Artificial intelligence companies and consumer electronics manufacturers are locked in a sudden battle for memory chips as global shortages push prices to extreme levels and threaten to derail the technology boom.
Key Takeaways:
- Memory chip prices have more than doubled since February in some segments, with suppliers running critically low on inventory
- Tech giants Microsoft, Google, and ByteDance are rushing to secure supplies while smartphone makers warn of 20-30% price increases
- The shortage affects everything from USB drives to advanced AI data center chips, with experts warning of macroeconomic consequences
Japanese electronics retailers are now rationing hard-disk drives. Chinese smartphone manufacturers are preparing customers for price hikes. Meanwhile, major technology companies are desperately courting memory-chip manufacturers Micron, Samsung Electronics, and SK Hynix for guaranteed allocations, three industry insiders report.
The crisis spans every memory category—from basic flash chips in USB drives and smartphones to sophisticated high-bandwidth memory that powers AI processors in data centers. Market research firm TrendForce reports prices in certain segments have soared by over 100% since February, attracting speculators who believe the rally will continue.
The ripple effects extend far beyond Silicon Valley. Economists and industry leaders warn that prolonged shortages could hamper AI-driven productivity improvements and postpone hundreds of billions in digital infrastructure investments. This adds unwelcome inflationary pressure as governments struggle to control rising costs and manage U.S. tariff policies.
“The memory shortage has now graduated from a component-level concern to a macroeconomic risk,” said Sanchit Vir Gogia, CEO of Greyhound Research, a technology advisory firm. The AI build-out “is colliding with a supply chain that cannot meet its physical requirements.”
A Reuters investigation involving nearly 40 interviews—including 17 chipmaker and distributor executives—reveals how the industry’s race to satisfy demand for advanced semiconductors has created a paradox. Chipmakers cannot manufacture sufficient high-end semiconductors for AI applications, yet their pivot away from traditional memory products is strangling supply to smartphones, personal computers, and consumer electronics. Several companies are now rushing to reverse course.
Tech firms are engaged in a global scramble, with price increases hitting electronics retailers and component suppliers across China and Japan. Dynamic random-access memory (DRAM) suppliers saw average inventory plummet from 13-17 weeks in late 2024 to just two-four weeks by October, according to TrendForce data. July inventories had stood at three-eight weeks.
This squeeze arrives as investors debate whether massive AI infrastructure spending has created a bubble. Analysts predict only the largest, financially robust companies will survive the price escalation.
One memory-chip executive mentioned that the shortage will postpone future data-center construction. New production facilities require at least two years to build, but manufacturers hesitate to overbuild, fearing idle capacity if demand suddenly collapses.
Samsung and SK Hynix have announced new capacity investments without specifying how production will divide between HBM and conventional memory. SK Hynix has informed analysts the memory deficit will persist through late 2027, Citi noted in November.
“These days, we’re receiving requests for memory supplies from so many companies that we’re worried about how we’ll be able to handle all of them. If we fail to supply them, they could face a situation where they can’t do business at all,” Chey Tae-won, chairman of SK Hynix parent SK Group, said at an industry forum in Seoul last month.
OpenAI signed preliminary agreements with Samsung and SK Hynix in October for its Stargate project, which demands up to 900,000 wafers monthly by 2029—approximately double current global HBM production, Chey revealed.
Samsung confirmed it monitors market conditions but declined to discuss pricing or customer relationships. SK Hynix stated it’s expanding production capacity to address increased memory demand.
Desperation Drives Negotiations
After ChatGPT’s November 2022 launch sparked the generative AI explosion, a worldwide rush to construct AI data centers prompted memory manufacturers to redirect production toward HBM, used in Nvidia’s high-performance AI processors.
Pressure from Chinese competitors manufacturing lower-tier DRAM, particularly ChangXin Memory Technologies, accelerated Samsung and SK Hynix’s transition to higher-profit products. The South Korean companies control two-thirds of the DRAM market.
Samsung notified customers in May 2024 of plans to cease production of one DDR4 chip variant—an older generation used in PCs and servers—by year-end. The company has since reversed this decision and will continue production, two sources confirmed. Micron announced in June it would stop shipping DDR4 and LPDDR4—a smartphone variant—within six to nine months.
ChangXin discontinued most DDR4 production as well, one source said.
This transition, however, coincided with a traditional data center and PC replacement cycle, plus unexpectedly strong smartphone sales—all dependent on conventional chips.
Looking back, “one could say the industry was caught off-guard,” said Dan Hutcheson, senior research fellow at TechInsights.
Samsung increased server memory chip prices by up to 60% last month. Nvidia CEO Jensen Huang, who announced partnerships and shared fried chicken with Samsung Electronics Chairman Jay Y. Lee during an October visit to South Korea, acknowledged the price surge as significant while noting Nvidia had secured substantial supply.
Google, Amazon, Microsoft, and Meta requested open-ended orders from Micron in October, telling the manufacturer they’ll accept whatever quantities it can deliver regardless of price, two people briefed on discussions said.
China’s Alibaba, ByteDance, and Tencent are pressing suppliers hard, sending executives to Samsung and SK Hynix in October and November to advocate for allocation.
“Everyone is begging for supply,” one said.
The Chinese companies didn’t respond to questions about the chip crunch. Nvidia, Meta, Amazon, and OpenAI didn’t reply to comment requests.
SK Hynix announced in October that all chips are sold out for 2026, while Samsung secured customers for next year’s HBM chip production. Both companies are expanding capacity for AI demand, but new conventional chip facilities won’t launch until 2027 or 2028.
Micron, Samsung, and SK Hynix shares have climbed this year on chip demand. Micron forecast first-quarter revenue above market expectations in September, while Samsung reported its largest quarterly profit in over three years in October.
Consultancy Counterpoint Research expects advanced and legacy memory prices to climb 30% through the fourth quarter and potentially another 20% in early 2026.
Smartphone Prices Set to Jump
Chinese smartphone makers Xiaomi and Realme have warned of potential price increases.
Francis Wong, Realme India’s chief marketing officer, told the dramatic memory cost increases were “unprecedented since the advent of smartphones” and could force handset price increases of 20% to 30% by June.
“Some manufacturers might save costs on imaging cameras, some on processors, and some on batteries,” he said. “But the cost of storage is something all manufacturers must completely absorb; there’s no way to transfer it.”
Xiaomi says it would counter higher memory costs through price increases and selling more premium phones, with other business units helping absorb the impact.
Taiwanese laptop manufacturer ASUS stated in November it maintains approximately four months of inventory, including memory components, and will adjust pricing as necessary.
Winbond, a Taiwanese chipmaker controlling roughly 1% of the DRAM market, was among the first to announce capacity expansion. Its board approved a plan in October to dramatically increase capital expenditure to $1.1 billion.
“Many customers have been coming to us saying, ‘I really need your help,’ and one even asked for a six-year long-term agreement,” Winbond’s President Pei-Ming Chen said.
Speculators Enter the Market
In Tokyo’s Akihabara electronics district, stores are limiting memory product purchases to prevent hoarding. A sign at PC shop Ark states that since November 1, customers can buy a maximum of eight items total across hard-disk drives, solid-state drives, and system memory.
Employees at five shops reported shortages have driven prices sharply upward in recent weeks. Some stores show one-third of products sold out.
Products like 32-gigabyte DDR5 memory—favored by gamers—now cost over 47,000 yen, up from approximately 17,000 yen in mid-October. Premium 128-gigabyte kits have more than doubled to around 180,000 yen.
Price increases are pushing customers toward the secondhand market—benefiting entrepreneurs like Roman Yamashita, owner of iCON in Akihabara, who reports his used PC parts business is thriving.
Eva Wu, a sales manager at component trader Polaris Mobility in Shenzhen, said prices fluctuate so rapidly that distributors now issue broker-style quotes expiring daily—sometimes hourly—versus monthly before the crisis.
A Beijing DDR4 seller stockpiled 20,000 units anticipating further increases.
Some 6,000 miles away in California, Paul Coronado said monthly sales at Caramon, his company selling recycled low-end memory chips from decommissioned data-center servers, have surged since September. Nearly all products now go to Hong Kong-based intermediaries reselling to Chinese clients.
“We were doing about $500,000 a month,” he said. “Now it’s $800,000 to $900,000.”
Written by Alius Noreika
