Since ChatGPT was unleashed on the world in November 2022, we’ve become used to summoning leading intelligence from a browser tab and interacting with it there.
asserted
we → unleash → it
More than a billion of us use ChatGPT every single week, asking it questions and acting on its answers.
asserted
billion → use → answers
The next generation of AI agents are starting to be unleashed, clicking around the web on our behalf and buying items — but still doing so from our devices.
asserted
generation → start → devices
However, the next phase of the tech boom may be much harder to keep inside our screens.
uncertain
phase → keep → screens
More than 500 scientists, engineers, founders and investors gather on Thursday, October 8, at 93 Feet East in Shoreditch for Deep Tech London — which brings the technology off our laptops and into the real world.
asserted
which → gather → world
Among the innovations being demonstrated and presented are humanoid robots, rockets, brain-computer interfaces, batteries, photonics and biotech breakthroughs.
asserted
robots → demonstrate → innovations
The term “deep tech” describes companies building tools that can make difficult advances in science or engineering, rather than just another app or software service.
asserted
that → describe → science
“I interpret deep tech as this research and development [R&D]-heavy, science-heavy research side of technology,” says Lachlan Chavasse, founder of the summit.
asserted
Chavasse → interpret → summit
And we Brits happen to be particularly good at that.
asserted
Brits → happen → that
The UK ranks third in the world for deep tech venture capital raised since 2019
Like our prowess in AI, the UK ranks third in the world for deep tech venture capital raised since 2019, behind the US and China, according to a recent analysis by Dealroom and the Royal Academy of Engineering.
uncertain
UK → rank → Engineering
Companies spun out of university research labs account for a third of UK venture-backed deep tech companies founded since 2010.
asserted
Companies → spin → 2010
And when they spin out, those companies tend to congregate in London.
asserted
companies → spin → London
Imperial College’s White City campus is one such hub, housing more than 100 science and technology businesses.
asserted
campus → house → businesses
This year the campus also opened a pilot manufacturing facility for deep tech firms trying to make the leap from lab prototype to something they can produce at scale.
asserted
they → open → scale
But companies are cropping up all over the capital.
asserted
companies → crop → capital
One company that’s presenting at the summit in Shoreditch is Prima Mente, a biotech start-up trying to use foundation models — the kind of AI technology that powers ChatGPT — to better understand neurodegenerative diseases such as Alzheimer’s, and ultimately help prevent the disease in the first place.
asserted
that → present → place
“Understanding biology has always been incredibly difficult,” says Hannah Madan, Prima Mente’s co-founder.
asserted
Madan → understand → biology
The company is using AI to analyse different kinds of biological information, from genomics to proteins, and looking for patterns that might distinguish different forms of dementia.
uncertain
that → use → dementia
It has around five petabytes of data that it’s already rifled through — equivalent to 1.7 million hours of the highest quality streaming video on Netflix.
asserted
it → have → Netflix
While the large language models behind our AI chatbots learn relationships between words, Prima Mente wants its models to do the same for biology.
asserted
models → learn → biology
“We want to use that technology to understand the sequences within biology and how they connect to each other, which we haven’t been able to do in the past,” Madan says.
asserted
Madan → want → past
Prima Mente says one model has achieved 89 per cent accuracy in detecting early-stage Alzheimer’s and Parkinson’s from blood samples — though turning research results into routine medicine is a much longer road.
asserted
turning → say → medicine
Prima Mente is one example of a whole host of deep tech companies starting to make an impact on our world.
asserted
Mente → start → world
South London-based Rivan is building machinery that uses solar power, water and carbon dioxide captured from the air to make synthetic methane, and deployed a one-megawatt system in August.
asserted
that → base → August
Its aim is to make synthetic fuels cheaper than fossil fuels.
asserted
fuels → make → fuels
It’s now working with Wales & West Utilities on what they say will become Britain’s first commercial synthetic natural gas facility connected to the national grid.
asserted
they → work → grid
Chavasse points to Rivan as the sort of deep tech firm that could become a part of our lives surprisingly quickly.
uncertain
that → point → lives
Tech at the speed of light
Deep tech could also, in turn, improve how AI works too.
uncertain
AI → improve → turn
Oxford spinout Lumai is developing processors that use beams of light to perform some of the calculations needed by large AI models, rather than relying solely on conventional silicon electronics.
asserted
that → develop → electronics
It raised more than $10m last year to try to reduce the energy needed for AI use.
asserted
It → raise → use
And Space Forge, a Welsh company trying to manufacture advanced semiconductor materials in orbit, has successfully generated plasma in low-Earth orbit on its ForgeStar-1 satellite.
asserted
Forge → try → satellite
Those involved in the sector recognise the challenges
It may all sound impressive, but those involved in the sector recognise the challenges.
uncertain
those → involve → challenges
Companies can pursue an idea only to realise in testing that it’s not feasible.
asserted
it → pursue → testing
And that risk is one reason why sometimes the sector can struggle to turn ideas into reality.
asserted
sector → struggle → reality
Britain is exceptionally good at generating the science, but weaker at funding companies once they need the cash to keep testing their ideas.
asserted
they → generate → ideas
Domestic investors supplied 57 per cent of early-stage UK deep tech funding between 2023 and 2025, but disappeared at later stages, according to Dealroom.
uncertain
investors → supply → Dealroom
They accounted for only 18 per cent of the money in funding rounds above $100m.
asserted
They → account → 100
Into that breach entered US investors, who supplied 59% at that stage.
asserted
who → enter → stage
For Chavasse, that’s why putting actual machines they’ve been working on in front of people — at something like Deep Tech London — matters.
asserted
they → ’ → London
He wants more people to realise that deep tech is something they can participate in — and see the benefits of.
asserted
they → want → benefits
…and 1 more, not listed.