DeepMind is a wise old sage compared to Europe’s current fresh-faced AI tech darlings like Mistral, Aleph Alpha and H.
Founded in 2010 and staffed by top UK researchers, the company has been behind some of the world’s most ground-breaking developments in AI and exited to Google for $400m in 2014 — at the time one of Europe’s largest tech acquisitions.
That kind of pedigree has bred a crop of employees-turned-entrepreneurs; Sifted has tracked down 12 DeepMind alumni who’ve gone on to launch their own startups in Europe.
The London-based company is not the only AI startup churning out founder talent in Europe; Sifted also tracked down four OpenAI alumni who’ve gone on to set up their own startups in the region too.
At companies launched by former DeepMind folk, almost all — unsurprisingly — are in the AI space, and there are some very recognisable names on the list.
We’ve not included non-profit companies or startups founded by people who worked at DeepMind as interns. If we’ve missed any DeepMind alumni who’ve gone on to become founders in Europe, please get in touch at kai@sifted.eu.
Mistral
Founded: 2023
Funding: $1.1bn, according to Dealroom
HQ: Paris
Mistral has been turning heads on Europe’s tech scene since it emerged from stealth in June last year and announced a €105m seed round led by US VC Lightspeed just four weeks after launching. But just months before that round its CEO Arthur Mensch was working as a research scientist at DeepMind, according to his LinkedIn.
The startup is building foundational models and went on to raise a €385m Series A at the end of 2023 — led by Lightspeed alongside new investor Andreessen Horowitz. It then topped that off with a €600m round of equity and debt in June this year.
H
Founded: 2024
Funding: $200m
HQ: Paris
H is building foundational models and hit the headlines in May this year when it announced it had raised a $220m seed round — made of equity and debt, according to reports. The startup’s founding quartet — Karl Tuyls, Laurent Sifre, Julien Perolat and Daan Wierstra — are all former scientists at DeepMind.
The company’s seed round featured US VC Accel, with participation from US tech behemoth Amazon and New York-based robotic automation software company UiPath. The fresh funds will be used to expand the startup’s research team and to fund the “vast amounts” of compute and data power that are necessary to build foundational models, the company said in May.
EquiLibre
Founded: 2022
Funding: $17m
HQ: Prague
EquiLibre is applying machine learning and reinforcement learning methods to algorithmic trading — which uses computer programs to automatically complete trades in financial markets. It was founded by ex-DeepMind research scientists Martin Schmid, Matej Moravcik and Rudolf Kadlec.
It has one partnership with US-based algorithmic trading-focused financial services firm Tower Research Capital, and is looking to scale its technology to traditional financial markets, says Schmid.
Glyphic
Founded: 2022
Funding: $5.5m
HQ: London
Glyphic is building an AI copilot for sales teams, which it says can listen in to customer conversations and give tactical and strategic insights on how to close more deals. It was founded by former DeepMind research engineers Devang Agrawal and Adam Liska.
Iconic AI
Founded: 2023
Funding: $4m
HQ: London
Founded by former DeepMind software engineer Piotr Trochim, Iconic AI is a video games studio, with a focus on deeply immersive, narrative-driven PC and console games. The startup is currently working on its first game.
Human Native AI
Founded: 2024
Funding: £2.8m
HQ: London
Human Native AI is a data marketplace which provides AI developers with a catalogue of licensed content to train their models. It was founded by James Smith, who was a product manager at DeepMind.
The startup tells Sifted its product is currently in beta and is undergoing testing with AI labs and publishers.
Inductiva.AI
Founded: 2021
Funding: €2.3m, according to Dealroom
HQ: Porto
Inductivia is building an AI-driven platform to help scientists, engineers and businesses to simulate, predict and optimise physical processes across industries. It was founded by Hugo Penedones, former research scientist at DeepMind.
Finster AI
Founded: 2023
Funding: $1.8m
HQ: London
Founded by Siddhant Jayakumar, who spent seven years at DeepMind working across research and engineering before leaving last year to found Finster — an AI-powered platform to speed up investment banking research and workflows.
Jayakumar tells Sifted Finster is working with early customers across hedge funds, asset management and investment banking and has a team of 10. It raised a $1.8m pre-seed round in December 2023 from investors including Hoxton Ventures.
Adia Thermal
Founded: 2023
Funding: £1m+
HQ: London
Adia Thermal is developing a smart installer and controller for heat pumps in the home, which provides room-by-room heating control and detailed insights into the home’s energy performance. It was cofounded by Sam Duke, who was a developer at DeepMind.
The startup is taking pre-orders for autumn this year and says its first manufactured units are coming off the production line now. Adia Thermal is currently raising its seed round.
Elefant AI
Founded: 2024
Funding: Undisclosed
HQ: US and UK
Elefant AI is building AI agents for the gaming world, and its first product is an agent which can play Minecraft. It was founded by Jonathan Hunt, who was previously a research scientist at DeepMind.
The startup has raised a round — which it’s not announced yet — from investors including Adverb Ventures, RTP Global and Squarepeg.
Reliant AI
Founded: 2023
Funding: Undisclosed
HQ: Berlin and Montreal
Reliant AI is an AI research software for business insights. It was founded by former DeepMind research scientist Karl Moritz Hermann.
Variable State
Founded: 2014
Funding: Undisclosed
HQ: London
Variable State is a game development studio that has produced two games, award-winning Virginia and Last Stop. It was founded by former DeepMind senior designer Jonathan Burroughs.
Read the orginal article: https://sifted.eu/articles/deepmind-founder-factory/
