Offline🔌

In September 1859, a solar storm known as the Carrington Event pushed electric current through telegraph networks so forcefully that some messages were transmitted with power supplies disconnected, carried by the induced current alone. Equipment sparked, stations caught fire, and operators reported electric shocks, forcing parts of the network to shut down. This occurred in a world that was only lightly electrified, where telegraphy was the main long-distance communication system and the wider economy was largely insulated from such failures. Nothing like the deeply networked, electricity-dependent world of today.

Contemporary accounts describe auroras visible far beyond their usual range, seen as far south as the Caribbean and parts of southern Europe, bright enough in some places to read newspapers at night. Magnetic observatories recorded extreme disturbances, with compasses swinging erratically or becoming temporarily unreliable. Railway signalling and scheduling were disrupted, dependent on telegraph coordination.

In the third decade of the 21st century, a Carrington-scale event would represent a very real threat to the infrastructure that keeps our global economy synchronised. We enjoy satellites (especially GPS), long transmission lines, data centres, and cables that carry all of our cat pics, bank transactions, GPS timing signals, stock trades, payroll files, shipping manifests, software updates, identity checks, medical records, emergency alerts, and the machine-to-machine messages that keep power grids, factories, and markets online and in sync.

In July 2012 a very large coronal mass ejection crossed our orbit. Earth dodged it. Researchers described it as the most powerful in “as much as 150 years” and warned if it had hit recovery would have taken a long time.

Everything everywhere all at once

The internet we spend every waking minute on is not a single network but a network of networks, a global data exchange system. It was born in the late 1960s as ARPANET, a US defence-funded research project. Universities, companies, then governments connected their own networks, linked by routers and fibre-optic cables, coordinated by a small set of technical bodies rather than a single authority. Over time, this research network became critical infrastructure. Email, the world wide web, cloud computing, digital payments and global markets were then layered on top of it. 

‘Switching off’ the internet is difficult because it has no one on–off switch. Decentralised by design, it is thousands of independently operated networks connected by shared technical standards. Traffic can therefore be rerouted when individual links fail. Even when major components break – cut cables, overloaded data centres, fallen satellites – some connectivity remains, making total collapse unlikely. The weakness lies in the modern world’s dependence on it working reliably, at speed, and everywhere at once.

All around the world

Undersea cables are the internet’s backbone. TeleGeography, an industry tracker, notes that as of 2025 there are around 570 cable systems, with more in the pipeline. This plumbing is a major strategic asset – critical national infrastructure – and anxiety over protecting them is a very real concern. European governments and NATO have raised concerns about ‘hybrid warfare’ threats to undersea infrastructure, and reports of a UK-Russia exchange in January 2025 saw Britain accuse Russia of activity around cables (which Russia denied, natch). 

A severe solar storm could create just the strategic cover an adversary could exploit. As networks struggle with satellite outages and power disruptions, a nefarious state actor could exploit the confusion to target undersea cables, where damage is hard to attribute and slow to repair. Cutting a small number of high-capacity Atlantic cables at key choke points would be enough to force traffic onto longer, congested routes, compounding the disruption caused by the flare. Slow internet. No signal. Worst case scenario. 

Cash is king

Payments look simple to consumers. But under the hood they operate continuous authentication, authorisation and fraud controls across multiple networks. When the network is iffy payment systems fall back to ‘offline’ modes built for short outages, or they stop. Either way, risk rises. Merchants may refuse transactions. No more food shopping, no more public transport, no more Netflix.

At that point cash stops being a cost of living budgeting tool and becomes essential for survival. The problem is cash itself is all logistics. When the internet works, banks forecast demand and coordinate deliveries. But when it doesn’t… Even if central banks can print enough notes, they still have to move them securely at scale while fuel supplies and staffing would inevitably be under strain.

Banking stress follows quickly. Modern banks depend on the continuous exchange of settlement messages keeping balance sheets in sync. The wholesale side of banking depends on synchronised communications for settlement and liquidity management.

In Europe, MiFID II’s RTS 25 sets timestamp accuracy requirements on the order of 100 microseconds for certain kinds of electronic and high-frequency activity, synced to UTC. Markets need an agreed sequence of events. Which order arrived first, which trade executed first, which venue reported what and when. A multi-week disruption that degrades GPS timing and network synchronisation severely undermines the basic paper trail of markets.

Trading halt

Exchanges would likely halt trading early in a prolonged outage. Coordinated market require shared information. If market participants cannot see the same price feed or route orders reliably then running an exchange becomes very difficult indeed.

Derivatives would be the immediate casualty. Clearing houses sit at the centre of modern market plumbing. Their job is to measure and mutualise risk. If you cannot mark positions to market, you cannot call margin fairly. If you cannot call margin, you cannot control counterparty risk.

Even if trading continues on reduced connectivity, it would not be business as usual. Liquidity would be thin. Bid-ask spreads would widen. Participants who normally act as market makers would step away because hedging and inventory management become impossible without reliable links. But the most important economic effects would not start on trading floors but in the corporate world.

A cloud on the horizon

Over the past two decades, firms have moved many core functions to cloud services, whether that’s inventory systems, payroll, supplier management, shipping coordination, customer support, even internal communications. In an extended outage, ‘the cloud’ becomes a single point of failure.

Supply chains are particularly exposed. A container without a manifest is just a metal box. A shipment without digital customs documentation is a delay. A factory without reliable scheduling data is idle. The macro result is a sharp drop in output.

The modern economy is service-heavy and services rely on connectivity. Rather than ‘too much money chasing too few goods’, we’d see bottlenecks, delays, and shortages thanks to an inability to coordinate supply with demand. Some prices would jump while others would collapse because buyers and sellers cannot find each other.

What’s the damage?

Lloyd’s of London, working with the Cambridge Centre for Risk Studies on systemic-risk scenarios, published analysis in 2025 suggesting the global economy could be exposed to trillions of dollars of losses over a multi-year period in extreme space weather scenarios (it cites $2.4 trillion over five years as an exposed-loss figure in one scenario set). 

NOAA’s Space Weather Prediction Center notes electric power transmission can be disrupted by space weather through induced currents and associated grid issues. In a severe event, the fear is not just local blackouts. It is damage to high-value equipment, especially large transformers.

Transformers are substantial bits of kit and globally in-demand industrial assets. The US Department of Energy’s resilience work stresses large power transformers have long replacement lead times and this creates serious challenges for grid reliability and resilience. If a Carrington-scale event damaged enough of them across multiple regions a few weeks offline could become months of partial recovery, depending on which parts of the grid were hit and what spares were available.

Have you turned it off and on again?

The world has built its critical systems based on ‘always-on’ international connectivity. As we’ve seen, undersea cables carry the vast majority of intercontinental data traffic. Satellite internet is increasingly useful, but does not come close to the capacity and economics of fibre, yet. Under stress – a solar storm, sabotage, or both – traffic is forced to reroute and becomes unreliable. So, how would the world adapt without the web?

When global connectivity fails, people and firms fall back on what works: face-to-face coordination, local radio, paper, cash. Something like an experiment in deglobalisation. The ‘winners’ are businesses that can take cash, hold stock, and operate with manual processes, communities with robust institutions, and supply chains with redundancy built in.

Much of the digital economy is anchored around trust. Identity verification, authorisation, secure messaging, audit trails. When those systems fail, transactions suffer because counterparties cannot validate each other. Credit tightens because lenders cannot monitor risk cheaply. All the ingredients for a real recession.

After the storm

When connectivity returns, markets would not just pick up where they left off. Expect them to reprice violently thanks to weeks of uncertainty compressed into a highly anticipated reopening window. Liquidity would be tight. Some firms would have gone under in the interim, others quietly negotiated emergency financing in private. That would show up as sudden dispersion in prices, especially in credit.

In the reconstruction phase redundancy would take precedence over everything else. Expect more domestic rails, more offline capability, more diverse timing sources, and more regulatory focus on operational continuity.

Every cloud…

A period offline would reduce some contemporary harms. Fraud, cybercrime, and misinformation campaigns would lose their main line of attack. No more ragebait, attention economy, or 24/7 news cycle. Not to mention the ability of bad state actors to manipulate the free world’s information flows to their advantage. 

A weeks’ long outage would damage near-term revenues for the most internet-native businesses, especially advertising and cloud services. After a real outage, markets would demand a higher risk premium for businesses that depend on continuous connectivity, precise timing, and complex third-party infrastructure. 

The internet feels permanent, embedded as it is in almost every aspect of our lives. But that permanence is an illusion. A Carrington-scale event and an emboldened rogue state actor wouldn’t end the digital world but would remind us modern prosperity is heavily reliant on it. Being offline wouldn’t just be inconvenient. It would expose how much of today’s economy is built on the assumption that everything, everywhere, is always on.

‍

Important Information

‍