Global semiconductor revenue is projected to climb to $1.6 trillion in 2026, a 92% jump from 2025's $809 billion, according to new forecasts from Gartner. The surge is being powered almost entirely by memory chips, as AI infrastructure buildouts reshape demand across servers, smartphones, and PCs — and reset pricing across the entire semiconductor supply chain.
Memory Leads the Surge
Memory chips are the single biggest driver of this growth wave, with revenue forecast to reach $837 billion in 2026 — 54% of total semiconductor revenue. Within that segment, DRAM is expected to grow 246.6% and NAND 371.9%, fuelled by both structural AI demand and an unusually strong pricing cycle.
According to Gartner Director Analyst Ben Lee, advanced AI accelerators require large volumes of high-bandwidth memory, while AI servers are also driving up demand for conventional server DRAM, and expanding AI datasets are pushing enterprise NAND SSD demand higher on the storage side.
Pricing is playing an outsized role too. Manufacturing capacity has been redirected toward higher-value products like HBM and enterprise memory, tightening supply for other markets — leading Gartner to describe the 2026 memory revenue surge as heavily price-led, even as underlying capacity needs continue to grow structurally.
"AI infrastructure is consuming much more memory per system."— Ben Lee, Director Analyst, Gartner
Resource Reallocation and OEM Pressure
Because AI is absorbing such a large share of the industry's growth and investment, chipmakers are aggressively redirecting DRAM capacity toward HBM, NAND capacity toward enterprise applications, leading-edge wafer capacity, advanced packaging, substrate and interconnect capacity, and semiconductor capital spending overall.
That reallocation is squeezing other parts of the market even where consumer demand hasn't dropped off. Higher memory prices raise the bill of materials for smartphones and PCs, which can suppress unit production or push OEMs to ship devices with reduced memory configurations — a cost of AI-driven resource reallocation that Gartner's Q2 2026 research flags explicitly.
The shift looks structural rather than temporary: the AI data centre ecosystem is expected to grow from 36.5% of total semiconductor revenue in 2026 to more than 53% by 2030.
Key Ecosystem Players
NVIDIA: As the primary driver of enterprise AI hardware demand, NVIDIA leans heavily on HBM3e/HBM4 integrations for its Blackwell-generation architecture, with its wafer and memory commitments shaping global supply chain allocations.
Samsung Electronics: As a leading integrated memory manufacturer, Samsung's reallocation of fabrication lines toward high-margin AI memory is directly shaping the 2026 pricing and capacity cycle across DRAM, HBM, and enterprise NAND SSDs.
TSMC: As the leading manufacturing partner for advanced silicon, TSMC produces the 3nm and 2nm chips behind today's AI accelerators, with its CoWoS advanced packaging capacity acting as the key bottleneck for shipping leading-edge AI chips to hyperscalers.