Interviewee: Daniel Marin, founder and CEO of Nexus
Interviewed and written by: Alex, OpenBuild content team
As AI computing continues to expand and Web3 infrastructure evolves rapidly, how to verify the authenticity of large-scale computing has become a core challenge facing global digital systems. Nexus is pushing forward a cutting-edge exploration - building a verifiable computing network that everyone can participate in and operate globally, integrating the zero-knowledge virtual machine zkVM with a distributed prover network to create a borderless 'World Supercomputer'.
In this interview, OpenBuild talks with Nexus founder Daniel Marin, delving into the founding intention of Nexus, the future path of verifiable computing, and the deep coupling of zkVM, distributed computing, and AI trusted execution. He also shares how Nexus drives high-performance general computing without sacrificing privacy through dual engines of technology and ecology.
Guest Introduction
Daniel Marin, founder and CEO of Nexus, graduated from Stanford University with a degree in Computer Science. He founded Nexus in 2022 with the aim of building a digital world where 'every action is verifiable, every model is auditable, and every interaction is verifiable'. He won a bronze medal in the International Physics Olympiad in 2018 and 2019, and was selected for Forbes' '30 Under 30' list in 2025. Nexus has completed $25 million in funding, with investors including Pantera Capital, Lightspeed, Dragonfly Capital, and others.
Founding Intention: A verifiable computing platform for everyone
OpenBuild: What prompted you to found Nexus? What was the problem you initially wanted to solve?
Daniel Marin: At that time, I was involved in a research project at Stanford's Cryptography Lab and witnessed the rapid evolution of zero-knowledge proofs, especially inspired by Microsoft Research's Nova paper. It made me realize that we could efficiently combine multiple zk proofs to build a new paradigm of 'general verifiable computing'.
On the other hand, I observe that the current blockchain ecosystem faces three major bottlenecks:
1. Users cannot easily contribute computing power
2. Developers need to have a professional background to participate
3. The scalability and performance of the chain itself are limited
I envision whether anyone can use any device (even an old phone) to connect to the network, contribute computing power, run AI-level tasks, and also prove the authenticity of these tasks.
This is the starting point of Nexus - a chain where 'more nodes result in a faster network'. Unlike traditional chains, the Nexus network gets faster and stronger with more participants. Each device that joins is not a burden, but an accelerator. My original intention in founding Nexus is simple - to make the internet safer and more trustworthy.
For many years, I have been thinking about whether it is possible to create a 'general verifiable computer' for this world, an infrastructure where anyone can verify the results of computations. With breakthroughs in zero-knowledge proof technology, this goal has finally become a reality. Especially during my involvement in research at the Stanford Cryptography Lab, I was inspired by Microsoft Research's Nova paper. It made me realize that we could efficiently combine multiple zk proofs to build a new paradigm of 'general verifiable computing'.
In addition, I have also observed multiple limitations in the current blockchain systems:
• Ordinary users are unable to contribute resources;
• High development thresholds, requiring mastery of complex tools;
• Scalability is severely limited.
Thus, I came up with an idea: to allow everyone to connect to the network with their own device (even a smartphone), contribute computing power, and ultimately collectively build a trusted supercomputer that supports large-scale computations.
This is the starting point of Nexus. Unlike traditional blockchains where 'more nodes mean slower', the design logic of Nexus is that more nodes mean a stronger network.
The Future of zkVM and Verifiable Computing
OpenBuild: What was your initial vision for 'verifiable computing'? How do you see its future now? Daniel Marin: We are at a turning point in computing paradigms. I believe verifiable computing will reshape our understanding of computing scalability, performance, and even application forms.
The scalability here is divided into two dimensions:
• Vertical: Speeding up zkVM with hardware acceleration such as GPUs;
• Horizontal: Connecting more devices to form a global computing network, bringing exponential aggregation effects.
And this is not just talk. We expect that within the next year, the operating speed of zkVM will reach 1GHz on an M2 MacBook Pro - meaning a laptop can verify a block every second. This is like the 'Moore's Law of the zk realm' playing out.
At the application level, we are particularly optimistic about the verification needs of AI, such as auditable model behavior and traceable AI-generated content, which is also the original intention of establishing the Nexus Verifiable AI Lab.
The fundamental difference from traditional blockchains
OpenBuild: What is the biggest difference between Nexus and other chains? How does it address the challenges of AI infrastructure?
Daniel Marin: Everyone is talking about scalability, but they all overlook the most critical point: insufficient underlying computing power.
Especially as AI applications grow wider and models become more complex, the required computing power increases exponentially, leading to extremely high-frequency transaction requests. So the question arises: how do we obtain enough computing power to support billions of transactions per second? And also, users must trust these transactions?
This is exactly the solution proposed by Nexus:
• It is a World Supercomputer.
• The network performance improves with each laptop or smartphone connected.
• Any user can easily contribute idle computing power without needing a technical background.
We have built a zkVM-driven Layer 1 blockchain that not only verifies transactions but also verifies arbitrary computations. It aggregates global idle computing power, allowing any user to join with a single click.
Ultimately, Nexus is not a traditional 'chain', but an infrastructure network optimized for general verifiable computing.
The concept of building a 'verifiable internet'
OpenBuild: You have proposed that Nexus is building a 'verifiable internet'. What does this mean? How is it different from the current internet?
Daniel Marin: The 'verifiable internet' we envision is a new paradigm:
• Every action can be proven;
• Every model can be audited;
• Every piece of data can be verified.
From AI to healthcare, from identity systems to on-chain finance, every component of the future digital world will be built on 'verifiability'.
This is not only an upgrade of the trust mechanism but will also bring about a transformation in business paradigms - in this internet, trust requires no 'belief', only 'verification'. With the support of verifiable computing, the digital society will be more transparent, efficient, and secure.
Roadmap and Product Planning
OpenBuild: From both technical and ecological perspectives, what are Nexus's key directions for the next 6-12 months?
Daniel Marin: We have just launched the global Testnet III, which will continue running until the mainnet goes live in the third quarter of this year. 👉 Welcome to visit: app.nexus.xyz to participate in testing and earn rewards.
At the same time, we are focusing on:
• Development of a new version of zkVM;
• Construction of the Nexus Verifiable AI Lab;
• Ecological expansion with over 50 cooperative projects.
In the future, we will continue to iterate, pushing 'verifiable computing' to become a core component of AI and Web3.
Conclusion: Building a Trusted Digital Civilization
From Daniel's sharing, we can see that Nexus is not just trying to 'improve blockchain', but is redefining the trust logic of the future digital society from the perspective of verifiable computing.
In this era of explosive data generation, 'verifiability' is becoming a more critical next-generation infrastructure than 'consensus'.