The future of the Internet is verifiable.
We are beyond excited to partner with industry giants
@ravi_lsvp
,
@veradittakit
,
@lstephanian
,
@hosseeb
to scale zero-knowledge proofs to unseen levels of compute power.
It's an honor to work with such a dream team
@JensGroth16
1/ We're excited to announce a $25M Series A round co-led by
@lightspeedvp
and
@PanteraCapital
, with participation from
@dragonfly_xyz
, and more, to bring zero-knowledge proofs to the fabric of the Internet.
We're scaling the Nexus 1.0 zkVM to 1 trillion Hz of compute capacity.
Excited to share huge internal alpha at
@NexusLabsHQ
The Nexus Network (v1) is live -- full E2E zkVM distributed proving. Speed is *super super* low, 10Hz global network throughput! 😅😅
Web, mobile and CLI provers are all live.
Supply compute now:
Excited to announce the Nexus 2.0 zkVM!
Substantial engineering improvements from the 1.0:
1. An integration with Jolt from
@a16z
&
@SuccinctJT
2. A new HyperNova prover, invented by
@srinathtv
and
@abhiramko
, built on
@arkworks_rs
3. The new programmatic Nexus SDK in Rust
1/ Today we’re announcing the Nexus 2.0 zkVM – a major leap from last month’s 1.0 release and the next step toward our goal of scaling ZK to 1 trillion Hz of compute capacity.
#zkvm
#zeroknowledge
Nexus was cited here, so replying:
Very true,
@StarkWareLtd
invented precompiles, they’ve been present in the Cairo zkVM for years:
Precompiles & non-uniform computation have been central to
@srinathtv
’s SuperNova and many other teams’ work
4/ And this. Succinct did not invent ZK precompiles
R0 has had them since our initial release; and Valida, which SP1 is based on, was designed to support them from the very beginning as well
But ofc Succinct wants you to think they invented the idea
(From their benches)
Excited to announce the Nexus 1.0 machine!
We are scaling ZK to 1 trillion Hz of verifiable compute.
It’s powerful to see decades of research by many great scientists finally coming together.
cc:
@NexusLabsHQ
@JensGroth16
@alexanderfowler
Wow seeing a lot of people in APAC, specially Indonesia, supplying compute to Nexus /
@NexusLabsHQ
The network has seen participation from at least 20 countries in < 15 hrs
Classical proofs 🤝 Probabilistic proofs
Advancements in formal methods are now guiding zkVM design -- from compilation, to CPU design & ISA extensions, to arithmetization and witness extraction techniques.
Excited to explore the frontier with
@Pi_Squared_Pi2
Excited to partner with
@NexusLabsHQ
! They’re using advanced techniques with their Proof Compression System. We're also leveraging their Memory Checking Mechanism to create mathematical proofs that will bring ANY programming language on-chain!
1/ We're super excited to work with
@PolyhedraZK
to accelerate the Nexus zkVM with Expander, GKR, and more.
We've known the incredible
@PolyhedraZK
research team led by
@I_NicoNiconi
and
@zhenfei_zhang
for years, and couldn’t be more excited to work with them on pushing the
1/ We’re aligning with Aligned Layer.
Together, Nexus and
@alignedlayer
will simplify and optimize zero knowledge proof generation and verification. Simply code in Rust, prove it with Nexus, and verify it on Ethereum.
Had a great convo on Nexus, zkVMs, IVC & formal verification with
@AnnaRRose
and
@JensGroth16
last week mid-SBC while at NYC
A few great takes from
@JensGroth16
on standardizing Groth16 with a Solidity / Rust (and maybe C++) verifier w/
@zkproof
Thanks for hosting!
➡️ Don't miss:
@danielmarinq
and
@jensgroth16
dropping knowledge about zero knowledge on the ZK Podcast.
@AnnaRRose
, thanks for having us and for the great conversation.
Beyond excited to be backing the world class team of
@danielmarinq
@JensGroth16
@alexanderfowler
to bring internet scale to verifiable computation
The future of general-purpose verifiable computation is truly an inspiring one, and proud to be part of the
@NexusLabsHQ
journey🫡
Hey San Francisco community, you're invited to a Happy Hour *happening tomorrow* in our Salesforce Tower office to meet the Nexus team and talk all things zero knowledge. Hope to see you there.
RSVP:
1/ We’re super excited to be supporting the
@StanfordCrypto
event at SBC – August 6 in NYC.
Plus,
@danielmarinq
and
@nexodonb
from our team are part of a really great lineup of speakers.
Excited to see many friends tomorrow at Stanford BASS
DMs open if you want to discuss partnership opportunities w/ the
@NexusLabsHQ
team
We'll be exploring the next 10 years of Ethereum scaling
📢 Happening tomorrow at 11:10 EST:
@danielmarinq
and
@nexodonb
are speaking at the
@StanfordCrypto
event at SBC.
Topic: "The Nexus zkVM: Democratizing Ethereum’s Execution Layer"
Here's our CEO
@danielmarinq
presenting on "The Nexus zkVM: Democratizing Ethereum’s Execution Layer" at the
@StanfordCrypto
SBC event earlier this month – lots of interesting thoughts about the future of our industry:
1/ We
@panteracapital
excitedly co-led the Series A investment in
@NexusLabsHQ
, a breakthrough in ZK cryptography, earlier this year.
Let’s learn about how Nexus is revolutionizing verifiable computation. 🔍🧵
It's a pleasure to partner with
@ravi_lsvp
and the rest of the
@lightspeedvp
team.
@lightspeedvp
is well known in the valley for making bold bets in deep tech and fundamental compute infrastructure
Verifiable computation couldn't be a better fit
Congratulations to
@NexusLabsHQ
on announcing their $25M Series A round, co-led by
@lightspeedvp
and
@PanteraCapital
with participation from Lightspeed’s joint venture,
@FactionVC
. Nexus is pioneering zero-knowledge proofs to advance critical digital infrastructure, optimizing
We are excited to back
@NexusLabsHQ
in their Series A alongside
@lightspeedvp
&
@PanteraCapital
. Nexus enables the next generation of verifiable computing, powered by zero knowledge proofs
Incredible work, from an incredible team. I know no one as dedicated and skilled to bring these ambitious goals to life as
@griffintier
The million appchain future is close, and we need companies like
@HallidayHQ
to make wallet UX feel like a breeze
🚀 Introducing the world's first APPCHAIN WALLET 🌐
The future of the EVM is in a million Appchains—yet asking non-crypto natives to switch networks, bridge, and swap tokens can be a huge barrier to entry and will prevent growth.
Change is here 🧵⤵️
The integration presents a nice pipeline and composition of primitives:
Sumcheck -> GKR -> Expander -> The Nexus zkVM.
Check our blog post for a bit more technical depth:
The GKR prover is powered by the sumcheck protocol, and operates over layered arithmetic circuits.
It was invented by the legendary Shafi Goldwasser in collaboration with Kalai and Rothblum, and later improved by
@SuccinctJT
, Xie, Zhang and others
The Halliday Network, a big move for
@HallidayHQ
in powering app-chain commerce in the rollup-centric scaling roadmap
Projects like this are much needed to expand the real-world capabilities of individual app-chains. Congrats
@griffintier
and team
Introducing the Halliday Commerce Automation Network—bringing commerce to modular chains.
Save thousands of engineering hours and increase your volume by empowering users to spend, manage, and transact with ease.
Blockchain commerce is incomplete.
One-click payments - as we
@0xkatz
@arbitrum
This is so true, Stylus has 2 benefits zk-wise:
(1) You can deploy a ZK verifier written in Rust super easily on Stylus (although gas cost might be high)
(2) Stylus contracts can be proven in a Wasm-compatible zkVM
2-way interop!
1/ We're now T-2 days out from BASS SBC in NYC!! 🌲🚀
As the event approaches we wanted to highlight a few more of our speakers and sponsors.
Today we are excited to highlight our MEV, RWA, and DeFi BASS speakers
@DCbuild3r
Just WebGPU + Wasm a memory efficient prover that is recursion-enabled, like Nova, and consumes < 4 GB of RAM
Trade off is large client side proofs :(
Reducing settlement costs for zkVMs through a middleware agg / composition layer is definitely in the path the whole community needs to reach the Proof Singularity
Very excited about this project
But co-processors are still a) expensive to prove offchain and b) expensive to settle on L1 frequently.
At
@0xHungryCats
, we're building a co-processor settlement stack, which reduces both offchain and onchain costs for compute-rich co-processors, and makes zkVMs practical.
It is awesome for revm to get audited and fuzzed
This will certainly push the Rust / Ethereum community forward many steps
Yet another example of how community-driven audits can benefit the whole ecosystem
To
@theBBFund
it was a no brianer to go all in on
@NexusLabsHQ
@danielmarinq
has been an inspirational founder since we worked with him in the Stanford Blockchain Accelerator. Since then his obvious potential has manifested
Huge shoutout to the
@NexusLabsHQ
team who is working night and day to improve the protocol, zkVM and Network software -- solving insane problems at the frontier
Wished I could share all the things that are coming soon -- but that'll have to wait for another day
Truly exciting advancements from the
@NexusLabsHQ
team coming soon, cc:
@JensGroth16
Parallelizing the modular zkVM stack:
Decoupling memory-checking from CPU execution for IVC-based zkVMs
1/
@JensGroth16
from
@NexusLabsHQ
introduced zkVMs and incrementally verifiable computation (IVC). In IVC-based zkVMs, instruction execution is local, but memory accesses may refer to data from earlier instruction cycles. He also discussed techniques to verify these memory
Introducing trevm:
trevm is a loving extension of
@rakitadragan
's revm, designed to simplify high-level EVM flows
trevm aims to simplify gas-estimation, transaction simulation, and block and bundle construction for Ethereum and rollups
There are few that can match the tenacity and shear brain power of
@danielmarinq
and the team he has put together.
Couldn't be more excited for this project. Soon enough, there will be immortal, provable computations. Congratulations!
Huge congrats to Nexus team on zkVM launch today, and also to
@RiscZero
on their zkVM launch yesterday! 🫡
It's a brave new world of fast, arbitrary verifiable compute... and our research team
@SpaceandTimeDB
is looking forward to integrating with the top zkVMs to bring them
Congrats to the amazing
@nocturne_xyz
team and
@luketchang
in particular!
Privacy is key for institutional adoption, and new primitives like these bring us closer to that future.
1/ For crypto to reach its next stage of growth, it must be used less for speculation and become more like money. We believe privacy is a necessary condition for this to occur.
@nocturne_xyz
we’re excited to introduce a new primitive to the Ethereum ecosystem—private accounts 🧵
@kostascrypto
Contrarian take: special purpose ZK is still exciting, and will get even more exciting, due to the context of zkVM precompiles and special-purpose ISA extensions
Couldn’t be more excited for what
@danielmarinq
@JensGroth16
@alexanderfowler
and team are building at
@NexusLabsHQ
!
All important computations will be verified, not just in crypto but across the entire Internet.
The future of the Internet is verifiable, proved by Nexus!
Verifying proofs on Aligned allows for benefits beyond cheaper proof verification:
(1) Tight-knit development cycles
(3) Iteration with different provers
(2) Architectural ability to be proof-system agnostic
Virtually, Aligned serves as a proxy universal verifier
@DCbuild3r
@srinathtv
@recmo
That’s nice,
@nexodonb
& team are doing something similar with modified Spartan for committed relaxed R1CS + Groth16 so Nova proofs can go on-chain. Both on the server, however.
Running Spartan on the client should be a bit heavy tho, as it’s a full monolithic SNARK
@VitalikButerin
@danfinlay
Sometimes the simplest route is the best one, e.g., binary trees with SHA256.
MPTs and Verkle Trees are hard to implement and understand, and Poseidon, while ZK-friendly, leads to further engineering overhead for non-zk systems, developers and standards.
I think we can bet on
@srinathtv
@DCbuild3r
That’s a neat trick to get client-side zero-knowledge without full zkSNARK compression
We have yet to enable zero-knowledge for HyperNova:
@alex_xiong_
@SuccinctLabs
@NexusLabsHQ
@RiscZero
@a16zcrypto
For the Nexus 1.0 zkVM we deviate from the TinyRAM arch to a RISC-V inspired arch (NVM 1.0) mainly so that (1) instruction decoding is cheaper, and (2) Merkle-tree based memory-checking is cheaper in the context of 256-bit fields and IVC / recursion (see SuperNova paper)
Advancements on the security of the verification layer itself are also possible, from cryptoeconomic (AVSs) to optimistic execution, to full zk-proving the Aligned Layer itself
The
@alignedlayer
team has certainly a lot of room for experimentation, see:
Collaboration mode: ON
We're wrapping up the 2nd Nexus Summit – a busy and productive week of teamwork and progress on all fronts building something big together. Stay tuned...
@alinush407
@Aptos
This is awesome! 1.3M constraints is pretty good as well.
We’re trying to integrate a Circom -> R1CS -> CCS -> Nexus zkVM compiler. Perhaps we can include a Keyless precompile within a larger program by integrating these circuits
@m2magician
@valardragon
Not yet! But we’ve been looking into eliminating registers altogether from the CPU arch as it might be cheaper to consider all register accesses as memory accesses through a global permutation argument technique. Certainly this puts more stress on the compiler
@GuilleAngeris
Using
@nextjs
and SWR from
@vercel
nowadays — fastest load times I’ve seen in the web in years with static rendering for React and caching already built in
@valardragon
@m2magician
We’re actually doing this — doing a compiler pass from RISC-V to NVM to simplify instruction decoding, and specialized pecompile protocols (e.g., matrix multiplication) in conjunction with the precompile system
2/ The Nexus <|> Polyhedra partnership centers around the celebrated GKR prover, built through
@PolyhedraZK
's Expander, and integrating it as a prover module into the Nexus zkVM
GKR offers massive benefits for hyperparallel large-scale verifiable computation -- which is exactly
Excited talk at Stanford among such an epic lineup of speakers.
We will talk about the latest in verifiable computing, cryptography, math and zero-knowledge proofs. Our team at
@NexusLabsHQ
has also prepared a surprise
The folding-scheme story continues with amazing work from {
@abhiramko
,
@srinathtv
}. Now we can reason about IVC schemes and their proofs without thinking explicitly about 2-cycles of elliptic curves, also simplifying the path to implementation
@0xkydo
This seems to have two very different ideas.
1) A bridge may have client diversity to decrease the chance of having implementation bugs.
2) If there was a standard bridging API then clients would be able to cross-check messages.
(2) may be too gas-expensive in practice?