Thesis · No. 04
Economics

The Luxury
Trap.

The subscription economy, manufactured necessity and how AI became the final link in a chain that turned everything — including thought itself — into a recurring charge.

Malta Insider
September 2026
13 min read

In 1980, the median American household income was $21,020. That income bought a modest home, a reliable car, a family vacation, and a retirement with reasonable security — on a single paycheck, for a family of four. You owned a television. You owned a radio. You owned your music, on vinyl or tape. You did not subscribe to anything except perhaps a newspaper and a telephone line.

By 2025, the median household income had grown to approximately $80,000 — nearly four times higher in nominal terms. But the median home price had risen from $64,600 to $410,000. Healthcare costs had multiplied. And somewhere along the way, ownership had been quietly replaced by access. You no longer owned your music. You no longer owned your software. You rented your entertainment, your productivity tools, your communication platforms, and increasingly, your thinking.

The question this report asks is not whether any individual subscription is worth the money. Most of them are, individually. The question is what happens when the model itself — the recurring charge, the automatic renewal, the frictionless extraction — becomes the dominant structure of economic life.

You no longer owned your music, your software, or your entertainment. Now you don't own your thinking either. You rent it.

How Ownership Became Access

The shift from ownership to access happened gradually, then all at once. Each step felt like progress — more convenience, lower upfront cost, always up to date. The aggregate effect is a consumer who owns less and less, while paying more and more, indefinitely.

1990s
Software as a product
You bought Microsoft Office for $150. You owned it. It ran forever. Updates were optional. The software industry's revenue was transaction-based — sell the product, move on.
2000s
Media as access
iTunes replaced the record store. You bought songs at $0.99 each — still ownership, but digital. Then Spotify arrived. Why buy when you can access everything for $10/month? The record labels, after initial resistance, embraced the model: predictable recurring revenue, no piracy, permanent lock-in.
2010s
Software as a service
Adobe Creative Suite: $2,500 to own, forever. Adobe Creative Cloud: $55/month, forever, or until you cancel — at which point your files become inaccessible without the software to open them. Microsoft followed. Every major software company followed. SaaS became the dominant business model of the decade.
2020s
Intelligence as a service
ChatGPT Plus: $20/month. Claude Pro: $20/month. Copilot: bundled into Microsoft 365. Gemini: bundled into Google Workspace. Your cognitive assistant — the tool that writes, reasons, analyses and decides on your behalf — runs on a recurring charge. Cancel, and the thinking stops.

The Numbers Behind the Trap

The data on subscription spending is remarkable — not because the individual numbers are large, but because of the gap between what consumers think they spend and what they actually spend.

$219
Actual monthly subscription spend per US consumer
$86
What US consumers think they spend — a 2.5× gap
8.2
Active subscriptions per consumer when all services counted

The average American consumer spends $219 per month — $2,628 per year — on subscriptions across 8.2 active services. They estimate they spend $86. The gap is 2.5 times. Three-quarters of consumers say it is easy to forget about recurring charges. 72% have all subscriptions set to auto-pay. 42% have forgotten about a subscription entirely while still being charged for it.

This is not consumer irresponsibility. It is the subscription model working exactly as designed. Small recurring charges, spread across multiple payment methods, billed on different dates, with friction deliberately removed from the payment experience — and deliberately injected into the cancellation process.

Subscription Spending Statistics 2026
Americans estimate $86/month on subscriptions but actually spend $219/month — a 2.5× perception gap. 74% say it's easy to forget recurring charges. 72% have all subscriptions on auto-pay.
Aggregated 2025–2026 data from Mastercard, Recurly, Zuora, West Monroe

The global subscription economy was valued at $557 billion in 2025 and is projected to reach $2.5 trillion by 2035. Subscription businesses have grown 3.4 times faster than S&P 500 companies over the last decade. The model is not a delivery mechanism for products. It is a mechanism for extracting recurring revenue from consumers who have been structurally discouraged from tracking what they spend.

The Manufactured Necessity

Every generation experiences the same phenomenon, which economists call lifestyle creep and sociologists call the ratchet effect: former luxuries become perceived necessities. What your parents considered a treat, you consider standard. What you consider standard, your children will consider a minimum.

But the current wave is different in a specific way. Previous luxury-to-necessity transitions were driven primarily by genuine utility gains — cars really were more useful than horses, indoor plumbing really was better than a well. The subscription economy manufactures necessity through a different mechanism: it replaces ownership with access, then makes the underlying capability unavailable without the subscription.

You cannot open a Photoshop file without Photoshop. You cannot access your Microsoft documents without a Microsoft 365 subscription. Your files, your work, your creative output — all of it is held hostage to a recurring charge. The necessity is not manufactured by making the product better. It is manufactured by making the alternative unavailable.

"Subscriptions are not just a pricing model. They are a control architecture. They convert a one-time transaction into a permanent dependency — and the dependency is reinforced by making the alternative (owning the thing you need) increasingly unavailable."
Just Pricing — Subscription Economy Statistics, 2026

The Rentier Economy

There is a term in economics for an entity that extracts income not from producing value but from controlling access to something others need: a rentier. Historically, rentiers were landlords — they owned land, and everyone who wanted to use the land paid rent. The economic critique of rentierism is that it extracts value without creating it; the landlord does nothing between rent payments to justify the charge.

The subscription economy is the digital extension of this model. The software company owns the code. The streaming platform owns the licences. The AI company owns the model. You pay rent — month after month, year after year — for access to capabilities you depend on. If you stop paying, the capability disappears. If the company raises prices, you pay or go without.

The AI subscription is this model at its most intimate. Previous rentier relationships were about access to external tools — software, media, platforms. The AI subscription is about access to a cognitive capability that, as we documented in the previous report, the user may be gradually losing the ability to perform independently. The dependency and the cognitive atrophy accelerate each other.

The AI subscription is rentierism at its most intimate: you rent access to a cognitive capability you are simultaneously losing the ability to perform without it.

Why 2026 Is Different

The subscription economy has existed for decades. What makes 2026 different is the convergence of several trends that have crossed a threshold simultaneously.

Four thresholds crossed simultaneously
1. Ownership has become genuinely unavailable. For most software categories, there is no longer a "buy once" option. Adobe, Microsoft, and most professional tools exist only as subscriptions. The choice is not "subscribe or buy" — it is "subscribe or go without."

2. Cognitive tools are now in the stack. AI assistants are no longer productivity accessories. For knowledge workers, they are increasingly core to daily work. The subscription is not for entertainment or convenience — it is for the ability to function competitively in the labour market.

3. Subscription fatigue is real but ineffective. 41% of consumers report subscription fatigue. Yet 44% spent more on subscriptions in 2025 than the year before. The fatigue exists. The behaviour does not change. The friction asymmetry — easy to subscribe, hard to cancel — ensures the model continues to extract even from exhausted consumers.

4. Wages have not kept up with subscription costs. The median home price is now five times median income. Every category of essential spending has grown faster than wages. The subscription costs add to a base of financial pressure that did not exist in the same form a generation ago.

Thinking as a Subscription

The final, strangest link in this chain is AI — and it deserves to be understood clearly.

Every previous subscription extracted payment for access to content someone else created, or tools that helped you create. Spotify charged for music others recorded. Adobe charged for tools you used to make your own work. The labour was still yours. The output was still yours. You paid for the instrument, not the performance.

AI subscriptions are different in kind. When you pay for Claude Pro or ChatGPT Plus, you are not paying for a tool that helps you think. You are paying for something that thinks instead of you — drafting, reasoning, analysing, deciding, and presenting you with conclusions. The cognitive labour is being performed by the system you are renting. Your role is to prompt and accept.

As documented in our previous report, this outsourcing is associated with measurable cognitive atrophy — a decline in the specific skills being delegated. The implications for the subscription model are significant: as the user's independent capability declines, their dependency on the subscription increases. The product creates its own demand by eroding the alternative.

This is not a conspiracy. It is a structural consequence of a model that extracts value from recurring dependency and a capability that atrophies with disuse. The incentives all point in the same direction. The outcome is foreseeable — and it is already happening.

Fortunly · Subscription Statistics 2026
44% of US consumers spent more on subscriptions in 2025 than the year before. $321/year wasted on unused subscriptions. Global subscription economy projected to reach $1.2 trillion by 2030.
Mastercard, Recurly State of Subscriptions 2026

The Exit Problem

There is an obvious response to all of this: cancel. Audit your subscriptions, eliminate what you don't use, resist new ones, buy software where ownership is still available, and maintain your own cognitive capabilities through deliberate practice.

This is correct advice. It is also advice that a small minority will follow, because the system is not designed to be exited. Every subscription service is optimised for retention. Cancellation flows are deliberately obstructed. Auto-renewal removes the moment of conscious recommitment. Price increases are timed for periods of low scrutiny. The "pause" function — usage of which rose 337% in 2025–2026 — exists specifically to interrupt cancellation without resolving the underlying dissatisfaction.

The subscription economy's most powerful feature is not the recurring charge. It is the psychology of continuation. The default is to keep subscribing. Changing the default requires an active decision, made repeatedly, against constant friction. Most people, most of the time, do not make that decision.

For AI subscriptions specifically, the exit problem has an additional dimension: by the time you consider cancelling, your independent capability may have atrophied sufficiently that the prospect of going without is genuinely threatening to your work. The product has created the dependency it then charges you for.

By the time you consider cancelling, your independent capability may have atrophied sufficiently that the prospect of going without is genuinely threatening. The product creates the dependency it charges you for.

This is the luxury trap. Not a trap set by malice, but by the perfectly rational logic of a system optimised for recurring revenue, operating on a human psychology optimised for convenience, in an economy that has replaced ownership with access at every level of daily life.

The 1980 household had fewer choices and lower comfort. It also had no subscriptions. Its income was not distributed across dozens of recurring charges to entities it was structurally unable to exit. It owned what it used. The things it owned did not get smarter, and did not atrophy when the payment stopped.

We are unlikely to return to that model. But understanding the trap is the first step toward navigating it — and toward building the political and regulatory frameworks that might eventually constrain the most extractive features of a system that, left unchecked, will continue to run precisely as designed.


Sources cited in this report
Subscription Spending Statistics 2026 — Aggregated data from Mastercard, Recurly, Zuora, West Monroe. March 2026.
45 Subscription Economy Statistics for 2026 — Just Pricing. April 2026. Includes Zuora SEI data.
Subscription Economy Statistics 2026: Revenue Surge — SQ Magazine. February 2026. Market size projections to 2030.
How Middle-Class Income in 1980 Compares to 2025 — Yahoo Finance / GOBankingRates. 2025.
The Hidden Cost of Mandatory Luxury — Money Talks News. February 2026.
Lifestyle Creep — Wikipedia. Economic concept of luxury-to-necessity ratchet.
Next in Thesis · No. 05
The Sentinels at the Gate
The AGI race, the Prisoner's Dilemma of frontier labs and why no single country or company can afford to stop — even if they wanted to.
← No. 03: Are We Getting Stupider? Back to Thesis