The discussion surrounding the artificial intelligence (AI) bubble continues to escalate, with analysts warning investors to proceed with caution. At the same time, the surge in AI-related chip demand has led to a unique phenomenon called “Chipflation,” with research institutions predicting that this will impact a broader range of economic sectors and may become a pivotal turning point this year.
According to a 66-page report by Morgan Stanley cited by Reuters on June 3, the soaring prices of memory chips, driven by the high demand for AI, may trigger Chipflation, forcing device manufacturers to choose between raising prices and accepting lower profit margins.
The report states that the initial impact was on AI infrastructure, but now it is spreading to hardware profits, device prices, cloud costs, currency inflation, and policy aspects, making this dilemma a macroeconomic issue.
Unlike previous cycles of prosperity and recession, Morgan Stanley suggests that the current surge in demand may represent a “lasting supply-demand reset.” While the direct impact on ordinary consumers may be limited, pressures are emerging in producer prices, corporate profit margins, cloud costs, capital expenditures, and delays in the introduction of new technologies.
Over the past year, memory chip prices have skyrocketed sixfold as manufacturers struggle to keep up with the hefty spending by major tech companies on AI infrastructure, prioritizing chips for high-profit data centers over those used in daily devices.
Memory chips are essential components for AI operations. According to data released by market research firm DRAMeXchange on May 29, the average price of PC DDR4 8Gb memory reached $20, a 25% increase from April and the highest value tracked by the institution since June 2016.
Global consulting and advisory firm IDC predicts that in 2026, the personal computer and smartphone markets will shrink due to the rising chip prices, hindering potential buyers, especially in the low-end segment.
There has been a frenzy of investments in the AI sector worldwide, with data showing that if investors hold S&P 500 index funds, over a third of the funds are concentrated in the stocks of the seven largest US tech companies. This concentration is even more pronounced if one holds Nasdaq 100 index funds, reaching around 40%.
The excessive focus on a few large tech stocks has prompted some market experts to warn of potential bubble formation. Speaking at an event on May 27, Director Hong Hao of the “China Chief Economists Forum” indicated that investments in chips have reached a peak, approaching historical highs. Each day of price increase brings the market closer to full-blown bubble territory, similar to the dot-com bubble of 2000. “This year, with the Red Horse and the Red Sheep, especially at the peak of the cycle, is the most suitable time for a dizzying market bubble to form.”
Political analyst Xia Yan believes that instances of investment bubbles are not uncommon in history, from the Dutch tulip mania in the 17th century to the dot-com bubble of 2000. Although significant capital is flowing into the AI sector, a substantial portion of these investments has yet to yield tangible or visible business returns. For instance, Meta (parent company of Facebook) released its first-quarter financial report this year, disclosing heavy investments in AI computing capabilities, data centers, and talent recruitment, yet their stock price fell, reflecting investor concerns about the substantial funds poured into AI.
A study by the UK-based Man Group indicates that the AI bubble is real and rapidly expanding. From railways and electricity to the internet bubble, technology itself endures, but its financing cycles often rupture as market expectations surpass the industry’s actual supply capacity. Therefore, whether the AI bubble will burst is a matter of time.
Renowned Chinese economist Fu Peng stated at the “2026 Phoenix Bay Financial Forum Financial Summit” that the next 18 months are a critical observation period. If AI can truly penetrate various industries, transform into actual productivity, fundamentally improve production relations, and drive overall economic growth, global distribution conflicts will significantly ease. On the other hand, if AI fails to boost the real economy, the world will face not just financial crises but deeper international institutional and order disturbances.
“There are already signs of such global asset performance this year,” Fu Peng continued, “If the gold cycle ends, and artificial intelligence spreads across all sectors, this year could be a crucial turning point. Once this turning point happens, we will enter a 10-15 year period of reconstructing production relations.”