Every AI capex chart published this year gets laid over the same overlay: 1999-2000. Capex-to-sales ratios, VC concentration, IPO froth, Fed warnings — the comparison is everywhere, and it is not baseless. But it is incomplete in a way that matters, because it measures this cycle using the only historical yardstick available and assumes the yardstick still applies.

The Comparison Investors Keep Reaching For

The parallels are real and worth taking seriously. Big Tech’s combined AI capex is expected to reach $650-700 billion in 2026, and one Wells Fargo analysis noted this drove 42% of Q1 GDP growth while representing 2.4% of total US GDP — a figure some analysts expect to surpass dot-com-era peaks by Q4 2026. MIT research found that 95% of organizations report zero measurable return on their GenAI investments so far, and the ratio of infrastructure spending to actual AI software revenue runs close to twenty-to-one. Morgan Stanley has pointed out that capex-to-sales ratios are on track to exceed the 32% peak hit at the top of the dot-com bubble.

The chip complex did not drift lower on Friday. It broke on a macro catalyst, and that distinction governs everything about the week ahead. The Philadelphia Semiconductor Index closed at 12,220.7, down 1,396.73 points, a 10.26 percent single-day collapse. A retracement back to the prior highs inside five sessions requires the catalyst itself to be unwound, and the catalyst was a jobs report that cannot be un-printed before the next data drop. The base case is stabilization and a partial bounce, not a clean rip to fresh records.

The consensus framing around Adobe is wrong. The market has spent two years debating whether Midjourney or Firefly or some yet-unnamed generative model will out-feature Photoshop. That is the wrong question. The right question is whether the category of human-directed creative production software retains its structural position in the content supply chain at all. The answer is increasingly no, and that answer is not cyclical.

The Category Is Being Hollowed Out, Not Competed Against

Adobe’s moat was never purely about features. It was about the irreducible requirement for a skilled human operator to sit between intent and output. A creative director needed Illustrator. A motion graphics artist needed After Effects. A print production manager needed InDesign. The software was expensive because the operator’s time was expensive, and the operator’s time was expensive because the skill was scarce. Adobe captured a tax on that scarcity.

Reddit, Twilio, Roblox, and Visa are among the companies reporting earnings on April 30. The date collision is not coordinated, but the coincidence is analytically useful — the same afternoon will produce signals from digital advertising, developer communications infrastructure, virtual economy engagement, and global payment volume simultaneously.

Each of these represents a distinct layer of the digital economy, and the proximity of their reports creates a rare opportunity to read the stack vertically rather than treating each as an isolated sector event.