The Amazon Tax: How Platform Rent-Seeking Threatens Developer Autonomy and Open Source

Amazon's advertising machine now extracts nearly a billion dollars a week from merchants who often have no choice but to pay. The same dynamics are creeping into developer ecosystems.

Seth Godin recently called it "legal theft." On his blog, the marketing author described how Amazon's search advertising business has ballooned into a system where sellers must buy ads just to protect sales they would have earned organically. His publisher started buying search ads on Amazon to promote his new book, not because the ads would find new readers, but because without them the book would be buried beneath competitors who did pay. The "Amazon tax" isn't a tax at all, Godin argues, because taxes fund public goods. This is pure rent extraction by a platform that controls the marketplace.

The framing resonated far beyond retail. For developers, open-source maintainers, and small software companies that build on top of cloud platforms, the pattern is uncomfortably familiar.

How the Amazon Tax Actually Works

Godin's core observation is simple: Amazon makes enough from search ads each week to give every one of its employees a $35,000 cash bonus and still have money left over, as he wrote on Seth's Blog. That revenue comes from degrading the very search experience Amazon built its reputation on.

When someone searches for an air fryer, Amazon already knows which model has the best reviews, lowest return rate, and best price. The ads don't help the shopper find that product. They exist to redirect the shopper toward a different product, or to force the maker of the best air fryer to pay for visibility it should already have. As Godin puts it, "the ads make the search worse."

This creates a compounding trap. Once enough competitors buy ads, everyone has to buy ads. Not because the ads work in any traditional sense, but because not buying them means disappearing. The platform captures the surplus. The merchant pays the tax.

Traditional advertising creates demand. Platform advertising, in Godin's framing, merely redistributes existing demand while siphoning off margin. And the dynamic isn't limited to retail.

From Storefronts to Software: The Developer Version

The Amazon tax has a direct analogy in cloud computing and developer platforms. AWS, Azure, and Google Cloud have become the infrastructure layer for most of the internet. As we previously reported in our coverage of Amazon's "Project Dawn" layoffs, AWS has become the backbone for roughly a third of the internet. When Amazon restructures its cloud business, it sends ripples through every startup and government system that depends on it.

That dependency is the leverage point. Here's how the developer version of the Amazon tax manifests:

Managed service markups. Open-source databases like Redis, Elasticsearch, and Kafka are freely available. But AWS wraps them in managed services, charges a premium, and often outranks the original project in search results and documentation. The open-source maintainers who built the software get nothing. The platform captures the value.

Marketplace positioning. Just as Amazon's retail marketplace buries products that don't buy ads, cloud marketplaces can privilege first-party services over third-party tools. A developer searching for a container orchestration solution on AWS will encounter Amazon's own offerings before independent alternatives, regardless of quality.

API lock-in as a moat. Platform-specific APIs, proprietary extensions, and tightly coupled tooling make migration expensive. Once you've built on a platform's proprietary services, leaving means rewriting code. The switching cost is the tax.

The pattern mirrors retail exactly: build a marketplace that attracts participants, then extract rent from those participants once they're too embedded to leave.

When the Tax Gets Physical

The rent-seeking instinct isn't confined to software pricing. It sometimes shows up in hardware, too, in ways that feel almost absurd.

In May 2026, 9to5Google reported that Motorola phones had started hijacking the Amazon app to inject affiliate codes. An app update caused Motorola devices, including the $1,900 Razr Fold, to route users through a browser redirect when opening Amazon from the app drawer, silently attaching an affiliate tag. The behavior was invisible unless you were watching closely. Motorola later told 9to5Google the behavior was "unintended."

This is the Amazon tax in miniature. A hardware vendor, presumably seeking a revenue share from Amazon purchases, inserted itself into the transaction between a user and a platform. The user got a degraded experience. The vendor got a cut. The only party that didn't consent was the one paying.

For developers, this kind of affiliate code injection is a cautionary tale about platform layers. Every intermediary between your software and your user is a potential rent-seeker. Every SDK, every app store, every managed service is a layer that can, and eventually might, extract value from traffic that flows through it.

The Open-Source Squeeze

Open-source projects sit at the most vulnerable point in this ecosystem. They provide enormous value, they're freely available, and they have limited ability to capture revenue from the platforms that commercialize their work.

The structural problem is straightforward. A small team maintains a widely used open-source database. AWS offers a managed version of that database. Enterprise customers choose the managed version because it's easier to deploy and comes with support. The managed service generates significant revenue for AWS. The maintainers continue to work for free, or for donations that cover a fraction of their costs.

Several open-source projects have responded by changing their licenses. MongoDB, Elastic, and Redis have all adopted variations of source-available licensing that restrict cloud providers from offering competing managed services. These moves are controversial. They protect maintainer revenue, but they also fragment the open-source ecosystem and create legal uncertainty for users.

The deeper issue is that "open source" as a sustainability model assumed a world where the primary value capture happened at the application layer, not the infrastructure layer. Cloud platforms inverted that assumption. When the infrastructure provider can wrap any open-source project in a managed service and sell it at scale, the people who write the code have almost no leverage.

Beyond Amazon: A Structural Pattern

It's worth noting that this isn't uniquely an Amazon problem. As Phenomenal World explored in a July 2026 analysis comparing Amazon and Costco's business models, Amazon's approach represents one end of a spectrum: infinite assortment, ultra-fast delivery, and a logistics infrastructure so complex that it becomes its own justification. Costco sits at the other end, with limited selection, bulk purchasing, and a model that has delivered more than 10 percent annual revenue growth over the past five years, according to the analysis.

The contrast matters for developers because it suggests the Amazon model isn't inevitable. Costco's success shows that curation and constraint can compete with infinite choice and platform extraction. In software terms, this might look like opinionated frameworks, curated package registries, or infrastructure cooperatives that prioritize sustainability over scale.

Some of this is already happening. Developer cooperatives, foundation-backed infrastructure projects, and alternative hosting providers offer paths outside the major cloud platforms. But they remain small relative to the hyperscalers, and the gravitational pull of AWS, Azure, and Google Cloud is strong.

What Developers Can Actually Do

The Amazon tax isn't going away. Platform rent-seeking is a structural feature of concentrated digital markets, not a bug that better regulation will fix overnight. But developers aren't powerless.

First, architecture choices matter. Building on open standards and portable abstractions reduces switching costs. Kubernetes, for all its complexity, exists partly because it makes cloud migration possible. Choosing platform-agnostic tools is a form of insurance.

Second, licensing choices matter. The shift toward source-available licenses is imperfect, but it reflects a real attempt by maintainers to capture some of the value their work creates. Developers who depend on open-source tools should understand these licensing changes and support them where they align with long-term ecosystem health.

Third, awareness matters. Godin's framing is useful precisely because it names the dynamic clearly. The Amazon tax isn't a fee for service. It's a toll extracted by a gatekeeper. Recognizing it in retail, in cloud pricing, in app store policies, and in affiliate code injection is the first step toward building systems that resist it.

The platforms will keep finding new ways to extract rent. The question for developers is whether they'll build in ways that make extraction easy, or in ways that preserve their autonomy. The answer, increasingly, is an architectural decision as much as a political one.