Interview Supra Co-founder: From Oracle to Vertically Integrated Multi-VM L1, Coinbase Invested

In this podcast, Joshua Tobkin, co-founder and CEO of Supra, discusses the project’s philosophy and technological breakthroughs. He explains how Supra integrates decentralized services like oracles, cross-chain protocols, and automation to create a high-performance Layer1 network. Joshua also shares recent updates, including the Supra dApp Competition, StarKey wallet, and Project Blast Off V2, and talks about Supra’s unique position and future direction in blockchain integration.

In terms of financing, Supra is backed by many VCs, including Animoca Brands, BCW, Coinbase Ventures, FiveT Fintech (formerly Avaloq Ventures), Galaxy Interactive, Hashed, HashKey, Huobi Capital, MARBLEX, No Limit Holdings, Prosus Ventures, Razer.com, Republic, Shima Capital, Signum Capital, SMO Capital, Sound Ventures, Sublime Ventures, UOB Capital Management (UOB), and Valor Equity Partners.

YouTube: https://youtu.be/yDbsF1QM5xA

Can you please introduce yourself and explain how Supra works?

My name is Joshua. I’m a co-founder and CEO here at Supra, formerly known as Supra Oracles. I’ve been involved in software development since I was 22, and even before that, at 12 years old, I was trying to figure out how websites work. Now at 38, I’ve spent a significant amount of time in online development and building products and applications. I got into blockchain in 2017 while working in San Francisco. At that time, I was involved in SaaS, and I had a product that still has paying users today, although I no longer focus on it and have others running it. Through this experience, I learned a lot about product-driven development.

So, what makes Supra different from other Layer1s? And why do we need another Layer1? We have a few major innovations. The first is our consensus algorithm called Moonshot. Moonshot is a type of randomized single-leader rotating BFT algorithm, part of the classical consensus algorithm family, and is considered state-of-the-art in terms of throughput and latency.

In addition, we’ve developed a directed acyclic graph (DAG) called Selfish. Theoretically, this DAG could outperform even the latest developments from Aptos and Sui. Although it’s still in the early stages and we are collecting data, the potential is very promising.

What sets Supra apart isn’t just being a Layer1. We have protocols for every step of the consensus process, from data transmission, data dissemination, transaction ordering, to consensus and even parallel execution strategies, ensuring end-to-end low latency. Furthermore, we have built-in native Oracle price feeds. Our Oracle protocol, Dora, has passed a double-blind peer review process and offers block-by-block price feed data, greatly enhancing the developer experience. By integrating Oracle price feeds directly into the blockchain, we can mitigate issues like Oracle extractable value and front-running.

Additionally, we offer on-chain randomness through VRF, decentralized automation akin to “if this, then that,” and native cross-chain communication protocols. In summary, we are fully vertically integrating multiple services into a smart contract platform, enabling applications to do things that are literally impossible on other networks.

Can you explain the reasons behind the shift to Layer1 and the current narrative of Supra?

The research behind Supra began in late 2017 when I was in San Francisco. The initial focus was on scaling blockchain throughput and performance, and even back then, I had the intuition that blockchains should include Oracles to enhance their capabilities. During the bear market from 2018 to 2020, my co-founder John Jones and I established a crypto lab called Unity Chain. As DeFi Summer emerged, we realized that our insights into Oracles and Layer1 could be utilized to provide a next-generation Oracle service.

The reality is, we always had a blockchain and Layer1 in the background, primarily using it for our Oracle protocol. We decided to enter the market focusing on the Oracle problem first because there was a lot of noise around Layer2 solutions, yet there were only a few viable Oracles at that time. We believed we could tackle the Oracle problem with our superior consensus algorithm and protocol.

Thus, the Layer1 aspect has always been part of the plan; it was not an afterthought or pivot. The vision has always included vertical integration, Layer1, multi-VM support, native Oracles, and native VRF. If you were to speak to some of our seed investors, like Coinbase Ventures, Hashed, and Animoca, they would confirm that we initially marketed ourselves as an Oracle-focused project, with the plan to eventually transition to a fully integrated Layer1 solution. Now, we are fully focused on this vision, encompassing both Oracles and Layer1.

What are the key characteristics of Supra Layer1, and how does it compare to other high performers like Solana and TON?

In terms of performance, we believe our algorithms will outperform both Solana and TON. Unlike Solana, which doesn’t use a sharded architecture, and TON, Supra offers execution sharding. Our nodes are randomly sampled into subcommittees called clans, within a network referred to as the tribe. These clans can execute various virtual machines, allowing us to run EVM, Move VM, Solana VM, and eventually, Kazan and WASM. This makes Supra a multi-VM platform.

In our last test, conducted across 125 nodes in five global regions, using our Moonshot consensus with data availability protocol, we achieved a throughput of 530,000 transactions per second. The latency was 500 milliseconds for optimistic finality and 1.5 to 2 seconds for full finality, including transaction ordering, i.e., consensus.

Additionally, our protocol stack offers full vertical integration with native price feeds, cross-chain communication, on-chain randomness, and automation. These features provide capabilities that other blockchains lack. It’s almost like combining the functionality of Solana or TON with Chainlink, but we believe we have the advantage in nearly every protocol.

How do the various components of Supra work together synergistically?

In the Layer2 world, you might have separate layers for execution, data availability, and settlement, and different third-party networks for Oracles, randomness, bridges, and automation. This fragmentation means you don’t have shared security, increased costs due to multiple networks charging fees, and higher latency because of the need to coordinate across different networks.

With Supra, we’ve realized that single nodes in a blockchain can and should perform multiple tasks, limited only by their bandwidth and compute capabilities. As computing power and bandwidth improve over time, we can leverage this to enable our nodes to do more. In our network, nodes are randomly sampled and assigned different tasks. While all nodes participate in consensus, specific subcommittees or “clans” can handle Oracle services, VRF, execution sharding with different virtual machines, etc.

This setup provides a shared security environment, a unified token for all services, a better developer experience, lower costs, and reduced latency. Essentially, it’s the same set of nodes shuffled every epoch to perform different tasks, creating a decentralized microservice architecture that can efficiently handle multiple services.

Let’s talk about the new consensus protocol, Moonshot. What is it, and how does it differ from other Byzantine fault tolerance mechanisms?

Supra’s Moonshot consensus algorithm is a Byzantine fault-tolerant (BFT) algorithm, belonging to the classical consensus algorithm family. In this system, a leader proposes a block, and the rest of the protocol follows. We’ve passed the Microsoft IV verifier, ensuring that our system and design are formally verified to prevent forks, thus guaranteeing safety.

What makes Moonshot unique is its efficiency in block production. Traditional algorithms, rooted in the classic BFT paper by Miguel Castro, require three rounds of communication before proposing the next block. However, Moonshot allows for a block to be proposed in every communication round, rather than one block every three rounds. This significantly increases throughput and reduces latency.

We’ve achieved this through advanced techniques like pipelining and decoupling voting from block dissemination. While large blocks take time to propagate through the network, votes on these blocks are much smaller and can be quickly disseminated. This enables us to propose and vote on blocks almost asynchronously, ensuring both safety and liveness at high speeds.

In summary, Moonshot represents the state-of-the-art for its class of consensus algorithms, with its single leader rotating randomly. It has recently been accepted into a top academic venue through a double-blind peer review process, further validating its position as a cutting-edge solution. We believe Moonshot offers the best performance achievable for this type of BFT algorithm.

Could you introduce the cross-chain communication protocol, HyperNova, and the cross-chain bridge protocol, HyperLoop?

We have two different protocols for cross-chain communication and bridges, HyperNova and HyperLoop, because different networks have varied requirements. Whether bridging to a Layer1, Layer2, Layer3, subnet, parachain, or Bitcoin, each setup requires a unique approach, as they might operate on proof of work or proof of stake mechanisms.

HyperNova is a cryptographically secure protocol that’s simple yet powerful. It leverages the fact that validators on proof of stake networks must register their identities on the blockchain. By knowing the public addresses of these validator nodes, we can aggregate their signatures and verify consensus on our side. This allows our nodes to check if a block from another chain has achieved supermajority agreement by verifying the aggregate signature, using BLS signatures similar to those used by Ethereum, Aptos, and Sui. This method enables us to rerun their consensus and verify it, ensuring the block’s authenticity.

HyperLoop, on the other hand, is suitable for networks like Layer2s and potentially Bitcoin, which may not have an active validator set. HyperLoop involves full nodes of both the interacting chains and Supra. It introduces smart contracts on both the origin and destination chains, implementing a sliding window rate limit. This rate limit restricts the maximum value that can be transmitted through the bridge to 51% of the bridge nodes’ stake within a specified time frame, ensuring game-theoretic security. For example, if 10 nodes have staked $1 million each, the maximum value handled in one sliding window is $5 million. This design prevents colluding nodes from gaining disproportionately by limiting the damage they can inflict.

Additionally, HyperLoop includes a mechanism with “whistleblower nodes” that can pause the bridge if they detect suspicious activity. These nodes can trigger an alert by staking $100,000, and if they are correct, they are rewarded with 50% of the colluding nodes’ stake. If they are incorrect, they lose their stake. This mechanism, along with smart contract checks and sliding window limits, ensures the bridge’s security and efficiency.

Overall, these protocols offer multiple layers of security, maintaining low latency and high value transactions while preventing malicious behavior. Our design and detailed analysis prove that cheating in this system is not profitable. Further details are available in our technical papers, with a new version coming soon.

What’s the next step for Supra, and what role will the $SUPRA token play in the ecosystem?

We have 512,000 KYC-verified participants in our Project Blast Off (PBO) Airdrop “learn and earn” program, which has been running for over a year. This initiative educates our community through gamified missions, ensuring that each participant is a verified individual, not a bot, with a cap of 30,000 to 50,000 participants per country.

Looking ahead, we’re launching Project Blast Off Version 2, which will focus not just on Supra’s technology but also on applications deploying on Supra. This platform, with half a million verified users, allows us to offer a robust marketing campaign for new applications, potentially featuring airdrops. Our goal is to launch a new application every week, offering specific missions for our audience, creating a virtuous cycle of growth and engagement.

The $SUPRA token serves multiple purposes in the ecosystem. It is used to pay for economic security (staking), gas fees, and access to various services like Oracle, cross-chain communication, and automation. Although we aim to minimize costs, including offering Oracle and VRF services for free initially, any fees activated in the future will be payable in $SUPRA tokens. This token unifies payments for all services, including price feeds, cross-chain communication, and verifiable randomness.

Currently, we service 80 different ecosystems without charging them. Even when competitors introduce fees, we plan to keep our services free for some time, supported by block rewards from our Layer1. In summary, the $SUPRA token will have extensive utility across various compute services, not just Layer1 consensus, making it one of the most versatile tokens in the ecosystem.

What projects and dApps are currently being built around Supra?

We have over 155 Memorandums of Understanding (MOUs) with various projects from different ecosystems interested in participating on Supra. Many of these projects are particularly eager to join our upcoming Supra dApp competition. A Supra dApp is an application deployed on Supra that utilizes multiple virtual machines, such as EVM and Move VM, alongside our Oracle price feeds, cross-chain communication, or on-chain randomness capabilities. By leveraging two or more of these services, these dApps can build functionalities that aren’t possible elsewhere.

For example, we’re in talks with several decentralized exchanges (DEXs), including perpetuals DEXs. They’re interested in Supra because we provide Oracle price feed data on a block-by-block basis, ensuring low latency and high throughput. Additionally, we offer an automation network — a decentralized “if this, then that” protocol — that allows node operators to deterministically execute various transactions based on factors like time and asset prices.

This automation network can enable new revenue models for dApps. For instance, in liquidation scenarios, typically, 1% of the liquidation fee goes to whoever triggers it. However, with Supra’s automation network, the process can be automated, allowing for revenue distribution: 70% to the dApp, 20% to node operators, and 10% to a decentralized treasury. This setup allows applications to capture revenue streams that they might not have access to on other blockchains.

Looking ahead, what are your long-term vision and goals for Supra?

Our long-term vision involves a fully vertically integrated blockchain ecosystem that supports multi-VM environments, native Oracle price feeds, cross-chain communication, automation, and on-chain randomness — all under a single network with shared security. We believe this setup enables new use cases and represents the endgame for blockchain technologies.

One of the concepts we’re pioneering is containerization, similar to Layer2 subnets or app chains, but directly on Supra’s Layer1. This approach allows projects to set their own rules and tokens while benefiting from shared security and efficient parallel execution strategies.

Our goal is to continuously improve our infrastructure. For instance, while Moonshot is already highly performant, we’re also exploring Selfish DAG, which may offer even better performance, especially as we scale beyond 200 nodes. We’re committed to testing and adopting the best available technologies, ensuring our network can support a large number of nodes with high decentralization and security.

Ultimately, the job is never done. We will continually study, test, and upgrade our protocols to maintain and enhance our network’s capabilities. This relentless pursuit of excellence ensures that Supra remains at the forefront of blockchain innovation.

What are some of the biggest challenges Supra has faced, and how have you overcome them?

Building a blockchain from scratch presents numerous challenges, especially when it’s not a fork of an existing project. One of the biggest hurdles is dealing with the unknowns, which is why it’s crucial to have protocols peer-reviewed and accepted at top academic venues. This process allows other experts to scrutinize and validate the concepts.

Another significant challenge has been funding. Although Supra has raised around $37 million, which might seem substantial, it pales in comparison to what some DeFi protocols and competitors have secured — often ten times more. This financial pressure has been a double-edged sword; while it has constrained us, it has also driven us to be more creative and resourceful.

We’ve also faced internal challenges, particularly during the bear market and events like the FTX collapse. These periods were emotionally taxing, and we even had disagreements within the team. However, our belief in full vertical integration and the value it can bring to Web3 has kept us going. We’re motivated by the potential to create Super dAPPs that offer functionalities other Layer1s can’t match. This vision, combined with our conviction that we have the best design and architecture, has inspired us to push through the difficulties and continue our journey.

What else do you want to share with us and our audience?

Supra has recently launched the StarKey wallet, which has already garnered over 110,000 downloads. StarKey is an omni-chain wallet, supporting not just Supra but also Ethereum, Solana, Sui, and many others, with plans to add Layer2s and more. The wallet’s standout feature is its innovative key recovery mechanism called Split Key Recovery. This system allows users to encrypt their private key with a password, then split it into pieces stored across various locations, such as Google Cloud, a local device, or even with a trusted guardian. This method enhances security and offers peace of mind, as users can reconstruct their private key only when necessary, reducing the risk of losing funds due to hacks.

The technology behind StarKey is led by Doctor Anaka Kate, a renowned cryptographer known for his work on distributed key generation (DKG). The reception for StarKey has been overwhelmingly positive, with many users switching from MetaMask to take advantage of the key recovery features. Supra anticipates that several hundred thousand users will stake their tokens and expects around 400,000 or more wallet installs. This development is particularly exciting as it broadens Supra’s appeal beyond just its blockchain services, making it an attractive choice for anyone in Web3.

Follow us
Twitter: https://twitter.com/WuBlockchain
Telegram: https://t.me/wublockchainenglish