A.I. and crypto hybrid Giza raises $3 million to bring machine learning to the blockchain

Founded in October by Cem Dagdelen, Fran Algaba, and Renç Korzay, the firm announced on Tuesday that it had received $3 million in a pre-seed round led by CoinFund. Other participants included StarkWare, TA Ventures, and Arrington Capital, the TechCrunch founder’s venture capital outfit. 

Algaba, Giza’s CEO, declined to comment on his startup’s implied valuation.

Einar Braathen, an investor at CoinFund, said in a statement that Algaba’s firm is on its way “to be one of the first projects that make A.I. available to smart contracts and dramatically expand their design space.”

The investment comes as venture capital funding for crypto startups continues to dwindle. While the A.I. gold rush has stalled amid a larger downturn in VC funding, according to Crunchbase, the almost $58 billion raised by A.I. startups in the second quarter of 2023 far outstrips the approximately $2.3 billion drummed up by crypto firms. It’s no wonder, then, that A.I.-themed cryptocurrencies outperformed Bitcoin through the beginning of the year. (The price gains have cooled, however, as A.I. fever has slightly subsided.)

Algaba comes from the world of A.I., not crypto. He used to be the head of machine learning at Banco Bilbao Vizcaya Argentaria, a Spanish multinational bank, before moving on to a similar role at the sneaker giant Adidas.

In 2019, he started dabbling in crypto, more as a hobby, but recently, as he was leading Web3 events in Madrid and Barcelona, he decided to take the professional plunge. He teamed up with Dagdelen and Korzay after meeting them through the Web3 gatherings he organized.

Blockchains like Ethereum are essentially slow, decentralized computers, and to run computationally intensive A.I. models on top of them is currently unfeasible. “At the moment, there is no clear way to bring A.I. into the blockchain space,” Algaba told Fortune.

That’s why some firms, including Giza, are looking to zero-knowledge proofs, a buzzy cryptographic technique popular in the world of Web3. Firms can run A.I. models on regular, faster computers and prove that the computation happened correctly through a proof, which requires significantly less computational power to verify on a blockchain.

However, for most developers with no knowledge of zero-knowledge proofs, this feat is a lift. That’s where Giza, which currently has 12 employees, comes in. What Algaba and his cofounders are building is a platform for A.I. developers to easily generate zero-knowledge proofs for  A.I. models. 

And with the new injection of capital, Algaba told Fortune, they plan to release the infrastructure behind their platform next week and fully launch their platform by the end of 2023.