The 'Memflation' Crisis: Why AI Infrastructure Costs Are About to Explode

The global semiconductor industry is expected to reach US$975 billion in annual sales in 2026, fueled by an intensifying AI infrastructure boom, with growth accelerating from 22% in 2025 to 26% in 2026. But that headline number hides something uglier.

The world will spend $1.3 trillion on semiconductors in 2026, marking the largest growth in two decades, but memory prices will increase 125% in 2026, while storage chip prices will climb 234%.

"Memflation will destroy, or at least delay, non-AI demand into 2028, to varying degrees depending on the application"—which means every non-AI tech project just became significantly more expensive.

While high-value AI chips now drive roughly half of total revenue, they represent less than 0.2% of total unit volume. Translation: the semiconductor industry is sacrificing entire product categories (smartphones, PCs, IoT) to fuel AI infrastructure.

There's also a packaging bottleneck. Advanced Semiconductor Assembly and Test (ASE), the world's largest outsourced semiconductor assembly company, sees advanced packaging sales doubling in 2026, with Nvidia having reserved the majority of TSMC's leading CoWoS technology. Capacity is fully booked. Prices will rise accordingly.

My take: AI infrastructure is cannibalizing the entire semiconductor market. What's not being discussed: this math works only if AI companies can monetize at an exponential rate. If they can't, we're building $200B worth of idle capacity by 2027. And when memory prices stabilize, the margin compression will be brutal.

Sources