation. To date, Brevis has generated hundreds of millions of proofs across 40+ protocols on 6 blockchains.

TL;DR: Brevis ($BREV ) Pico Prism now achieves 99% real-time Ethereum proving with just 16 RTX 5090 GPUs across two machines, down from 64 GPUs in our original release. Average proving time holds at 6.91 seconds. GPU costs drop from $128K to $32K, with total hardware setup estimated at ~$100K. The breakthrough comes from a redesigned dual-machine distributed architecture that keeps all GPUs fully utilized through intelligent scheduling and data locality optimizations.
Recap: Where We Started
In October 2025, we announced Pico Prism and its real-time proving results for Ethereum L1 blocks. Using 64 RTX 5090 GPUs across eight servers, Pico Prism proved 99%+ of Ethereum blocks within 12 seconds, with an average proving time of 6.9 seconds and a hardware cost of $128K.
Those results marked a significant step toward the Ethereum Foundation’s real-time proving goals. But the original announcement also included a commitment: achieve comparable results with 16 GPUs in the coming months.
We delivered on that commitment.
The Results
Tested on the same benchmark set of 7,200 Ethereum blocks used in our original Pico Prism announcement, the 16 GPU configuration achieved:
The distribution tells a clear story: out of 7,200 blocks, 7,165 were proven under 12 seconds. The median proving time landed at 6.76 seconds, with 90% of blocks completing under 9.25 seconds.
How We Got Here: Dual–Machine Architecture
The original Pico Prism setup distributed proving across eight servers. That configuration delivered strong results, but the coordination overhead across that many machines left room for architectural improvement.
The new version consolidates everything down to two machines, each equipped with 8 RTX 5090 GPUs, connected by a 100Gbps interconnect. A lightweight scheduler coordinates work between them while all heavy computation stays local.
Two areas of improvement drive the efficiency gains:
Data Locality. The previous architecture spent significant resources moving data between eight machines. The new design keeps data where it’s needed, so machines spend their time proving rather than transferring information back and forth.
GPU Utilization. With eight machines, keeping every GPU busy at all times was a coordination challenge. The new architecture ensures all 16 GPUs are working continuously with no downtime between tasks, squeezing maximum performance out of every piece of hardware.
The result is virtually identical proving performance with 75% fewer GPUs.
Computing Infrastructure
What This Means
When we announced Pico Prism in October, GPU costs alone totaled $128K across 64 RTX 5090s. That figure is now $32K with just 16 GPUs, a 75% reduction. Total hardware cost for the full two-machine setup, including CPUs, memory, and networking, comes in at approximately $100K, right at the Ethereum Foundation’s target for real-time proving infrastructure.
The Ethereum Foundation has now declared the performance race effectively won and shifted its focus toward security foundations for L1 zkEVM integration, setting milestones for 128-bit provable security by end of 2026. With Pico Prism now delivering 99%+ real-time proving at ~$100K total hardware cost, the performance side of that equation is settled. We’re actively working alongside the EF’s security roadmap to ensure Pico Prism meets the soundness requirements for mainnet-grade deployment.
The path from 64 GPUs to 16 also tells a broader story about proving efficiency. Raw GPU count was never the real bottleneck. Architectural decisions around how work gets distributed, how data flows between machines, and how GPUs stay utilized matter as much as the hardware itself. Smarter coordination delivered a 75% reduction in GPU requirements and cost with virtually identical proving performance.
We’re continuing to optimize Pico Prism and will share updated benchmarks as the architecture evolves further.
Our benchmarks remain fully reproducible. The binaries are available at: https://github.com/brevis-network/pico-ethproofs
About Brevis
Brevis is a verifiable computing platform powered by zero-knowledge proofs, serving as the infinite compute layer for Web3. Applications can offload expensive computations off-chain while proving every result on-chain. The Brevis stack includes Pico zkVM for general-purpose computation, the ZK Data Coprocessor for trustless access to historical blockchain data, Pico Prism for real-time Ethereum block proving (99.6% coverage, 6.9s average), and ProverNet, a decentralized marketplace for ZK proof generation. To date, Brevis has generated hundreds of millions of proofs across 40+ protocols on 6 blockchains.