There is a number sitting on the Bank of England's balance sheet that shouldn't exist. In 2010, it read £50 billion in outstanding banknotes. Today it reads roughly £99 billion — nearly double. Stop there and you'd conclude Britain is in the grip of a quiet cash revival, that households spooked by digital fragility are stuffing mattresses the way my grandmother's generation did in the war.
Now read the second number. Cash accounts for just 8% of UK payments. And of that £99 billion in notes, close to 94.9% — roughly £94 billion — is held outside the country entirely.
So the "British cash hoarding" headline is, on its face, a mirage: the nominal figure doubled partly because inflation inflated it and mostly because foreign hands hold sterling as a store of value, not because Mancunians are queuing at ATMs. And yet. Strip the noise and something genuinely strange remains — a payment technology with three centuries of track record is being abandoned for daily use while simultaneously being retained as insurance. That contradiction is the real signal. Reading between the code to find the human story, what emerges is that the world hasn't stopped wanting permissionless money. It has stopped trusting it for convenience, and started trusting it for survival.
To understand why, you have to stop thinking of cash as a payment method and start thinking of it as a technology stack — one of four competing stacks for moving and storing value, each with its own trust assumptions, failure modes, and permission structure. There is physical cash: control by possession, no third-party approval, fully offline. There is digital fiat: a convenience layer, but dependent on electricity, connectivity, and an institution deciding to release your funds. There is gold: a thousand-year store of value, perfectly offline, terrible as a payment rail. And there is the fourth stack — crypto — which this argument never names, and that omission is the most interesting thing about it.
I have spent years mapping this spectrum. When I left traditional finance in 2017 to join a boutique research firm in Zurich, I built a habit of cross-referencing developer activity with social sentiment — a crude "Narrative Velocity" metric that told me capital flows tended to move about two weeks before price did. That discipline taught me something that technical analysis alone never could: money is never merely a medium. It is a bundle of fears, and each technology on the spectrum answers a different fear. Cash answers the fear of system downtime. Gold answers the fear of debasement. Digital fiat answers the fear of friction. And crypto, in theory, answers the fear of censorship — the fear that someone, somewhere, can flip a switch and decide you don't get to transact.
The industry has spent a decade selling that last fear as the master narrative. This piece, published without a single mention of blockchain, tokens, or protocol, quietly demonstrates that the demand is real — while also demonstrating, without meaning to, that the current answers to it are all incomplete. That is worth taking seriously precisely because it comes from the mainstream, not from our own echo chamber.
The historical pattern is consistent. Gold standard, then fiat, then digital rails — each transition removed friction and added a dependency. The permissionless money spectrum is the collection of technologies that refuse to add dependencies, and that refusal always carries a cost.
Here is where the technical argument gets sharp. The function of a banknote is not that it is anonymous or private in the romantic sense. Its function is that it removes a dependency from a transaction. A note in your hand needs no electricity, no network, no counterparty approval, no clearing institution. When part of the system seizes up, life continues. That is not a payment feature; it is a systems-engineering property. It is what reliability engineers call fault isolation — the capacity to keep one failure from cascading into total failure.
This is the same instinct that animates blockchain design, but arrived at from the opposite direction. Crypto removes the intermediary through mathematics; cash removes it through physics. Both aim at the same target — eliminating a single point of trust — but the engineering tradeoffs are inverted, and that inversion is where the honest analysis lives.
Consider the phrase that struck me hardest as I worked through the data: three cards are not three independent backups. If a single outage knocks out every nearby merchant terminal, holding Visa, Mastercard, and Amex in your wallet gives you nothing. You have three credentials sharing one failure domain. I have come to call this pseudo-redundancy — the illusion of diversification when the components actually fail together.
Once you see pseudo-redundancy in physical payments, you cannot un-see it in crypto. When multiple Layer 2 networks settle to the same Layer 1, they share a fault domain. When five cross-chain bridges route through the same custodian or the same signing quorum, they are three cards in one wallet. The DeFi Summer of 2020, which I documented in real time with a viral thread arguing liquidity would consolidate into three hubs, was ultimately a study in exactly this: dozens of forks that looked like independent pools but shared the same oracle, the same incentive reflex, and the same cascade risk when yields turned. When they broke, they broke together.
That is why I have never bought the "liquidity fragmentation" panic that venture capital keeps trying to sell as the next must-solve problem. It is a manufactured narrative, not a structural one. Fragmentation is what healthy competition looks like from the inside. The genuine fragility was never the number of pools — it was the number of shared dependencies underneath them, and no new routing product fixes a shared oracle.
Now bring that lens back to the sound money question. Why would anyone hold an asset with a negative carry — one that pays no yield, actively loses to inflation, and even carries physical storage cost? The data answers cleanly: because they are not buying a return, they are buying an option. Cash is an insurance premium against institutional failure. The holder of a £20 note is not expecting it to appreciate; they are paying a small, slow, certain cost in exchange for a discontinuous, large, uncertain benefit — access to value at the exact moment the digital system says no.
That is a crucial distinction for anyone analyzing the broader market, because it tells us there is a standing, non-speculative demand for assets whose entire appeal is optionality rather than yield. The 2022 collapse of TerraUSD, which I dissected over three weeks by interviewing former validators in Seoul, was compelling for the opposite reason: it was an asset that promised yield and offered no option. When belief systems concentrate rather than diversify, they do not merely fall — they fall instantly. "The Death of Algorithmic Faith" was not really about an algorithm. It was about the danger of holding one dependency and calling it safety.
The official confirmation of this demand is the detail most people are skipping. The Dutch central bank now advises households to keep around €70 per adult and enough supplies for three days, specifically for when electronic payments fail. Read that twice. When a central bank publicly tells citizens to prepare for the moment the digital plumbing stops working, it is not hedging — it is formally acknowledging that its own preferred payment architecture has a single point of failure. In the language of fault isolation, the operator has published a failover plan, and the failover plan is a competing technology.
The empirical record backs it up. Demand for cash spikes in almost every systemic shock: 2008, the Greek crisis, the pandemic, the invasion of Ukraine. April 2025's Iberian blackout produced a visible, measurable jump in cash usage across Spain and Portugal. These are not sentiment events. They are stress tests, and cash keeps passing them while cards keep failing them.
The gold data tells a subtler story, and here the mainstream narrative is being overstated too. Global retail investment demand for bars and coins rose roughly 16% in 2025 to about 1,374 tonnes, the highest since 2013. Impressive — until you disaggregate. US demand fell by volume over the same period, and the growth is concentrated in China and the Middle East. "The whole world is buying gold" is a lazy generalization. What is actually happening is that specific jurisdictions with the largest geopolitical exposure to the dollar system are quietly rotating into a neutral reserve — a monetary-hedging alignment, not a global consensus. Unearthing value where others see only chaos means separating the headline from the histogram.
And the deeper truth in the middle of all this is that the demand being described is fear-driven, not greed-driven. That matters enormously for how long it lasts. Greed-driven narratives inflate and pop on a quarterly rhythm. Fear-driven narratives grow slowly, persist for years, and never fully deflate, because the fear never fully disappears. This is a structural, not a cyclical, bid.
Which brings us to the place the article refuses to go. The demand it documents is real: for offline access, for privacy, for censorship resistance, for a store of value independent of any single government. It then presents two solutions — cash and gold — that each fail on a different axis. Cash fails on debasement and on payment convenience; the prose is unusually candid about this, noting that the same cash that saves you during an outage cannot restart the supermarket's refrigeration or the fuel pump's electricity. Gold fails on divisibility and settlement; selling it can mean a discount at the dealer and a moving price when you need liquidity most.
What is missing from the spectrum is the third answer: something digital, divisible, censorship-resistant, and portable across borders. The article's own logic points straight at it, then walks past it. And here is the part the crypto community will not enjoy hearing. When you apply the article's own test — does it survive a physical infrastructure failure — crypto scores worse than cash, not better. Cash at least works with the lights off. A self-custody wallet needs power, a network connection, and a device in working order. For all the talk of being the ultimate disaster hedge, crypto is strictly more dependent on functioning physical infrastructure than the paper it claims to replace. Reading between the code to find the human story means being honest about which story the code is not equipped to tell.
The same reckoning applies to the grand branding exercises. A large share of what gets marketed as "Bitcoin Layer 2" is, on inspection, Ethereum architecture wearing a Bitcoin ticker — projects that the actual Bitcoin developer community does not recognize as Bitcoin at all. That is not a technical quibble; it is a dependency mislabeled as sovereignty. Likewise, exchange-led launch programs that once returned a hundredfold now struggle to return ten, a decay curve that says more about the exhaustion of traffic monetization than about the underlying chains. Each of these is a case of a shared dependency dressed up as independent value. The pattern is everywhere.
So if the demand exists — and the banknotes, the gold bars, and the central bank advisories all insist it does — why has the digital, permissionless answer not simply won? Because it has not yet resolved the tension that cash resolves by being physical and gold resolves by being inert. It must be simultaneously convenient and unpermissioned, and those two properties pull against each other the way friction and finality always have.
The most honest place to look is the least glamorous. Stablecoins — digital, fast, cross-border, and at least partially resistant to the kind of institutional switch-flipping that the article implicitly worries about — sit precisely in the gap between convenience and permissionlessness. They are not a rebellion; they are a bridge. And bridges, historically, are where the largest flows actually cross. If you are positioning in a sideways market rather than chasing a breakout, the signal to watch is not the next narrative token but the quiet net inflow into the infrastructure that makes digital money feel as immediate and dependable as the note in your pocket.
Here is the contrarian read that most crypto readers will resist. The instinct is to file this article as a bull case for the industry — proof that the world is hungry for what we build. I think that is a misreading, and a costly one. Bitcoin is not an insurance asset; it is a high-beta risk asset, and in genuine liquidity crises it tends to be sold first, not hoarded. The 2020 COVID crash proved it. March 2020 liquidations, the 2022 credit squeeze — in each case, the "digital gold" thesis met a market that wanted dollars and cash, not a volatile risk curve. Holding a narrative that refuses to admit this is itself a pseudo-redundancy: faith in the story is not diversification of belief.
If fear-driven demand for permissionless money persists — and the fear is structural — then the assets that capture it will not necessarily be the ones with the loudest aesthetics or the biggest launch programs. The one thing the whole spectrum shares is that its value comes from removing a dependency, not adding one. Any project whose appeal depends on a shared oracle, a shared custodian, a shared sequencer, or a single brand's continued solvency is not selling permissionlessness. It is selling one card in a wallet full of them.
Which is exactly the trap the Dutch central bank avoided when it told citizens to hold a bit of physical fallback. So here is the question worth carrying into the next quarter of chop: if even a central bank now tells households to keep a failover plan outside its own system, which asset in your portfolio is genuinely offline-independent — and which one have you convinced yourself is, because holding two versions of the same dependency feels safer than admitting you have only one?