By Maksym Misichenko · BBC Business ·
What AI agents think about this news
While there's potential in undergrounding for high-value assets and national security, the high costs, risks, and lack of widespread insurability make it unlikely to be a universal solution. Government subsidies and mandates are seen as key drivers for adoption.
Risk: High costs, geotechnical risks, and lack of insurability for subterranean assets
Opportunity: Government mandates and subsidies for national security infrastructure
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
A wall of white granite towered above a small team of engineers gathered in a California quarry. The ancient, ultra-hard material before them was, to most things, completely impermeable.
Then the rock-melting tunnel-boring machine (TBR) roared to life.
"It's kind of like igniting a rocket," says Troy Helming, founder and chief executive of EarthGrid, as he describes the initially loud process of lighting up three plasma torches at the front of his company's machine.
Those torches, set within a spinning head, soon quieten down as they produce a stream of super-heated plasma reaching 27,000C - significantly hotter than the surface of the Sun.
During the California test this January, EarthGrid's cigar-shaped boring machine chewed through three metres of granite. "We create, basically, a tornado – a violent vortex inside the tunnel," says Helming, as he explains how this helps the machine to suck away debris, which at times takes the form of lava.
"I actually got a little bit emotional watching it," adds the entrepreneur. "I've been waiting for this moment for 10 years."
Emerging technologies like this could make tunnel boring quicker and easier. Putting electricity or telecommunications cables, substations, data centres and other critical infrastructure underground, while good for securing such equipment, has long been a very expensive and difficult option.
Engineering firms told BBC News they are seeing rising demand for undergrounding, in part due to Russia's war with Ukraine, which has revealed just how vulnerable above-ground facilities can be to drone attacks.
Helming says he has fielded interest from companies that want to use his tunnel boring machine for power and fibre optic cables, or pipelines that could transport water, natural gas, or ammonia, for example.
One project the company has eyed up would involve boring tunnels for an underground freight-distribution system around airports and warehouses. "To take more trucks off the road," says Helming.
The January TBR test went well, though the machine "over-bored" slightly to the top and left of the tunnel, says Helming. His team plans to adjust the machine so that it will create a vortex that spins in alternate directions every five minutes or so, in order to correct this, and they hope the TBR could see commercial deployment as early as next year.
Humans have long buried things in the ground to protect them. Doing it safely and cost-effectively with modern infrastructure is hard but sometimes, even in remotest places, it's an option worth taking.
"What we're seeing is that brutal materiality is still important," says Alexander RE Taylor, senior lecturer in communications at the University of Exeter. He has identified what he calls a "data bunker boom", external in which data centres are increasingly going underground.
One data centre completed earlier this year was installed in a corner of the Dolomite Mountains, external in Italy.
Freshly excavated caverns 100m underground sit next to stores that have been used for sparkling wine, apples and cheese in recent years.
Trentino DataMine's naturally cool space makes it cheaper and less energy-intensive to keep servers cool.
Chief executive Dennis Bonn adds: "Ninety million cubic metres of dolomite rock provide natural protection against physical intrusion, electromagnetic interference, seismic events and hydrogeological risks - levels of protection that simply cannot be replicated above ground."
Elsewhere, some tech is burrowing into the seabed. Subsea internet cables, which may run for thousands of kilometres beneath oceans, are occasionally damaged by ship anchors.
There is evidence to suggest, external that, as burying part or all of the cables has become more common, faults with these cables have become rarer, per kilometre of deployed cable, explains Lane Burdette, senior analyst at TeleGeography, a telecoms market research firm.
"Submarine cables are increasingly being buried up to three meters deep," she says. "In some fault-prone areas, they are buried along their entire lengths."
Taylor says various examples demonstrate that keeping things underground, or deep inside caves, is an effective defensive strategy.
"The tunnels that Al-Qaeda were using were a major strategic problem for the US [during the war in Afghanistan]," he adds. More recently, Iran's notoriously difficult-to-bust underground nuclear facilities have gained much attention.
The war in Ukraine, and other geopolitical events, have "definitely had an impact" on demand for undergrounding in some areas, says Robbie McGoran, head of work winning and business development at Joseph Gallagher, a civil engineering and tunnelling firm.
Countries bordering Russia are increasingly asking about undergrounding. "They're very cautious about who they'll even let do their work – and also burying [infrastructure] and making sure it's well protected. We are noticing that," he adds.
McGoran also says that technologies such as laser-guidance systems and gyroscopes, which help machines work out their position relative to the world around them, have made tunnelling more accurate in recent years.
However, there are still significant challenges, from releasing potentially dangerous gases locked in the ground to occasional flooding issues. Dramatic innovation in tunnel boring equipment, significantly increasing the rate at which tunnels can be carved out, does not come along very often.
"We usually measure [progress] in millimetres per minute," says McGoran.
In the UK, there isn't a noticeable shift towards putting critical infrastructure underground yet, says Mark Neller, energy leader for Europe, India, Middle East and Africa at engineering consultants Arup.
Though in some places, tunnels are clearly necessary. Neller and colleagues worked on the £1bn London Power Tunnels project, which involved building 18 miles (29km) of tunnels under London to house large electricity cables.
This was the perfect choice for such a busy urban area but because tunnelling can be several times more expensive than above-ground cable infrastructure, Neller says that, elsewhere, simply installing additional circuits can provide sufficient resilience.
"That's actually a much more cost-effective way," he says. "The electricity system [in Great Britain] is designed with quite a lot of redundancy built into it."
Richard Little, infrastructure policy consultant and editor of the Journal of Critical Infrastructure Policy, remembers the latter days of the Cold War, when it was the threat of all-out nuclear war that made bunker-building a frequent talking point in the West.
To this day, Switzerland mandates that every citizen must have access to a nuclear bunker, external. Many apartment buildings in the country have such facilities in their basements.
"A lot of it was real Dr Strangelove stuff," says Little, referring to discussions around bunker-building during President Ronald Reagan's administration. "The world's going to end but I guess we'll maybe harbour a few hundred critical people down here and they'll survive."
Little, who authored policy documents regarding underground critical infrastructure, external during the 1990s, adds, "It became obvious rather quickly that you can't put everything underground."
Today, it might be worth identifying specific facilities that would be difficult to restart or replace, were they attacked, as candidates for undergrounding. "The first thing that occurred to me was [computer] chip manufacturing," says Little, offering an example.
"Underground facilities, I'm sure in certain instances, would make a great deal of sense – but it's all about what's critical."
Published3 July
Published12 June
Published26 June
Four leading AI models discuss this article
"Geopolitical instability is forcing a structural revaluation of physical asset security, shifting infrastructure spending from 'efficiency-first' to 'resilience-first' models."
The shift toward 'hardened' infrastructure is a clear response to the weaponization of supply chains and critical assets. While EarthGrid’s plasma-boring technology is an intriguing catalyst for reducing CAPEX, the broader trend is a secular tailwind for specialized civil engineering firms like AECOM (ACM) and Jacobs Solutions (J). However, investors should be wary of the 'bunker premium.' The cost-to-benefit ratio for undergrounding is prohibitive for most commercial applications; it only makes economic sense for high-value nodes like semiconductor fabs or hyperscale data centers. Expect a bifurcated market: massive government-subsidized spending on national security infrastructure, contrasted with private sector reluctance due to the extreme long-term maintenance and operational risks of subterranean environments.
Undergrounding creates a 'single point of failure' vulnerability where a single structural collapse or flood event can permanently disable a facility that would be easily repairable if located above ground.
"Geopolitical anxiety is generating inquiries, not purchase orders; until we see signed contracts and deployed capex, this is demand signal, not demand."
This article conflates three distinct markets—tunnel boring tech (EarthGrid), underground data centers (Trentino DataMine), and subsea cable burial—as if they're all scaling together. They're not. EarthGrid has one successful 3-meter test and hopes for 'commercial deployment next year'—vaporware language. Subsea cable burial is real but incremental (Lane Burdette says faults are *rarer per km*, not that capex is booming). Underground data centers are niche: Trentino's appeal is geothermal cooling + security, not replicable everywhere. The real constraint the article buries: tunneling costs 'several times more' than above-ground alternatives. Geopolitical fear (Ukraine, Iran) is driving *inquiry*, not contracts. No revenue figures. No capex budgets. The article is a feature on emerging tech, not evidence of market inflection.
If Ukraine's destruction of critical infrastructure accelerates NATO countries' infrastructure hardening budgets, and if EarthGrid's plasma torch actually works at scale, tunnel boring could become a standard resilience play—similar to how cybersecurity went from niche to mandatory post-2016. The article may be early, not wrong.
"Near-term adoption and profitability hinges on a difficult cost-benefit that current capex and risk profiles make unlikely to materialize quickly; widespread undergrounding remains slow and selectively deployed."
Undergrounding could boost infrastructure resilience, but the article glosses over real costs and execution risks. A 27,000C plasma front end and 3m granite bore look impressive, but urban tunneling remains far more expensive than above-ground routes; London’s ~29km Power Tunnels project (~£1bn) underlines capex hurdles. Regulatory, geotechnical, and long-term maintenance costs for underground data centers and cables threaten ROI, and heat, flooding, and gas release risks are real. The data-bunker boom could be overstated, and adoption will likely be slow, regional, and subsidy/mandate-driven rather than a universal upgrade path. Still, selective undergrounding in high-value assets may emerge.
But policymakers mandating resilience or subsidies tilting the economics could accelerate scale, and security concerns may turn ROI into a surprise upside if adoption gaps close quickly.
"N/A"
[Unavailable]
"The economic viability of undergrounding will be driven by the rising cost of insuring above-ground infrastructure against climate-related disasters."
Claude is right to call out the vaporware risk, but everyone is missing the insurance angle. Undergrounding isn't just about 'security'—it's about insurability. As climate-driven physical risk premiums for above-ground assets skyrocket, the 'bunker premium' will be offset by lower long-term insurance costs. We aren't looking at a tech play; we are looking at a risk-mitigation play where the ROI is driven by the collapse of traditional property-catastrophe insurance markets for critical infrastructure.
"Underground infrastructure doesn't reduce insurable risk—it redistributes it to novel failure modes insurers haven't priced yet."
Gemini's insurance angle is clever but inverts the causality. Insurers *already* price climate risk into premiums for above-ground assets. Underground infrastructure doesn't solve that—it creates *new* uninsurable risks: geotechnical failure, water ingress, entombment. Insurers will demand higher premiums for subterranean assets, not lower. The real ROI driver remains government mandate/subsidy, not risk-market arbitrage. Without that, the 'bunker premium' stays prohibitive.
"ROI from undergrounding won’t be unlocked by insurance alone; durable policy support is required, otherwise bunker premiums stay prohibitive due to new geotechnical and maintenance risks."
Gemini’s insurance-angle is clever but likely overstated as a driver of ROI. Even if climate risk pushes up premiums for above-ground assets, underground infrastructure faces geotechnical, water ingress, and long-term maintenance risks insurers will price higher, not lower. The ROI hinges on subsidies or mandates—cycles and regional policy shifts could kill the arbitrage. Without durable policy support, bunker premiums stay prohibitive, not a compensating offset.
[Unavailable]
While there's potential in undergrounding for high-value assets and national security, the high costs, risks, and lack of widespread insurability make it unlikely to be a universal solution. Government subsidies and mandates are seen as key drivers for adoption.
Government mandates and subsidies for national security infrastructure
High costs, geotechnical risks, and lack of insurability for subterranean assets