Gartner Lifts 2026 Chip Forecast to $1.6T as AI Memory Demand Tightens Supply

Gartner now expects worldwide semiconductor revenue to reach $1.6 trillion in 2026, up from $809 billion in 2025. The forecast puts AI infrastructure at the center of a market whose fastest-moving pressure point is memory: the chips that AI servers need in growing quantities alongside processors and accelerators.

The new estimate is higher than Gartner’s previous $1.3 trillion forecast. It is a projection, not a reported sales total, but its scale reflects a sharp change in what the industry expects to sell into AI buildouts over the next two years.

The new arithmetic

AI clusters are becoming larger and faster, according to Gartner, increasing their need for CPUs, accelerators, interconnects, power equipment and memory. The firm expects AI data center revenue to account for more than 30% of the semiconductor market in 2026, rising to 53% by 2030.

That mix matters because the forecast is not driven only by the specialized processors associated with AI. Gartner expects memory to account for more than 54% of semiconductor revenue growth in 2026. It projects NAND flash revenue to grow 372% and DRAM revenue to grow 247% during the year.

Memory becomes the limiting input

Gartner ties the memory surge to continued AI infrastructure deployments, higher memory content in each AI server, and sustained demand for high-bandwidth memory, or HBM. Added fabrication capacity is expected next year, but Gartner still expects supply-demand conditions to remain tight as AI deployments use more memory.

The demand side is also expanding. Gartner projects hyperscaler spending on AI infrastructure will rise by more than 50% in 2026, increasing demand for GPUs, tensor processing units and other accelerators. That spending forecast is a key condition beneath the broader $1.6 trillion estimate.

The purchasing problem outlasts the surge

For enterprise buyers, the forecast carries a less welcome implication: Gartner does not expect meaningful pricing relief until late 2027. It expects higher prices in the first half of 2026, followed by persistent but moderating increases through the rest of the year. Gartner senior principal analyst Rajeev Rajput advised CIOs to be cautious about supply agreements with unfavorable pricing that extends beyond 2027.

Gartner expects the market to reach $1.9 trillion in 2027. Whether that trajectory holds will depend heavily on the same forces that lifted the forecast: sustained AI infrastructure spending, the amount of memory required per server, and whether new capacity can ease a market Gartner expects to stay constrained through most of next year.