Micron Technology is seeing a sharp increase in demand for memory used in artificial intelligence infrastructure, with long-term supply agreements rising to $32 billion. The company expects tight supply conditions to persist as AI data centers require increasingly large amounts of high-bandwidth memory.
Micron’s AI memory demand hits a new level
Micron Technology has reported record fiscal fourth-quarter and full-year 2026 results as demand for memory used in AI infrastructure continues to accelerate. The latest results highlight how the semiconductor industry’s growth is increasingly being driven not only by processors but also by the memory required to keep AI accelerators operating efficiently.
The company said its long-term supply agreements, known as strategic customer agreements, had reached $32 billion by the end of September. That represents a substantial increase from the $22 billion level reported in June.
The agreements are important because AI infrastructure companies are looking for greater certainty over future memory supplies. Micron said customers are making longer-term commitments because access to advanced memory has become increasingly important to their AI roadmaps.
The development comes as cloud companies and AI developers continue to expand data-center capacity around the world.
AI accelerators are driving high-bandwidth memory demand
AI accelerators, including GPUs and specialized AI processors, require large amounts of high-performance memory to process data efficiently.
High-bandwidth memory, commonly called HBM, has therefore become a critical component in modern AI systems. Unlike conventional memory, HBM is designed to provide extremely high data-transfer rates while maintaining the bandwidth required by demanding computing workloads.
Micron says its HBM3E 12-high products are being used in advanced 2026 AI accelerators and can deliver bandwidth of up to 1.2 terabytes per second.
This rising demand is changing the economics of the memory industry. AI data centers require substantially more memory per accelerator, while increasingly complex AI models require more computing and data movement.
Micron has positioned its HBM business around this shift, alongside competitors such as Samsung Electronics and SK hynix.
Long-term contracts give Micron better demand visibility
The increase in strategic customer agreements is significant because traditional semiconductor demand can fluctuate sharply between periods of shortage and oversupply.
Long-term agreements provide Micron with better visibility into future demand while giving customers greater certainty that required memory supplies will be available.
Micron said in June that it had completed 16 strategic customer agreements covering data-center, consumer and automotive markets. The company described these agreements as part of a broader change in its business model as customers increasingly seek dependable access to advanced memory.
The latest $32 billion figure indicates that this approach has expanded quickly.
For AI infrastructure companies, securing memory supply can be particularly important because a shortage of one component can delay the deployment of expensive data-center systems.
Micron forecasts tighter supply through 2028
Micron’s latest outlook suggests the current memory tightness is not expected to disappear quickly.
The company expects supply-demand conditions to remain tight through fiscal 2027 and fiscal 2028. It also expects 2027 to be a record year, supported by continued demand from AI infrastructure and other markets.
The company’s comments point to a wider issue facing the semiconductor industry. Building advanced memory capacity takes significant capital, specialized equipment and considerable time. Demand can therefore increase faster than manufacturers can add new production.
Micron has responded by accelerating investment in manufacturing capacity.
In July, the company increased its planned US investment to more than $250 billion through 2035. The spending is intended to support additional semiconductor manufacturing capacity, with Micron targeting a significant increase in US-based DRAM production.
Data centers are becoming the core growth engine
Micron’s latest financial results demonstrate how important data centers have become to its business.
The company reported record fiscal fourth-quarter revenue of $54.23 billion, while data-center demand played a major role in the increase. Reuters reported that the company’s data-center revenue jumped sharply year over year as AI servers and HBM demand accelerated.
The growth reflects a broader transformation in the semiconductor industry.
AI models require large clusters of accelerators operating together. Those accelerators need high-speed memory to store and move the enormous volumes of data involved in training and inference.
As AI companies build larger computing systems, the memory attached to those systems becomes an increasingly important part of the overall infrastructure investment.
That has helped move memory from being viewed primarily as a supporting component to becoming a strategic part of AI system design.
Micron is expanding its technology portfolio
Micron’s AI opportunity extends beyond HBM.
The company recently announced its 512GB DDR5 high-capacity memory module for next-generation servers. The module supports speeds of up to 9,200 MT/s and is designed for demanding workloads including AI, analytics and in-memory databases. AMD and Intel are among the companies validating the technology.
Micron also said its high-performance PCIe Gen 5 and Gen 6 SSD products are shipping to major customers for applications involving AI key-value cache, or KV-cache, workloads.
This matters because AI infrastructure increasingly requires multiple layers of memory and storage.
The market is therefore expanding beyond the traditional GPU and accelerator race. Memory bandwidth, capacity, storage performance and energy efficiency are becoming important factors in determining how efficiently AI systems can operate.
Competition remains intense across the memory market
Micron is not the only company benefiting from AI-driven memory demand.
Samsung Electronics and SK hynix remain major competitors in the global memory industry. The three companies are competing for a growing market created by AI data centers and increasingly sophisticated accelerators.
The pressure is particularly visible in HBM, where limited production capacity has created supply constraints.
Recent reporting has also highlighted the effect of the HBM shortage on Chinese AI chipmakers. Reuters reported in September that companies including Huawei and Cambricon had increased prices for AI processors as higher HBM costs affected production economics.
That illustrates how memory availability can influence the broader AI chip supply chain, rather than affecting memory manufacturers alone.
AI infrastructure spending keeps expanding
Micron’s growing customer commitments come against a much larger wave of AI infrastructure spending.
AI companies are signing increasingly large agreements to secure computing capacity and semiconductor supplies. Anthropic, for example, disclosed plans involving hundreds of billions of dollars in long-term AI infrastructure commitments in an IPO filing reported by Reuters.
Other technology companies are also developing custom AI chips and expanding data-center capacity.
The result is increased demand across the semiconductor supply chain, from processors and memory to networking equipment, storage and power infrastructure.
For Micron, this creates an opportunity to lock in demand before new memory capacity comes online.
Supply constraints remain the key market risk
The strong demand outlook also creates risks.
If AI infrastructure spending continues to accelerate faster than semiconductor manufacturers can expand capacity, memory shortages could persist and prices could remain elevated.
At the same time, semiconductor companies must invest heavily to meet future demand without creating excess capacity if AI spending eventually slows.
Micron has responded with major investments in the United States and additional technology development, while also expanding long-term relationships with major customers. The company expects its investment programme and capacity expansion to support demand over the coming years.
The latest results therefore provide another indication that memory has become one of the central components of the global AI infrastructure boom.
Key Takeaways
- Micron’s long-term supply agreements increased from $22 billion in June to $32 billion by September.
- AI accelerators are driving strong demand for high-bandwidth memory used in advanced data centers.
- Micron expects tight memory supply-demand conditions to continue through fiscal 2027 and 2028.
- The company has raised planned US semiconductor investment to more than $250 billion through 2035.
FAQs
Why is AI increasing demand for memory chips?
AI accelerators process very large datasets and require high-speed memory to move data efficiently. As AI models and data centers become larger, the amount of memory required per computing system also increases.
What is HBM and why is it important for AI?
High-bandwidth memory, or HBM, is a type of advanced memory designed to provide very high data-transfer rates. It is used alongside AI accelerators because these processors need rapid access to large amounts of data.
How large are Micron’s AI-related supply agreements?
Micron said its long-term supply agreements had reached $32 billion by September 2026, compared with $22 billion in June. The agreements provide greater visibility into future demand and supply requirements.
Is Micron expanding production capacity?
Yes. Micron has announced more than $250 billion in planned US investment through 2035 and is expanding manufacturing capacity to meet expected long-term demand for memory, particularly as AI infrastructure grows.
