Omdia has sharply revisited its forecast for the global semiconductor market in 2026. The consultancy now expects sector revenue to grow by 62.7% this year, driven by a very specific combination: sustained demand for artificial intelligence infrastructure, strong pressure on DRAM and NAND memory, and component shortages that are unlikely to ease clearly until well into 2027.
The figure is striking even for an industry accustomed to intense cycles. Growth will not come solely from selling more chips, but primarily from higher average prices. Omdia warns that the current expansion is mainly fueled by increasing memory costs and components closely linked to data centers, servers, and storage systems, rather than an equivalent increase in units shipped.
Memory returns to the center of the market
The main driver of the change is memory. Omdia forecasts that the DRAM market will nearly double its value compared to 2025, while NAND, a smaller segment, could quadruple. This jump is explained by demand from AI servers, enterprise storage systems, accelerators, high-performance networks, and cloud platforms that need more capacity to train, tune, and run models.
The consultancy also highlights an important side effect: the production of High Bandwidth Memory (HBM), essential for GPUs and AI accelerators, is absorbing capacity within the memory industry. HBM offers more value per unit but also consumes manufacturing resources and does not necessarily translate into more volume available for conventional memories. The result is a dual tension: high-end chips are expensive, and at the same time, the supply of traditional memory is tightening.
This shift has consequences across the entire market. A manufacturer needing conventional DRAM for servers, PCs, smartphones, storage, or consumer electronics might face higher prices because available capacity has been redirected toward more profitable AI-related products. The pressure no longer limits itself to GPUs or large training clusters.
The graph published by Omdia helps illustrate the phenomenon. The data center and data storage segment will lead revenue growth, with a 90% year-over-year increase in 2026 and sales exceeding $700 billion. IC memory is also identified as one of the main growth engines, with a much faster growth rate than the overall market. The clear takeaway is: data centers have become the primary economic driver of semiconductors.
Data centers, servers, and corporate upgrades
Omdia links this surge to a cycle of server refreshes in businesses and exceptional investment levels by hyperscalers. Many organizations are retiring older hardware to run more demanding workloads, from databases and analytics to AI inference and agile applications. Meanwhile, major cloud providers continue buying capacity to sustain their AI services.
This combination pushes designs towards higher-value systems. More memory, greater connectivity, additional accelerators, faster storage, and platforms with better energy efficiency increase the cost per server. Not all additional semiconductor revenue results from increased physical volume; a significant part derives from more expensive configurations and rising average prices.
This point is vital to distinguish between healthy growth and supply tensions. If the market grows because more units are sold to more customers, the industry might interpret the expansion as a broad increase in end-demand. If it primarily rises because prices spike in a scarcity environment, profits become concentrated among fewer providers, while equipment manufacturers and end consumers bear higher costs.
Omdia specifically warns about this risk. Senior principal analyst Myson Robles-Bruce notes that AI’s evolution beyond question-and-answer applications has exponentially increased the demand for memory and processors. But he also raises a fundamental question: which applications will generate enough return to justify current levels of AI infrastructure investment?
Smartphones and consumer electronics will also pay more for memory
The rise won’t be confined to data centers. Omdia sees positive prospects for semiconductor revenues in consumer electronics and wireless applications, though with a caveat: smartphone shipments are expected to remain relatively flat, but chip revenues will grow due to higher memory prices and more expensive bills of materials.
This means a mobile market with limited volume growth could still generate more revenue for the semiconductor industry. Upcoming flagship launches, new foldables, and models with AI-powered photography, local processing, or integrated assistants will demand more memory and higher-value components. Smartwatches, health wearables, and connected devices will also contribute to growth, though far less than data centers and storage.
The cost pressure on materials might eventually impact manufacturers and consumers. Brands could absorb part of the increase, adjust margins, raise prices, or reduce features on more price-sensitive models. In a context of moderate consumer spending, not all will have the same capacity to pass on cost increases.
The risk of a price-driven expansion
Omdia’s report issues a cautionary note for the industry. The current revenue growth relies more on high and rising prices than on volume increases. This dynamic has been seen in other cycles, such as the crypto mining boom or previous memory supercycles, but the consultancy believes the scale and scope of the current phenomenon are unprecedented.
The difference is that AI impacts not just a niche. It affects memory, logic, storage, networks, servers, energy, cooling, and infrastructure software. External factors like tariffs, energy costs, and geopolitical tensions also mingle with the trend. If AI demand remains strong, manufacturers might prioritize higher-margin products and delay easing supply for more conventional components.
The outlook for chip suppliers could become less comfortable for customers. Hyperscalers, large enterprises, and system integrators will need to decide how much they’re willing to pay to secure capacity. Companies not directly competing in AI might also be impacted if their servers, devices, or storage systems incorporate the same types of memory and components affected by shortages.
Omdia’s forecast confirms 2026 as an exceptional year for semiconductor revenues but raises an unanswered question: whether this price increase reflects a new structural demand baseline driven by AI or if parts of the market are experiencing a temporary inflation of components that will adjust once capacity catches up with demand.
Frequently Asked Questions
How much will the semiconductor market grow in 2026 according to Omdia?
Omdia has raised its revenue growth forecast for 2026 to 62.7%, driven by demand for AI, memory, and data centers.
Which segment will see the largest revenue increase?
Data center and data storage will lead growth, with a 90% year-over-year increase and over $700 billion in revenue.
Why are DRAM and NAND prices rising so much?
Due to demand from AI servers, enterprise storage, and high-performance memory, coupled with limited supply and capacity shifts toward HBM.
Will the increase only affect data centers?
No. Omdia anticipates impacts on smartphones, consumer electronics, and wireless devices, mainly because of memory cost rises and material price increases.
via: omdia.tech.informa