Author: Yutaka Tanoue
The semiconductor industry, particularly the memory market, is currently undergoing an extraordinary expansion that can only be described as "explosive growth." Terms like "booming" fall short of adequately describing this phenomenon. A mere glance at the charts would stun anyone with its sheer magnitude.
Figure 1 shows the three-month moving average shipment value for various semiconductor categories, based on World Semiconductor Trade Statistics (WSTS) data from 1991 to May 2026. Over the past three decades, despite economic fluctuations, the four major categories—micro components, memory, logic, and analog—have generally shown a steady upward trend. This represents the "normal development path" of the semiconductor industry.
However, since around 2024, logic circuits, and more recently memory, have shown nearly vertical growth, completely overturning conventional understanding. The growth in memory, in particular, has been astounding, almost breaking through the chart's peak. This growth is so rapid that the growth history accumulated over the past 30 years seems like mere warm-up.
Based on my years of closely observing the semiconductor market, I can confidently say that the memory market has never experienced such a steep rise. This is a historic anomaly shaking the very foundations of the semiconductor industry.
A roughly tenfold increase in about a decade, with a staggering annual growth rate of 285%, is mind-boggling. Figure 2 highlights this explosive growth, focusing on the period from 2016 onward. Monthly MOS memory shipments in 2016 were about $5.6 billion. After a brief boom during the 2017-2018 "memory bubble," shipments plunged to just $5.8 billion by the early 2023 recession. This marked the lowest point of the semiconductor industry's downturn.
However, the recovery since then, or more accurately, the surge. As of May 2026, monthly memory shipments have skyrocketed to $63.3 billion. Compared to 2016 levels, this is more than an 11-fold increase in just a decade. Compared to the early 2023 trough, it's roughly a 10.7-fold increase in just a few years.
Similarly, Figure 2 shows that the logic business, which includes NVIDIA GPUs, also grew from $13.3 billion to $31.6 billion, clearly showing the "NVIDIA GPU effect" and "AI effect." However, even logic business growth appears relatively modest compared to the explosive growth of the memory business.
The anomaly becomes even more apparent when looking at growth rates. Figure 3 shows the year-on-year growth rate for the same data. The recent year-on-year growth rate for memory has reached a record high of 285%.
For reference, even at the peak of the previous "memory bubble" (around 2017), the annual growth rate was about 60%. In comparison, you can see how staggering the current growth rate is. Looking at the other three categories in Figure 3, even logic circuits are only around 40%, while microcontrollers and analog circuits are at 14%-19%. In other words, only memory growth has truly reached "another level."
DRAM and NAND Are Growing Explosively
Looking at the memory market, the main drivers of its explosive growth are the two major memory products: DRAM and NAND flash memory (hereafter NAND). Figure 4 shows the global DRAM and NAND market quarterly overview based on TrendForce data.
In early 2023, at the height of the semiconductor industry recession, DRAM prices hit rock bottom at just $9.7 billion, and NAND prices also fell to $8.7 billion. It was a dark period where all memory manufacturers incurred losses and were forced to cut production.
However, by Q2 2026, the DRAM market size is expected to jump to $145 billion, and the NAND market to $81 billion. Compared to the early 2023 lows, the DRAM market has grown about 15-fold, and the NAND market about 9-fold. Adding them together, the quarterly total would reach $226 billion, with an annual total exceeding $900 billion—a scale that seems almost unbelievable at first glance.
What was once considered a "cheap standard" or "commodity" product—memory—is now poised to seize the semiconductor industry's leadership from microcomputers (like logic and MPUs).
The Reason Is the Abnormal Surge in Memory and Wafer Prices
Why has the memory market expanded to such a scale? The key point is that this explosive growth is not solely due to increased shipments; the biggest factor is the abnormal surge in memory prices themselves.
Figure 5 shows the trends in DRAM (DDR5 16Gb 2Gx8) spot prices and NAND (1Tb TLC wafer) wafer prices. In early 2025, the DRAM spot price was only $4.70. Recently, the price has skyrocketed to $46.00, an increase of about 10 times. NAND wafer prices also jumped from $2.40 to $25.00, also an increase of about 10 times.
In other words, the tenfold growth in market size is not primarily due to a tenfold increase in shipments, but because the "unit price" has surged nearly tenfold. Even if memory sales remained unchanged, revenue would grow tenfold. This is the mechanism behind the explosive growth of the memory market size.
For memory manufacturers, this is an ideal environment. After all, prices automatically rise without requiring significant capital investment, thereby significantly increasing profit margins. As will be discussed later, the soaring stock prices of memory manufacturers are also due to this rapid profit growth fueled by abnormal price increases.
Reasons for the Price Surge
So why are memory prices soaring? It's because demand far exceeds supply, but tracing the source of this demand reveals that it is thanks to the astonishingly huge investments made by hyperscale data center operators.
Figure 6 shows the capital expenditure trends of the four largest hyperscale data center operators (Amazon, Google, Microsoft, and Meta). In 2015, the combined capital expenditure of these four companies was only $21 billion. Even then, it was considered a "huge amount."
However, since OpenAI released ChatGPT in November 2022, a wave of generative AI fever has emerged, and the capital investment curve has also risen sharply. Total investment by these four companies is projected to reach $355 billion in 2025 and an astounding $755 billion in 2026. This represents an unprecedented investment boom, with investment growing about 36-fold in just over a decade since 2015.
$755 billion far exceeds 120 trillion yen. Just four companies plan to spend an amount equivalent to Japan's national budget (general account) in a single year on data centers and AI infrastructure. This shows how extraordinary the situation is.
AI Data Centers Are Like "Black Holes"
The ultimate destination of this massive investment is AI data centers. Figure 7 illustrates the current situation schematically.
As hyperscale data centers race to invest in AI data centers, the semiconductors essential for AI learning and inference—namely GPUs from companies like NVIDIA, the High Bandwidth Memory (HBM) used in GPUs, and large-capacity storage SSDs equipped with NAND flash—are being sucked into the "black hole" of AI data centers one after another.
Memory manufacturers prioritize producing high-margin HBM and high-performance DRAM and NAND flash for data centers. This is a natural business decision. Consequently, production line capacity is shifted to the AI domain, leading to a drastic reduction in capacity available for other applications.
The most severely impacted sector is DRAM and NAND flash for digital consumer electronics like PCs, smartphones, and game consoles. Memory for these products is in severe shortage, to the point of being "utterly insufficient."
With limited supply capacity, if demand is concentrated only on AI data centers, memory for consumer electronics will inevitably run out. This leads to fierce competition for the limited supply, driving up prices. This is directly linked to the abnormal surge in memory prices mentioned earlier.
Indeed, PC and smartphone manufacturers have begun sending distress signals, stating they cannot obtain the needed memory, and procurement costs are soaring, forcing them to pass these costs onto product prices. Ironically, in the shadow of the AI boom, the digital devices we use daily are becoming increasingly expensive and even scarce.
Semiconductor Market Forecasts "Completely Wrong"
In 2023, I made a forecast of the global semiconductor market up to 2032, as shown in Figure 8. By analyzing the history of the semiconductor industry and incorporating the "PC effect," "Internet effect," "smartphone effect," and the upcoming "AI semiconductor effect," I predicted the semiconductor market would grow at a rate of roughly "doubling every decade."
This forecast projected the market size to reach about $1.2 trillion by 2032. At the time, I considered this a rather optimistic forecast.
However, this forecast has proven completely wrong. In fact, it was wrong because it was far too conservative. Please refer to Figure 9. According to the WSTS (World Semiconductor Trade Statistics) Spring 2026 forecast, the global semiconductor market size will reach $630.5 billion in 2024, $795.6 billion in 2025, and break through $1.5 trillion in 2026 to reach $1.5112 trillion. Furthermore, it is expected to grow rapidly to $1.9137 trillion in 2027, approaching $2 trillion.
Within just a few years, the market size has easily surpassed the $1.2 trillion level the author previously predicted for 2032. The 2032 forecast was achieved in 2026. This does not mean the earlier forecast was "too optimistic," but rather that the disruptive power of the AI frenzy has completely overturned all previous conventions of the semiconductor industry.
Figure 10 clearly shows that this rapid growth is driven by memory (including DRAM and NAND) and logic circuits (including GPUs). Memory is expected to break through $1 trillion by 2027, reflecting the explosive growth of DRAM+NAND mentioned earlier. The logic circuit market is also projected to exceed $500 billion.
On the other hand, analog circuits and micro semiconductors are almost flat. In other words, the overall semiconductor market growth is not balanced; instead, only these two AI-related fields are growing at an unprecedented pace, pulling the entire market up. This is an extremely distorted growth structure.
How Long Will This Boom Last?
Now, let's address the question everyone is most concerned about: "How long will the AI boom and the growth it brings last?"
The key to understanding this lies in the long history of the memory market. Figure 11 shows the annual growth rate of the memory market over nearly 35 years since 1991. This chart reflects the countless ups and downs the semiconductor industry has experienced.
From the boom brought by Windows 95 in 1995, to the 2000 IT bubble and its subsequent burst, the 2017-2018 memory bubble, the 2008 Lehman Brothers collapse, and the COVID-19 pandemic boom from 2020 onward, the semiconductor industry has experienced a rollercoaster of booms and busts.
The point to note here is not the height of individual peaks, but the "duration" that positive growth persists. Analyzing Figure 11 carefully reveals a key fact: in any historical period, the longest period the memory market has maintained an average annual positive growth is only five years. Over the past 35 years, there has never been a case of consecutive positive growth for six or seven years.
Why does the market stabilize after five years? Because the memory market is inherently subject to the "silicon cycle," which includes a surge in demand, price increases, companies increasing production investment, oversupply, and price crashes.
If the economy continues to boom, companies inevitably rush to invest in equipment, eventually leading to oversupply and price crashes. This is an inevitable mechanism determined by the nature of memory as a product.
The current AI boom is expected to drive significant memory market growth starting from the 2023 trough and peaking in 2024. Following historical patterns, this boom should end by 2028 at the latest, or possibly peak as early as 2027.
While inevitably some will say, "this time is different" or "AI is special," we must not forget that the same was said during every past bubble economy.
The Higher the Peak, the Deeper the Valley
The history of the semiconductor industry also teaches us another undeniable rule: "The higher the peak, the deeper the subsequent trough."
If you look at Figure 11 again, this rule becomes evident. The IT bubble peaked in 2000 with over 50% annual growth, but was immediately followed by a plunge to -49.5% in 2001. Similarly, the 2017-2018 memory bubble experienced a peak of over 60%, followed by a sharp 33% drop in 2019. In this industry, the higher the boom's peak, the more severe and prolonged the subsequent recession tends to be.
Let's examine the current AI boom again. A 285% annual growth rate is an unprecedented peak, far exceeding any previous bubble. Its rise is so high that the peaks of the IT bubble and memory bubble seem negligible in comparison.
Applying this principle, the conclusion is obvious. Following this unprecedented boom, the ensuing "trough" is likely to be deeper and more severe than any the semiconductor industry has experienced before. We should prepare for an extremely severe recession expected to start around 2027-2028, one that will not be overcome half-heartedly.
Memory Manufacturer Stock Price Surge and "Billion-Yen Investors"
Currently, memory manufacturers' stock prices are soaring across the board. This is entirely expected, as rising prices lead to substantial profit growth for them.
A symbolic example is Kioxia. The company's rising stock price has resulted in 600 investors, so to speak "billion-yen investors," gaining unrealized profits exceeding 1 billion yen from their holdings. The entire industry is immersed in an unprecedented celebratory atmosphere.
However, we need to stay calm now. Stock prices are like a mirror, reflecting expectations for the future. When these expectations become excessive, if reality falls even slightly short, stock prices can crash instantly.
The higher the peak, the deeper the valley. This applies to stock prices just as it does to market size. In every past bubble burst, we've seen scenes where billionaires who became wealthy at the peak of the bubble had to drastically reduce their assets as the recession arrived.
Now, during the boom, is precisely the time to make practical preparations for the recession. This prosperity will not last forever. History shows the memory market's positive growth lasts at most five years, followed by a downturn (at least on the surface). The higher the peak, the deeper the trough. Therefore, I strongly urge companies currently enjoying the boom: "While the boom continues, you should make practical preparations for the recession."
