The Cloud Is a Lie: What We Lose When Everything Becomes a Service

An exposé on how "the cloud" shifted from decentralization to dependency — and who profits from the illusion of simplicity.

There Is No Cloud

Let's start with the most important truth about cloud computing: there is no cloud. There's just someone else's computer.

When you save a document to "the cloud," you're not uploading it to some ethereal digital realm that exists everywhere and nowhere. You're putting it on a specific hard drive in a specific data center owned by a specific corporation in a specific legal jurisdiction. The cloud is a metaphor—and like all metaphors, it shapes how we think about what it describes.

The cloud metaphor is brilliant marketing. Clouds are soft, fluffy, benign. They float above us, accessible from anywhere. They're natural, almost magical. Who wouldn't want their data stored in something that sounds so pleasant and abstract?

But metaphors can obscure as much as they reveal. When we think of our files as existing "in the cloud," we stop thinking about where they actually are, who controls them, and what it costs us to access them. We stop asking basic questions about ownership, privacy, and dependency. We accept the cloud the same way we accept the weather—as something that simply exists, beyond our control or comprehension.

This isn't an accident. The cloud metaphor was chosen specifically to make us stop thinking critically about what's actually happening to our data, our software, and our digital autonomy.

The Great Rebundling

To understand how we got here, we need to go back to the beginning of personal computing. In the 1980s and 90s, when you bought software, you bought software. A physical box with physical disks that installed programs on your physical computer. You owned it, completely. You could use it forever, install it on any computer you wanted, and the company that made it couldn't take it away from you.

This model had problems, of course. Piracy was rampant. Support was expensive. Upgrades required convincing people to buy new versions. But from a user's perspective, it meant something fundamental: you owned the tools you used.

The internet changed the economics. Suddenly, software companies could update products continuously, deliver services instantly, and—most importantly—charge recurring fees instead of one-time purchases. The shift from products to services wasn't primarily about making software better for users. It was about creating permanent revenue streams.

"Software as a Service" (SaaS) emerged as the dominant business model. Don't buy Microsoft Office—subscribe to Microsoft 365. Don't own Photoshop—rent it through Creative Cloud. Don't install anything locally—just log in to our web app and trust us to keep it running.

The pitch was compelling: always have the latest version, access your files anywhere, let us handle the technical complexity. No more installing updates, worrying about backups, or dealing with compatibility issues. Just pay a monthly fee and everything works seamlessly.

What they didn't emphasize: you're not a customer anymore. You're a subscriber. And subscribers can be unsubscribed.

The Subscription Trap

Here's the uncomfortable math of subscription services: they're almost always more expensive than ownership over time.

Adobe Creative Suite used to cost about $2,600 for a permanent license—expensive, yes, but you owned it forever. Adobe Creative Cloud costs $60/month, or $720/year. After four years, you've paid more than the perpetual license cost, and you own nothing. After ten years, you've paid $7,200 for the privilege of temporarily accessing software that could be taken away at any moment.

Microsoft Office tells a similar story. Office 2019 cost $150 for a permanent license. Microsoft 365 costs $70/year. After three years, you've paid more than ownership would have cost. And Microsoft is no longer selling perpetual licenses for Office—subscription is becoming mandatory.

The pattern repeats across the industry. Autodesk, the maker of AutoCAD, discontinued perpetual licenses entirely in 2016. Maya, 3ds Max, Revit—all subscription-only now. The industry called this transition "inevitable," as if it were a natural law rather than a business decision.

For software companies, the economics are irresistible. A customer who pays once is worth far less than a customer who pays forever. Subscriptions provide predictable recurring revenue that investors love. They create lock-in—once your workflow depends on a service, canceling becomes painful. They enable price increases that customers have little choice but to accept.

But there's a deeper shift happening here beyond just pricing models. We're moving from a world where we own our tools to a world where we perpetually rent them. And rental always benefits the landlord more than the tenant.

The Hidden Costs

The true cost of "everything as a service" goes far beyond monthly fees. It's a comprehensive transfer of control from users to providers, with implications most people never consider until it's too late.

You can lose access at any time. Miss a payment? Lose your files. Account flagged by an algorithm? Locked out. Company changes terms of service? Accept or leave. Service shuts down? Everything disappears. Your access to tools you depend on is contingent on maintaining good standing with corporations who set and can change the rules unilaterally.

Your data isn't private. When files live on someone else's servers, they can be scanned, indexed, analyzed, and monetized. Cloud providers routinely use customer data to train AI models, serve targeted ads, and build user profiles. They can be compelled by governments to hand over your data. They can be breached by hackers. The terms of service you agreed to without reading probably grant them extensive rights to your content.

You're dependent on internet connectivity. Cloud services require internet access to function. In theory, most offer "offline modes," but they're often limited and buggy. If your internet goes down, if you're traveling somewhere with poor connectivity, if your service provider is experiencing an outage—you lose access to your tools and files. Your ability to work becomes dependent on infrastructure you don't control.

Features can disappear. When software is a service, the company can remove features at will. That plugin you relied on? Discontinued. That workflow you perfected? Updated away. Your tools change underneath you, and you have no recourse. With owned software, what you have keeps working. With services, you're always at the mercy of someone else's product roadmap.

You can't use older versions. Sometimes software updates make things worse. They introduce bugs, change interfaces, remove features, or increase hardware requirements. With owned software, you could simply not upgrade until bugs were fixed. With services, you get whatever version they're currently serving. No rollbacks, no choice, no control.

Lock-in is structural. Once your workflow and data are built around a service, leaving becomes extremely painful. Your files might be in proprietary formats. Your integrations might depend on specific APIs. Your team might be trained on particular interfaces. Switching costs increase over time, giving providers more pricing power and less incentive to serve you well.

Enshittification is inevitable. Cory Doctorow coined this term to describe the predictable lifecycle of platforms: first they're good to users to build market share, then they abuse users to serve business customers, then they abuse business customers to serve themselves. Every successful service eventually becomes worse because the incentives demand it. And you can't switch back to a version that worked better.

The Illusion of Simplicity

Defenders of cloud services argue that they're easier, more accessible, and more democratizing than traditional software. Let anyone create without needing technical expertise! Free people from the burden of maintenance! Make sophisticated tools available to everyone!

This argument deserves serious consideration—but it's built on a false premise. The simplicity of cloud services isn't inherent to the technology. It's a design choice enabled by controlling the entire stack and limiting what users can do.

Local software can be simple too. Apps like Notion and Obsidian demonstrate that sophisticated tools can run locally with elegant interfaces and minimal technical requirements. The difference isn't capability—it's control. Cloud services are simple because they don't let you do anything complex. They're easy because they've removed your options.

Consider what "simplicity" actually means in practice:

Simple for whom? Cloud services are simple for casual users doing standard tasks. They're often actively hostile to power users who want to customize workflows, automate processes, or integrate tools in unconventional ways. The simplicity is achieved by removing depth, which is fine until you need that depth.

Simple until it isn't. Cloud services work seamlessly until they don't. When something breaks, you're completely helpless. You can't troubleshoot, can't access underlying systems, can't fix problems yourself. You file a support ticket and wait. Your "simple" tool has made you dependent on someone else's customer service department.

Simplicity as dependency. By making everything "just work," cloud services ensure you never learn how anything actually works. You don't understand file systems, networking, databases, or any of the underlying technology. This isn't simplicity—it's learned helplessness dressed up as user-friendliness.

Deliberately degraded alternatives. Many companies intentionally make local alternatives worse to push people toward cloud services. Microsoft has systematically degraded the standalone version of Office while improving Office 365. Adobe makes Creative Cloud the only way to access new features. The "simplicity" of cloud services is often manufactured by making everything else more difficult.

The illusion of simplicity serves the provider, not the user. It keeps you dependent, prevents you from understanding your tools, and ensures you can never leave. Real simplicity would mean tools that work well whether you host them yourself or use someone else's server. The cloud doesn't offer simplicity—it offers constraint disguised as convenience.

Who Profits from the Cloud

Follow the money, and the true nature of cloud services becomes clear. This isn't about making technology more accessible or freeing users from complexity. It's about consolidating control and extracting maximum value from that control.

Amazon Web Services (AWS) generated $90 billion in revenue in 2023, with operating margins around 30%. Microsoft Azure and Google Cloud together add another $100 billion. These aren't just profitable businesses—they're becoming the infrastructure layer of the entire digital economy. Every startup, every enterprise, every government agency increasingly depends on cloud providers who can set prices, change terms, and dictate conditions.

The concentration is accelerating. AWS, Azure, and Google Cloud together control roughly 65% of the cloud infrastructure market. Everything above them—SaaS companies, platforms, apps—depends on this oligopoly. And everything below them—physical data centers, networking equipment, power generation—is increasingly optimized for their needs.

This creates unprecedented leverage. Cloud providers can:

• Raise prices and force customers to accept them (switching costs are enormous)
• Mine customer data for competitive advantage (see how AWS cloned MongoDB's business model)
• Use their position to favor their own products over competitors
• Set de facto technical standards that others must follow
• Effectively tax every digital transaction that flows through their infrastructure

The big three cloud providers are the landlords of the digital economy, collecting rent on every server request, every API call, every file stored. And like all landlords, they benefit from scarcity and dependency.

But the profit extraction goes deeper than infrastructure. The shift to services enables business models that were impossible with owned software:

Usage-based pricing means costs scale unpredictably. That app that costs $10/month for basic use might cost $1,000/month if you actually use it heavily. The pricing is deliberately opaque, making it difficult to predict costs or compare alternatives.

Data monetization turns users into products. Free cloud services aren't charitable—they're extracting value from your data, your attention, and your network effects. If you're not paying for the product, you are the product. And even if you are paying, you might still be the product.

Vendor lock-in creates captive markets. Once you've built your business on AWS or your creative workflow on Adobe Cloud, switching becomes a multi-million dollar proposition. Providers know this and price accordingly. The longer you use a service, the more expensive leaving becomes.

Planned obsolescence becomes frictionless. Don't want to support old versions? Just shut them down and force upgrades. Want to monetize a feature that was previously free? Just change the pricing tier. The service model eliminates all user power to resist changes.

The cloud isn't a neutral technological development. It's a business strategy designed to maximize extraction and minimize user power. And it's working brilliantly.

The Bargain We Didn't Know We Made

When we moved our files to the cloud, our software to subscriptions, and our infrastructure to service providers, we thought we were getting convenience. What we actually got was a comprehensive transfer of power from users to platforms.

Consider what we've given up:

Permanence. Nothing lasts in the cloud. Services shut down, companies get acquired, business models change. The files you stored, the tools you relied on, the workflows you perfected—all contingent on continued corporate existence and goodwill. Meanwhile, software I bought in 1995 still runs perfectly.

Privacy. Your files, your communications, your creative work—all visible to the service provider, their employees, their algorithms, and anyone they choose to share with or are compelled to share with. End-to-end encryption exists but is rarely implemented, because it interferes with data monetization.

Control. You can't modify cloud software to suit your needs. Can't automate tasks the provider doesn't support. Can't integrate with tools they haven't approved. Can't preserve workflows they decide to change. Your tools work the way the provider decides they work, and that's that.

Ownership. You own nothing. Not your software, not really your files (check the terms of service), not your account. Everything is a revocable license contingent on continued payment and acceptable behavior as defined by the provider.

Agency. When something breaks, you're helpless. When something changes, you adapt or leave. When prices increase, you pay or find an alternative. The service model systematically removes your ability to solve your own problems.

Resilience. A single point of failure—the service provider—determines whether you can work. Their outages become your outages. Their security breaches become your security breaches. Their business problems become your business problems.

We didn't consciously agree to this bargain. It happened gradually, one convenient decision at a time. Upload this file, subscribe to this tool, move this workflow to the cloud. Each decision made sense individually. Together, they've created a form of technological serfdom.

The Alternative Infrastructure

The cloud isn't inevitable. It's not even particularly efficient for many use cases. The centralized infrastructure model makes sense for certain scenarios—massive scale, geographic distribution, variable workload—but it's been oversold as a universal solution.

For many individuals and organizations, local and self-hosted infrastructure offers better economics, better performance, and better alignment with actual needs:

Personal servers have become incredibly cheap. A $200 used business desktop or a $50 Raspberry Pi can run file storage, photo management, document editing, password management, and dozens of other services. The computing power sitting unused in people's closets exceeds what the early internet ran on.

Network-attached storage (NAS) devices make personal cloud infrastructure accessible to non-experts. Companies like Synology and QNAP sell elegant, user-friendly systems that provide all the functionality of cloud services with none of the recurring costs or privacy compromises. Set it up once, use it forever.

Self-hosted software has exploded in quality and variety. Open-source alternatives to nearly every cloud service exist: Nextcloud for file storage, Bitwarden for passwords, Jellyfin for media, PhotoPrism for photos, Calibre for ebooks. These aren't hobbyist projects—they're sophisticated, actively maintained software that works.

Peer-to-peer systems offer genuine decentralization. Technologies like IPFS, BitTorrent, and various blockchain-based storage systems enable file sharing and collaboration without central servers. Syncthing lets you synchronize files directly between your devices with no intermediary.

Local-first software keeps data on your devices while maintaining the benefits of connectivity. Apps like Obsidian and Linear store files locally but sync when internet is available. You get offline access, data ownership, and privacy while retaining collaboration features.

Hybrid approaches combine the benefits of both models. Use a VPS for a small number of always-on services, but keep large files and personal data local. Use cloud services for collaboration but maintain local backups. Use open protocols that work across providers so you're never locked into one.

The economics often favor self-hosting, especially over time. A $300 NAS with $200 of drives provides 4TB of storage you own forever. The equivalent in cloud storage costs $10-20/month—$120-240/year—and you never own anything. After two years, you've paid more than the hardware costs. After ten years, you've paid thousands of dollars for storage you could have owned.

The Technical Debt We're Creating

The shift to cloud services isn't just changing who owns infrastructure—it's changing what knowledge exists and who has it.

An entire generation is growing up that knows how to use apps but not how computers work. They can tap icons but don't understand file systems. They can upload to the cloud but can't set up a local network. They can use services but can't host them.

This isn't about gatekeeping or claiming people should all be sysadmins. It's about the difference between literacy and consumption. When everyone who used computers had at least basic understanding of how they worked, users had agency. Now, technological literacy is concentrated in a professional class that builds and maintains services for everyone else.

This concentration of knowledge mirrors the concentration of infrastructure. Just as cloud providers become digital landlords, tech workers become digital clergy—intermediaries between users and the technology they depend on. This isn't healthy for democracy, for innovation, or for individual autonomy.

The decline in technological literacy also makes it harder to envision alternatives. If you've never installed software, never configured a network, never hosted a service, the idea that you could do these things seems absurd. The cloud becomes "just how things work" rather than one option among many.

We're also losing institutional knowledge. The skills to run email servers, host websites, manage databases, and maintain infrastructure are atrophying outside the major tech companies. Smaller organizations increasingly have no choice but to use cloud services because they lack the expertise to do anything else. This further entrenches the power of major providers.

The technical debt compounds. Every organization that moves infrastructure to the cloud is one less organization maintaining local expertise. Every student who learns only cloud platforms is one less person who knows how to build alternatives. Every tutorial that assumes cloud deployment is one less guide for self-hosting.

Eventually, we risk reaching a point where running your own infrastructure seems as archaic as running your own power generation. Technically possible, but so unusual that it's functionally unavailable to most people. The cloud stops being a choice and becomes the only reality.

Breaking the Cloud

Resistance to cloud dependency is emerging from unexpected quarters. Privacy-conscious Europeans are building self-hosted alternatives. Security-focused organizations are repatriating data from cloud providers. Cost-conscious startups are discovering that cloud economics don't work at scale. Hobbyists are creating surprisingly sophisticated homelab setups.

These aren't coordinated movements—they're parallel discoveries that cloud computing's promises don't match its reality. And they're discovering that alternatives aren't just possible but often superior:

Dropbox's outgrowing cloud costs story is instructive. In 2017, they completed a massive migration away from AWS, moving 90% of their infrastructure to their own data centers. The result? They saved $75 million over two years. The cloud isn't actually cheaper at scale—it just shifts capital expenses to operating expenses and masks the true cost until you're locked in.

European universities and governments are increasingly mandating that sensitive data stay within their jurisdictions, leading to growth in local and regional cloud providers. The dominance of US-based providers creates sovereignty concerns that many are no longer willing to accept.

The IndieWeb movement advocates for personal websites and self-hosted services. Members run their own infrastructure, own their data, and use open protocols to interoperate. It's small but growing, demonstrating that alternatives to platform dependency are viable.

Homelab communities on Reddit and elsewhere share knowledge about self-hosting, creating a DIY culture around personal infrastructure. What was once arcane sysadmin knowledge is becoming accessible to enthusiasts willing to learn.

Local-first software is gaining traction among developers tired of the complexity and fragility of cloud-dependent applications. Tools like Electric SQL and Replicache make it easier to build apps that work offline and sync peer-to-peer.

The path away from cloud dependency isn't all-or-nothing. It's gradual, pragmatic, and contextual:

Start with backups. Don't trust cloud providers as your only copy. Maintain local backups of everything important. This gives you options and insurance.

Self-host one thing. Pick a service you use regularly and find a self-hosted alternative. Learn how to run it. The first service is hardest; each additional one gets easier.

Use open protocols. Prefer email over platform messaging, RSS over algorithmic feeds, and open file formats over proprietary ones. Open protocols can't be taken away.

Choose providers that enable export. If you must use cloud services, use ones that make it easy to get your data out. This preserves your ability to leave.

Support alternatives. Open-source projects and ethical cloud providers need financial support. If you use their services, pay for them. Infrastructure isn't free, but it can be aligned with user interests rather than extraction.

Reclaiming the Means of Computation

The cloud promised liberation from the constraints of local hardware and the burdens of maintenance. What it delivered was a new form of dependency, more total and more profitable for providers than anything that came before.

We don't need to accept this. The technology for personal and community-owned infrastructure exists. The software is available. The economics often make sense. What's missing is collective awareness that alternatives are possible and the will to build them.

This isn't about returning to some imagined past where everyone ran their own servers and nothing was easy. It's about creating a future where the digital infrastructure we depend on serves human needs rather than corporate profits. Where we own our tools rather than rent them. Where we control our data rather than surrender it. Where technological competence is widespread rather than concentrated.

The cloud is a lie—not because cloud computing doesn't exist, but because the story we've been told about it obscures more than it reveals. "The cloud" makes it sound natural, inevitable, neutral. "Someone else's computer" reminds us that it's infrastructure owned by specific entities with specific interests that don't necessarily align with ours.

Every time you move data to your own storage, every time you run a service locally, every time you choose open protocols over proprietary platforms, you're reclaiming a small piece of digital sovereignty. These aren't heroic acts of resistance—they're practical choices about who controls the tools you use.

The cloud wants to be the only option. It wants you to believe that self-hosting is archaic, that ownership is impossible, that dependency is inevitable. It wants you to pay forever for things you could own. It wants your data, your attention, and your continued subscription.

You don't have to give it to them.

The computers in your home have more power than the servers that ran the early internet. The software you need is available, often for free. The knowledge required can be learned. The infrastructure can be owned.

The cloud is a business model that profits from your dependence. But dependence is a choice, made one subscription at a time. And it can be unmade, one self-hosted service at a time.

There is no cloud. There's just someone else's computer. And you can have your own.