Proving Ethereum in Real Time with OpenVM 2.0

Transcript

[applause] >> All right, so my name's Yi. I'm one of the founders and CEO of Axiom. Uh today today I'm going to tell you about how zero knowledge proofs are getting ready for production and how they are the key to forward-looking scaling of Ethereum. So first of all, the reason that we're all together here in New York today is that Ethereum is having an institutional moment in 2026. So there are a number of high-stakes use cases for institutions that are taking off.

Payments, tokenization, and trading. And all of these institutions are choosing Ethereum for the reason that Ethereum has the best security, neutrality, and okay scalability to support these financial applications. And so as these use cases grow, it really behooves Ethereum to scale throughput while preserving the core properties of security and neutrality that are actually bringing these use cases to the chain.

And the core challenge here is that if we just naively scale the gas limit, what's going to happen is that validators will have much higher hardware requirements and start impacting the decentralization and therefore the security and neutrality of the chain. And the core issue here is that in the current design of Ethereum and essentially all blockchains, every validator has to re-execute every transaction on the chain. And so the only way around this scaling issue is to address this bottleneck by removing the need for re-execution in chain validation.

Now Ethereum's road map to do this, and this is unique among blockchains, is to prove execution with zero knowledge proofs so that validators only need to verify proofs instead of re-execute all blocks. Now, to dive a bit deeper into how this works, a zero knowledge proof is a piece of cryptography that allows anyone to prove that a program, in this case the state transition function of Ethereum, was executed correctly so that anyone else can verify that proof very cheaply.

What this allows Ethereum consensus to do is have the execution of Ethereum proven once and then verified at constant cost by all validators. And the core thing that this gives Ethereum is the ability to decouple gas limits from the validator hardware requirements. And so, this will allow Ethereum to scale without compromising the decentralization and neutrality that is so important. Okay, so how do we actually achieve this in ZK?

The core innovation in zero knowledge proofs over the last two or three years is the introduction of what's called a ZK virtual machine. This is a piece of technology that abstracts away the cryptographic complexity of deploying zero knowledge proofs and enables developers to verify the execution of arbitrary computer programs. In particular, for Ethereum clients, this allows the existing state transition functions, which have been battle tested um in production for years, to be directly proven in zero knowledge.

And for other applications building on Ethereum, this allows developers to write their application logic in a normal programming language and directly get it verified in zero knowledge. And the consequence is that is that ZKVMs make zero knowledge proofs accessible to developers both of Ethereum and of applications on Ethereum. Now, we're in New York City and so I think the broader trend across zero knowledge in the last couple years is that ZK is now becoming institutional ready.

And so at Axiom we've been developing a ZKVM called OpenVM. And we think that the key characteristics that matter for institutions to rely on this type of cryptography are performance, security, and generality. So for the rest of the talk, I'll walk you through how we're doing on each of those fronts and how we're measuring up against the highest institutional standards. So in terms of performance, one of the biggest milestones for a ZKVM and in particular OpenVM is the ability to keep up with Ethereum itself.

What that means is to prove the Ethereum block execution within the block time of Ethereum, namely in real time. And with OpenVM, I'm happy to say that we're now comfortably within this time frame. What you see here is a histogram of how long it takes OpenVM to verify the execution of a full day of Ethereum blocks. You can see that the proving time ranges from a little over 1 second to a maximum 10 seconds. And the average is 4.7 seconds, which is well within the 12-second Ethereum slot time.

So at OpenVM, we're quite comfortable proving Ethereum. But we think that ZK can do more. In fact, OpenVM's able to prove general-purpose computations and we ran a benchmark of general computer processors. On that mark benchmark, Open VM is able to prove at 964 MHz. And so, what that means for developers is that if you're able to run something on your processor, and if it was good enough to run your application on a GHz processor, it's good enough to be proved with Open VM.

This means that performance is really no longer a bottleneck for deploying zero-knowledge proofs in production. What's more, this performance has been scaling quite rapidly. The beginning of last year, we were over 75 times slower than Open VM is today. And we see a very direct cadence of performance improvements in the coming months. The upshot of this for Ethereum is that with Ethereum's current scaling roadmap, improvements in zero-knowledge proof performance directly translate to improvements in Ethereum throughput.

So, we're very excited at Axiom to start powering increased gas limits for Ethereum. Now, zero-knowledge historically has been dominated by performance concerns. It was just too slow and too expensive. But that's resolved, we can turn to something that's perhaps even more important for institutions, which is security. So, at Axiom, we believe that the ultimate zero-knowledge proof systems that will be used in production must be built on extremely rigorous mathematical and cryptographic foundations.

That's why we've taken a verticalized approach to developing Open VM. Started with a new zero-knowledge proof system called Swirl. Proof system has provable security, meaning that the cryptographic foundations only depend on mathematical theorems and not on any conjectures, uh some of which were actually proven to be false in the last year. In addition, as many people have seen, there's some rapid developments in quantum computing that are threatening certain cryptographic primitives.

We designed Swirl to be post-quantum at its foundation. What that means is we only use the assumption that hash functions are random oracles. And this is the most minimal cryptographic assumption one can make, and is expected to stay robust under quantum computing. And finally, we target a rigorous security level. Open VM has 100 bits of provable security today, and we expect to achieve 128-bit security by the end of the year. Now, on top of this foundation, as men- Oh.

Even if you have good math, um you might still get hacked by AI. And so, we believe that it's necessary for anyone building cryptographic technology to innovate on defensive security in today's environment. We apply a multi-layered approach to actually secure Open VM. Of course, we have traditional internal and external security audits through our partners at Cantina and ZK Security. But we're also starting to innovate on new techniques to secure the system. First, we've developed a continuous AI auditor that runs internally and scans all updates to the system to catch security vulnerabilities before they

even enter our production process. And second, we're starting to formally verify each piece of Open VM in the Lean Theorem Checker. We believe this is the only way to get fundamental programmatic correctness guarantees that cannot be hacked by AI. As everyone is there are various mythos tier attackers coming both for DeFi and the underlying cryptographic primitives. And at OpenVM, we are excited to innovate to defend against them.

So now that we've addressed performance and security as the main features that a ZK system can provide, we started to talk about some of the applications that we see them bringing to Ethereum. Luckily for application developers, the maturity of ZK allows developers to benefit from the same performance characteristics that are used for Ethereum L1. Uh through our partnership with Lyra, we've been working to develop Lyra EVM, which is an EVM equivalent chain that can tap into the tremendous liquidity and markets that are available across Lyra spot and perps.

And we believe this can offer users kind of the best of both worlds between Ethereum L1 and the Lyra exchange. Lyra is a super high-performance ZK verified perpetuals exchange, um and we bring programmability to Lyra through the addition of an Ethereum Virtual Machine environment. Perhaps more traditionally, we're also powering more general-purpose rollups like Scroll, where by integrating with OpenVM, Scroll is able to become a type one ZK rollup and incorporates um EVM upgrades by uh writing them in normal Rust code.

And in both of these use cases, the key feature that we find exciting is that the developers of these projects were able to focus more on their application logic and features and less on just implementing them both performantly and securely in ZK. So, with the emergence of performant and secure ZK systems, we really believe that this year is the year that zero-knowledge proofs will be ready for institutional use cases.

We see the benefits that uh they can bring institutions across security for scaling on-chain applications um that inherit the security of Ethereum to modality like trading, bridging, and roll-ups accommodated on Ethereum L1. Incorporating greater privacy in applications for confidential payments or identity verification. And also for expanding the scope of cryptographic very verification outside of what one might traditionally consider blockchain applications.

And so, our view is that zero-knowledge proofs are finally reaching institutional maturity. And at Axiom, we're excited to help institutions actually deploy them. So, in conclusion, there're two real takeaways I want you to take from this talk. First, zero-knowledge ZK virtual machines are now one of the core drivers behind scaling Ethereum both securely and neutrally. And with OpenVM at Axiom, we're bringing the benefits of ZK value institutional use cases surrounding Ethereum and even outside of blockchains entirely.

So, if you'd like to learn more, you can check out our website or email me directly. Thanks. >> [applause]