Tony Lowe is the CEO of Delta.g, a University of Birmingham spinout developing quantum gravity sensors for underground mapping and navigation.
What’s your journey to becoming CEO of Delta.g?
I had a long career in corporate before getting involved in startups about 15 years ago. The last six years have been in deep tech. I was chief operating officer at two Oxford University spinouts. First at Oxbotica (now OXA), where I spent three and a half years overseeing growth from 100 to 325 people and through Series B and C funding rounds totaling $218 million. Then I spent a year at Oxford Quantum Circuits, rolling out quantum computers to colocation data centers globally. About 18 months ago, I was offered the CEO role at Delta.g. My personal passion is the near-term commercialization of quantum technology, and this was the ideal role. We have groundbreaking technology protected by IP that’s actually ready for real-world deployment.
A weekly dispatch featuring exclusive interviews with deep tech founders & a roundup of the most important deep tech news.
What drew you specifically to Delta.g?
Delta.g spun out from the University of Birmingham in 2023 with IP generated from 10 years of end-user focused research in the U.K. government’s Quantum Hub for Sensing, Imaging and Timing, which received over 20 million pounds in public funding. The trigger for spinning out was a 2022 Nature paper, which demonstrated for the first time that quantum gravity gradient sensing could detect a buried tunnel in field conditions, proving that this technology could find underground structures using only the gravity gradient signal.
What attracted me was the team’s mindset. Professor Mike Holinski and co-founders Dr. Andrew Lamb and Johnathan Winch want to see their technology used commercially in the real world. When I joined, they were already on the seventh iteration of the sensor. They’d been working iteratively rather than following the typical academic route of “create an experiment, prove it works, then start over.” Technology isn’t a product. We have to deliver customer value by packaging groundbreaking technology into something that solves real problems.
How would you explain the technology in a pub?
We use a quantum gravity gradiometer, essentially a device that measures how gravity changes between two points. Inside a vacuum tube are two clouds of rubidium atoms moving around at the speed of fighter jets. We use a laser to cool those atoms down to microkelvin levels — colder than space — until they’re practically stationary. Then we drop them and use that same laser ruler to measure how far they’ve fallen.
Here’s what makes us different: Environmental noise drowns out gravity signals. Even waves lapping on U.K. shores create noise many times stronger than the gravity signal we’re measuring. Because we have two identical gravimeters stacked on top of each other, both measurements capture the environmental noise. When we calculate the gradient — the difference between them — we get a pure gravity reading. We’ve created the noise-canceling headphones for subsurface measurements.
What problems does this solve in practice?
Today, if a civil engineering company wants to build something, they need to understand what’s underground like voids, subsidence, changes in ground conditions. Currently, they dig 10 to 20 boreholes in a 20-square-meter area. That takes up to four weeks, and can cost up to half a million pounds, and then they wait another four weeks for data analysis.
We can scan that same area in half a day and provide real-time answers. That’s about 84 times better on a cost-efficiency basis. The global industry tells us gravity data is the best data. The problem is it’s currently slow and expensive. We provide the same data with better fidelity, better insights, in real time on site.
You describe this as building “Google Maps for the underground.” How does that work?
Our quantum sensors turn gravity into what we call spatial intelligence, which is reliable, real-time insights for mapping, navigation and subsurface discovery. The initial model is sensing as a service, but as we reduce size and cost, we’ll move to sensor sales powering a data platform. That platform will use data fusion, machine learning, and AI to constantly improve underground feature identification. Eventually, it becomes externally accessible datasets, not just for people doing the actual measurements. That’s why we call it Google Maps for the underground.
Beyond mapping, you mentioned navigation capabilities. Could this replace GPS?
Yes, it can be hyper-precise. The same sensor that maps underground can tell you where you are. If it can map, it can navigate. There’s real concern globally about the vulnerability of GPS. We’re seeing GPS denial happening regularly in Ukraine and across Europe. For autonomous shipping, our technology could complement inertial systems to maintain navigation accuracy when GPS is denied or degraded. In principle, the same gravity-based navigation methods could enhance submarine positioning by referencing local gravity maps, providing a resilient alternative when satellite signals are unavailable. It’s compelling for defense, civil engineering, energy, mining or anyone who needs to see underground without breaking ground or know exactly where they are.
How much is dual-use capability playing into your roadmap?
We’re in a global situation where bad actors are at play and Europe is rearming. Our technology can provide real-time data for protecting critical infrastructure and borders. We could scan vehicles at full speed, and eventually have drone-mounted and handheld scanners that can detect buried objects, potentially including IEDs.
The technology is as broad as anyone who wants to see what’s beneath their feet or where they are in the world. When we run value proposition exercises with customers, their pain points are always false positives, false negatives, and that gravity data takes too long and costs too much. We’re eliminating those pain points.
How do you define deep tech?
Deep tech for me is technology that is hardware-based but software-enabled. Having completely different sensors for different applications is difficult. The more generic we can make the sensor itself and the more software-configurable we can make it, the better.