Cube3.AI Secures $8.2M to Launch AI-Powered Blockchain Security App

Blockchain security firm CUBE3.AI is rolling out its real-time AI-powered blockchain security application after raising $8.2 million in a seed round led by Blockchange Ventures.

The seed round for the launch included participation from Dispersion Capital, Symbolic Capital, Hypersphere Ventures, ICLUB, and TA Ventures. CUBE3 plans to use the funds to further develop its products.

AD

The release includes three distinct products, all of which use machine learning (ML) algorithms in real-time to help secure blockchain protocols. The products are called Detect, Protect, and Manage. Both Detect and Manage are being offered for free.

13 Biggest DeFi Hacks and Heists

Decentralized finance (DeFi) refers to blockchain applications that cut out middlemen from financial products and services like loans, savings, and swaps. While DeFi comes with high rewards, it also carries plenty of risks. Since just about anyone can spin up a DeFi protocol and write some smart contracts, flaws in the code are common. And in DeFi, there are many unscrupulous actors ready and able to exploit those flaws. When that happens, millions of dollars are put on the line, often with no…

Deep DivesLong Reads10 min read

Stephen Graves, Ekin GençApr 18, 2022

The products aim to plug a gap, technically called transaction security, in the industry that has allowed cybercriminals to exploit billions of dollars, including at least $3.8 billion from DeFi protocols, according to Chainalysis.

What is Cube3.AI?

Typically, the smart contract security stack has two tiers: there’s code security and remediation.

Code security focuses on pre-empting threats, employing methods like auditing code, offering bug bounties to white hat hackers to help identify vulnerabilities, and simulating attacks. The remediation layer focuses on steps that can be taken after an attack, including sending out alerts and comprehensive reporting.

CUBE3.AI CEO Einaras Gravrock told Decrypt over email that his team comprises AI, cybersecurity, and cloud engineers, many of whom have been building these types of algorithms for over a decade in Web2 security.

Gravrock also said that the team spent a year building both supervised and unsupervised ML models for Web3 “before we felt they were ready for production.”

In layman’s terms, supervised models use labeled data sets to guide the AI to classify information and predict outcomes accurately. Unsupervised models use unlabelled datasets and instead rely on algorithms to find hidden patterns and insights in the given data.

“We recognize that we are only at the beginning of this journey, and we will continue to improve our existing models, as well as introduce new models as threats evolve,” said Gravrock. “This process allows us to pick up exploits on newly deployed contracts or fresh wallets and stop the transaction before the exploit happens. This is what we are most passionate about.”

Smart contract deployers will now have the ability to block malicious transactions as they occur, meaning their applications can keep on running, even while they actively deal with threats.

CUBE3’S free Detect application taps algorithms to monitor and rate the risk of every smart contract, wallet, and transaction in three risk categories: cyber threats, fraud, and compliance.

Protect—the priced application—monitors everything from the underlying smart contract code layer to the user interface to block entities that exceed a company’s risk threshold in any of the three above categories.

Manage enables protocols to track wallets, view analytics and reports, and issue alerts via supported applications like Discord, Slack, and Telegram.