London-based startup Tromero has closed a £1.5 million seed round led by BlueYard Capital to launch a decentralised crypto mining platform for training AI models in a cost-effective way.

The startup, launched by AI PhD researchers Sophia Kalanovska and Charles Higgins while on the Conception X deeptech venture programme, aims to intercept the massive compute power behind crypto mining and redirect it to solve useful machine learning problems that benefit more people – across drug discovery, battery optimisation and more.

Tromero’s system is built around a proprietary proof-of-training algorithm that allows miners to compete to solve AI problems provided by organisations in the startup’s network wanting to train AI models. By combining two computationally intensive processes – AI training and crypto mining – into one, called Proof of Useful Work, the startup is able to offer large crypto rewards for miners, as well as payments in fiat money per job completed as additional incentives. 

This approach substantially reduces AI training and hosting costs for organisations, introducing a cheaper alternative to cloud providers such as Amazon’s AWS, Microsoft’s Azure, Google Cloud and most general distributed AI solutions available on the market.

“In an industry that struggles to keep pace with the soaring demand for machine learning training, our new way of mining cryptocurrency ensures that GPUs are no longer wasted on meaningless hashing, like they are currently, but are utilised to solve meaningful problems requiring AI solutions,” said Kalanovska. 

“Our system breaks down some of the barriers that inhibit the creation of impressive AI models, fostering a future where AI is created by everyone for everyone.”

Optalysys secures £21m for quantum encryption chip

Higgins added: “Efficient allocation of resources is rarely a bad idea. Proof of Work mining is trusted by a large community, but it’s hard to argue that it is an efficient allocation of resources. Tromero offers a path by which Proof of Work mining becomes fundamentally useful.

“Web3 has an inherent drive towards and offers a suite of tools for decentralisation, democratisation and privacy – eliminating the reliance upon institutional players for daily transactions.

“As AI becomes an integral part of everyday life, we believe it’s necessary to ensure the AI ecosystem can be just that: decentralised, democratic and private. This offers a clear route by which AI can be cost-effective and easy to use for all.”

The funding will go towards building a team to deliver a functioning beta for a closed set of early-adopter clients and miners in 12 months.

The FCA investigated by National Audit Office