Who controls the cable?
How cryptographic trust becomes power over infrastructure.
This blog by Giorgos Georgopoulos first appeared on the Quantum Foresight Substack on Nov 16 2026.
In short: Whichever technology operators migrate first to post-quantum cryptography capture a trust premium, becoming the preferred route and host for sensitive traffic, regardless of who owns the infrastructure they run on. The bodies that set the pace of migration control whose infrastructure counts as trusted, and that control is unevenly spread. Traffic is being collected now and full migration is years away, so early movers gain power over infrastructure owners and sovereigns.
From February 2016, for about six months, traffic from Canada to Korean government websites took a detour through China because of a routing change. It was called a perfect scenario for long term espionage, because the victim’s local protections wouldn’t raise alarms.
Incidents like this raise the question: who really controls an undersea cable (or any other comms link)? The default answer is whoever owns it, but I argue that post-quantum cryptography (PQC) migration is turning cryptographic trust into a new axis of control, independent of ownership.
This piece grew out of Quantum Foresight’s Northern Ireland quantum strategy work, which raised the question of who captures value from infrastructure investment.
Owning a cable no longer always means controlling it
Almost every submarine cable, terrestrial backbone, satellite link and inter-datacentre link still depends on classical public-key cryptography to authenticate and encrypt traffic. The mere prospect of a future cryptographically relevant quantum computer that could break that encryption changes threat models well before it materialises. Threat actors are collecting encrypted traffic today to decrypt once such a computer exists, a doctrine known as harvest now, decrypt later (HNDL). PQC only protects traffic collected after the switch, so data harvested today stays exposed however quickly anyone migrates.
While full PQC migration is still years away, state actors are assessed to be collecting encrypted traffic now, and the incidents below show they have had the means since at least 2010, using low-cost, high-reward tactics such as BGP manipulation and more resource-intensive undersea cable taps. In a BGP attack, the autonomous systems that make routing decisions are induced to send traffic through surveillance-friendly infrastructure. In April 2010, China Telecom announced incorrect paths to about 50,000 IP prefixes for about 18 minutes, affecting traffic to and from US .gov and .mil sites. The traffic then flowed on to its real destination, so users typically noticed nothing. GCHQ’s TEMPORA programme is an example at the resource-intensive end of the scale. Revealed in 2013, it attached intercept probes to transatlantic fibre-optic cables where they land on British shores. By summer 2011 it had probes on more than 200 internet links, each carrying 10 gigabits a second, after a trial at its Bude, Cornwall station.
The implications change who controls a link. The orthodox view puts ownership at the centre, seeing ownership of cables, data centres and the like as control, with the owner deciding who profits and who is exposed to surveillance or disruption. But think of this as an infrastructure space beyond property – as the standards, protocols and zones that quietly govern how the world runs – and you will discover power exercised as statecraft, independent of formal property or state law. PQC migration pushes that further, shifting effective control away from owners and toward the states that set or enforce cryptographic standards.
Whoever sets the timeline amasses power and offloads costs
Control over PQC migration, then, is a form of power on par with owning or operating infrastructure. Whoever controls the standards and vendor roadmaps for the migration decides which, and whose, infrastructure stays trusted. They also decide which infrastructure is written off as outdated or treated as a risk, regardless of ownership. That makes cryptographic trust a third axis of power alongside legal ownership and the physical reality of where cables run and what hardware actually carries the signal. Trust is what lets the operators of routing, hosting and platforms function as a single system at all, so a cable owned by a sovereign-aligned operator that runs unmigrated classical crypto is untrusted infrastructure, while a foreign-owned but PQC-hardened link may be more trustworthy.
Right now, this hits hardest on sovereign AI. National AI sovereignty scorecards measure compute capacity, data access, talent and model ownership. This ownership-first view misses how communications infrastructure is controlled. A state that leads on every conventional sovereignty metric but runs its AI stack over cryptographically stagnant infrastructure vulnerable to HNDL isn’t as sovereign as it would seem. And even sovereign AI infrastructure that does move to PQC gets there on a schedule set largely by the US.
In the US, the National Institute of Standards and Technology (NIST) finalised its first PQC standards in August 2024, triggering a large-scale mandated cryptographic migration, with the NSA’s CNSA 2.0 timeline requiring new US national security system acquisitions to be compliant from January 2027. It mandates exclusive PQC use across operating systems and cloud services by 2033, and quantum resistance across all national security systems by 2035. Internationally, everyone has to pay the bill over the next decade on a US-decided timeline.
That bill is substantial and unequally divided. In 2025, Cloudflare put PQC adoption above 50% of human traffic, against 3.7% of the origin servers sitting behind its network. The gap comes from market concentration on the client side. A handful of major browser vendors and CDN operators can switch on support for billions of connections at once. On the origin side, every server operator has to upgrade individually, pushing more of the cost onto the long tail of origin operators.
Smaller telecom operators and legacy satellite constellations sit further out on that same tail, as migrating older, heterogeneous backbone can take years or even decades. It’s even worse for rural telecom operators who are more likely to depend on bandwidth-constrained cellular and microwave backhaul and on satellite links. PQC’s overhead is an order of magnitude larger, hurting constrained networks most. For these operators the bandwidth-hungrier replacement protocol adds to their legacy migration cost, which disadvantages them against urban operators on fibre.
If you run, own, or use infrastructure, time before PQC migration runs against you
Realistically, we’re looking at two decades or more of internet-scale collection, documented from 2010. The UK NCSC’s milestones are 2031 for early, highest-priority migrations and 2035 for completion. Every day of delay benefits the collector, with no offsetting cost. Potential future value extraction, for example through industrial espionage, is all but assured against anyone outside the sovereignty-protected core.
Operators of routing, hosting and platforms that migrate early gain the same trust premium. They become the preferred route or host, with leverage over the owners of the infrastructure.
For anyone on either side of that divide, a few practical questions will determine winners and losers.
Who sets the migration timeline for the infrastructure you depend on, and are you on the early or the late side of it?
Who pays for PQC migration in your supply chain, directly or indirectly, and how much of that cost can they externalise?
How do you mitigate against exposure of the traffic that has already been and is still being collected today? (PQC only protects traffic collected after the switch, so anything until then is exposed)
Written by Giorgos Georgopoulos, Quantum Foresight and SAMI Associate. Quantum Foresight works with governments and industry on what quantum technologies mean for them. Our Northern Ireland strategy work covered PQC migration alongside quantum sensing, photonics and computing adoption.
The views expressed are those of the author(s) and not necessarily of SAMI Consulting.
Achieve more by understanding what the future may bring. We bring skills developed over thirty years of international and national projects to create actionable, transformative strategy. Futures, foresight and scenario planning to make robust decisions in uncertain times. Find out more at www.samiconsulting.co.uk




Comments