AI is riding high on a curve of hype, which means some of the practical questions that might grow louder over time — the many costs (financial and social), the business models — may not be the most front of mind right now. Today a startup called Fetch.ai is announcing $40 million in funding in anticipation of when that shifts.
The startup, based out of Cambridge, England, says it is building tooling that focuses on “autonomous agents, network infrastructure, and decentralised machine learning” that help enable communication and actions between AI applications, the idea being to make the work produced by them more actionable.
CEO Humayun Sheikh said in an interview that he also believes there is a role for Fetch.ai to play in the creation of learning models, providing a more equitable and traceable approach to AI by way of distributed ledgers for entities to feed data into those models.
Up to now, a large part of the work Fetch has been doing has been in the area of IP and technology development (it already has a number of patent applications and patents issued in the U.S. and Europe).
But there are also some aspects of Fetch.ai’s platform already live — Fetch.ai is built on blockchain technology and it has created a FET token that will be used on its platform. Earlier this week, it launched a “Notiphi” notification feature for using in connection with the Fetch wallet. But these are just parts of what looks like a very ambitious roadmap. Sheikh said the capital it’s announcing today will be invested in that work as Fetch gears up to launch commercial services later this year.
The funding is a substantial amount in the current market, and it is coming from a single company, DWF Labs, an incubator that is connected to an entity called Digital Wave Finance. The latter company is described by DWF Labs as a “top 5 trading entity by volume in cryptocurrency” although, ironically, I have been unable to find any trace of it on the internet and a spokesperson also was unable to provide me with any online links.
DWF Labs meanwhile was in the news last week also for putting $40 million into a startup called Tomi aiming to build a distributed “alternative internet”, and in November 2022, it also backed TON, the blockchain that was originally part of Telegram but resurfaced as a separate project after its original effort collapsed.
The capital behind Fetch.ai may not be the most traditional, but Sheikh insisted to me that his startup is far from being another crypto startup. Using Fetch.ai to build AI models, he believes, is one example of how blockchain can be used for non-financial purposes.
“The ability for people to train models on their own is difficult because you need money,” he said, referring to the divide in the world of AI between the well-capitalized ‘have’s’ and those with less funding who are the ‘have-nots.’ Building a model trained by multiple entities is the solution, in his opinion. “It’s not about decentralization per se but multiple people creating a model with the ownership sitting with multiple stakeholders who trained it.”
An earlier pilot Fetch.ai worked on back in 2020 around parking solutions — it used AI to both help determine free spots in city parking lots, but also to change pricing, and reward people with free public transport tickets when they chose not to drive at all — is another.
There will also be financial elements to this, though: the startup has plans to build services to take results from applications powered by generative AI and to turn them into transactions. For example, asking a chatbot for the best flights from London to Istanbul might in the future not just be a list of flights and prices, but results sorted and personalized to how you prefer to travel. Fetch.ai wants to provide the tools to actually connect those results with the purchase of tickets. And those purchases, Sheikh said, will be made with FET tokens.
Whether companies will have the appetite to be a part of this has yet to be seen, but Fetch.ai’s approach is significant nonetheless: the idea of building beyond basic “wow” applications will have to be explored sooner or later, and explorations in the space is a sign of how hype might ultimately land in the world of real usage.
“Fetch.ai’s platform provides a comprehensive solution for building and deploying peer-to-peer applications with automation and AI capabilities,” said Andrei Grachev, managing partner of DWF Labs, in a statement. “Their technical architecture and decentralised approach to machine learning create a new paradigm for developers and entrepreneurs, and we’re thrilled to support their growth and development.”