The End of the Six-Figure Engineer?

Twenty years ago, Silicon Valley had one story it loved to tell: we're different.

Traditional industries needed factories, equipment, physical labor — all that dead weight dragging on margins. Tech companies just needed a few smart people, some laptops, and a garage. Serve a billion users without meaningfully increasing costs. Print money at the margin. Wall Street loved this story and rewarded it with 30x, 40x, sometimes absurd multiples.

Then AI came along and started quietly rewriting the rules.

A recent piece from Sherwood News caught my attention. Needham analyst Laura Martin was listening to Meta's earnings call and reading between the lines of Zuckerberg's comments about engineering productivity gains of 30%. Her take: if you're only delivering 20-30% efficiency gains as an engineer, your job might be at risk.

But the layoffs themselves aren't what unsettles me — tech does layoffs every year. What's unsettling is the structural reason underneath: these companies are undergoing a fundamental shift from asset-light to asset-heavy. And in that shift, the role of the employee has quietly changed from core asset to optimizable cost.

The Accounting Problem

The financial logic here is simpler than most people realize.

Over the past 18 months, Meta, Microsoft, Google, and Oracle have been on a historic spending spree — AI chips, servers, data centers. All capital expenditure. But here's the accounting reality: when you spend $100 billion on hardware, you don't expense it all in the quarter you bought it. It depreciates over several years, showing up as a steady charge against earnings every single quarter.

Morgan Stanley estimates that just those four companies will rack up $680 billion in AI-related depreciation over the next four years. And that's before you count Amazon, which is spending the most aggressively of all.

That depreciation is fixed. You can't negotiate it away. The equipment is already bought; the charges will hit like clockwork every quarter. So if you need to protect margins, you find savings somewhere else. And the biggest "somewhere else" has always been headcount.

A Bank of America semiconductor analyst put it plainly: these companies' biggest expense used to be employees — especially stock-based compensation. Now they're buying GPUs instead of hiring people. GPUs don't need equity grants.

A Quiet Transformation

The head of accounting research at Morgan Stanley framed it well: for twenty years, these companies told the story of being asset-light, generating cash, buying back stock, no need for debt. All of that changed in the last 18 months.

The reason this matters goes deeper than layoffs. Tech's premium valuation was always built on the asset-light model. Near-zero marginal costs. Profits that could scale exponentially. That's the story Wall Street paid up for.

Now look at what these companies are actually doing: spending hundreds of billions on data centers, signing massive chip contracts with Nvidia, investing in optical networking companies to solve bandwidth bottlenecks. That's not a software company playbook. That's what utilities and industrial conglomerates do.

Evercore found that from 2023 to 2025, revenue per employee jumped 22% at Amazon, 29% at Google and Microsoft, and 48% at Meta. Extraordinary efficiency gains. You'd expect profits to surge.

But the efficiency gains are being eaten by depreciation. Morgan Stanley's point: Wall Street's consensus estimates are essentially pricing in a free lunch. Free lunches don't last.

Tech Is Eating Its Own

In every previous tech boom — dot-com, mobile, the unicorn era — the disruption flowed outward. E-commerce disrupted retail. Streaming disrupted cable. Ridesharing disrupted taxis. Tech itself was always the beneficiary. High-paying jobs multiplied. Engineers became the most sought-after workers on the planet.

This time is different. Goldman Sachs noted in January that AI's impact on the broader labor market is still limited — but inside the tech industry itself, the pressure is already visible.

For the first time, the revolution is eating its own.

Tech companies spent decades replacing retail workers, bank tellers, and newspaper typesetters with software. Now they're discovering that their own engineers can be partially replaced by software. Google says 50% of its code is already written by AI agents, with humans reviewing. Amazon has deployed over a million robots in its fulfillment centers.

This isn't a typical layoff cycle. It's an asset substitution. Capital that used to flow to carbon-based assets (people) is systematically moving to silicon-based assets (GPUs and data centers).

What This Actually Means

None of this means engineers are finished. But I do think the social contract of working in tech is being rewritten.

The old contract: the company provides high salaries, equity, free food, and ping pong tables. You provide your brain and your availability. A relatively fair exchange, because you were the core asset. Without great engineers, nothing got built.

The new contract looks different. The core asset is hundreds of billions in AI infrastructure. Your role is to maximize what that infrastructure produces. You're no longer the irreplaceable center of the operation — you're an operating cost to be optimized. Your value is less about what you can build and more about how effectively you can amplify what the machines can already do.

Two things follow from this.

First, raw execution ability is depreciating fast. If half of Google's code is AI-generated, the half-life of "I write good code" as a value proposition is shrinking. The engineers who will be fine are the ones who can make AI infrastructure produce things it couldn't produce without them.

Second, don't treat tech compensation as a permanent feature of the landscape. When a company's cost structure shifts from people-driven to capital-driven, it starts behaving like every other capital-intensive industry: tight headcount control, higher output expectations per person, less tolerance for redundancy. The era where "can breathe and write code" was sufficient for a six-figure salary was a specific historical window, not a permanent condition.

The deepest irony is this: these companies spent two decades telling Wall Street they were nothing like traditional heavy industry. Now they're becoming exactly that — businesses that continuously burn capital on equipment, absorb massive depreciation charges, and squeeze labor costs to protect their margins.

The disruptors got disrupted. Just not by anyone else. By the technology they built.