AI infrastructure spending is drawing fresh market attention after Micron and SK Hynix reached trillion-dollar valuations. Anthropic also raised $65 billion, adding another signal that large AI firms need more computing power. Meanwhile, memory-chip demand has moved into focus as data centers require faster systems for AI training and deployment.
AI infrastructure spending is rising as cloud service providers expand data centers for advanced computing. TrendForce now expects the world’s top nine cloud service providers to spend about $830 billion in 2026. That forecast marks a 79% annual increase, above its earlier estimate of 61%.
The spending covers major companies including Google, Amazon Web Services, Meta, Microsoft, Oracle, ByteDance, Tencent, Alibaba, and Baidu. Much of the budget targets GPU clusters and custom AI processors. These systems need high-bandwidth memory, known as HBM, to move large datasets quickly.
Micron Technology supplies DRAM, NAND, and HBM products used in data centers, servers, and AI systems. As AI chips carry more memory, demand for advanced memory products has increased. NVIDIA’s latest Vera Rubin GPU also points to higher memory use, with more HBM capacity than older AI chips.
Counterpoint Research expects demand for HBM in custom AI processors to rise sharply from 2024 to 2028. Moreover, HBM production requires more wafer capacity than standard memory chips. That supply need has created tighter industry conditions across the memory market.
Micron recently crossed a $1 trillion market valuation, joining SK Hynix and Samsung in a group of large memory-chip companies benefiting from AI demand. Reuters reported this week that Micron and SK Hynix reached trillion-dollar valuations for the first time.
The rally also followed stronger analyst attention. UBS raised its Micron price target to $1,650, citing stronger long-term supply agreements, better pricing visibility, and higher demand visibility across the memory industry. However, analysts also track whether earnings growth can support the sharp share-price move.
Micron’s previous reported quarter showed strong growth in revenue and profit. Revenue nearly tripled to $23.8 billion, while net income rose to $13.8 billion. Operating margin also reached 67.6%, helped by tight supply and higher demand for memory products.
Management has also pointed to stronger market conditions. At a JPMorgan Chase investor conference, Micron executive Manish Bhatia said the company’s outlook had strengthened since its last earnings report. He described the supply-demand gap as “structural, rather than cyclical,” while noting tightness across DRAM, HBM, and NAND.
Bhatia also said some major customers can fill only about 60% of their memory needs. That comment showed how supply constraints remain a key factor for Micron before its fiscal third-quarter earnings report scheduled for June 24.
Anthropic confirmed that it raised $65 billion in Series H funding at a $965 billion post-money valuation. The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. The funding places Anthropic near a $1 trillion valuation as competition across AI models grows.
The fundraising adds to the broader demand for computing infrastructure. AI firms need large amounts of chips, memory, networking equipment, and data center capacity to train and run advanced models. As a result, memory suppliers remain closely tied to the pace of AI infrastructure spending.
NVIDIA has also pointed to tighter memory conditions. Its finance chief, Colette Kress, said the company saw memory prices rising early and “ordered a long time ago.” She also said Nvidia works with suppliers on “what to build,” showing closer planning between AI chip companies and memory producers.
Nevertheless, memory remains a cyclical market. Prices can fall when supply catches up or when customer inventory rises. For now, investors are watching Micron’s June 24 earnings report, customer demand, HBM supply, and AI data center spending to measure whether the current memory cycle has more room to run.
Also Read: Windows 11 Overhaul: Why is Microsoft Replacing Decades-Old Dialog Boxes?