The Chip Industry's First Trillion-Dollar Year: Why Compute Demand Just Broke Its Own Curve

June 17, 2026
From cyclical to exponential — global semiconductor sales, 2015–2026E. Mar 2026 hits $99.5B (+79.2% YoY); Q1 2026 totals $298.5B, putting the industry on track for its first $1T year.
Quarterly global semiconductor sales, Q1 2014 – Q1 2026. Source: Semiconductor Industry Association (SIA).

The Semiconductor Industry Association (SIA) reported on May 4, 2026 that global chip sales hit $99.5 billion in March — up 79.2% year-over-year. Q1 2026 totaled $298.5 billion, putting the industry on track to clear $1 trillion in annual sales for the first time, roughly two years ahead of the most aggressive 2024 forecasts.[1] The number itself matters less than what it tells us about the shape of the curve.

A cycle that stopped cycling

For most of the last forty years, semiconductors have been a textbook cyclical industry. 18–24 month up-cycles driven by a new end-market (PCs, mobile, cloud), followed by 12–18 month corrections as capacity overshot demand. Founders and investors learned to underwrite the cycle, not the trend.

What the 2024–2026 data shows is that the cycle, for the first time, has not corrected. Inventory burned down in 2023, restocked through 2024, then — instead of plateauing — accelerated. March 2026's +79% YoY is not a base-effect bounce; it comes on top of a +28% comparable in March 2025.[2]

In my view, this is the most important structural shift the industry has gone through in a generation.

What's driving it

One end-market explains nearly all of it: AI training and inference infrastructure.

NVIDIA's data-center segment alone now runs at an annualized revenue rate that would have qualified as a top-five global chip company in 2022. Hyperscaler capex guidance for 2026 — over $400B combined across the four U.S. names — is structurally underwritten by GPU and custom-silicon procurement, much of it on multi-year take-or-pay contracts. TSMC, SK Hynix, and Samsung HBM are sold out through 2027.

The constraint is no longer demand. It's fab capacity, HBM packaging, and the power envelope around it.

What this means for venture

For VC, the takeaway isn't about chip companies themselves. It's that the cost of running AI — training and inference — is falling faster than any prior compute shift, because the underlying chip supply is compounding at 70%+ annually. That cost decline flows downstream to every company building on top of it. In the cloud era, infrastructure costs flattened within a few years; here, they're still dropping. Anything built on this base inherits that.

The takeaway

The chip industry's first trillion-dollar year is not really the story. The story is that it no longer looks cyclical.

Sources

  1. SIA — Global Semiconductor Sales Increase 79.2% Year-to-Year in March (May 4, 2026). semiconductors.org
  2. SIA — March 2025 global semiconductor sales (prior-year comparable). semiconductors.org