[Summary]
Fueled by the AI boom, the semiconductor sector continues to experience unprecedented market prices.
Based on WSTS data, SIA explains that global semiconductor sales will reach a record high of $791.7 billion in 2025, and will reach approximately $1 trillion in 2026.
On the other hand, among investors,
Will the AI semiconductor bubble burst in 2026?
There is also a growing sense of caution.
The reason is simple.
This is because the growth of the semiconductor market is highly dependent on AI capital investment by cloud giants such as Microsoft, Amazon, Alphabet, Meta, and Oracle.
TrendForce predicts that the 2026 CAPEX of eight major companies, including Google, AWS, Meta, Microsoft, Oracle, and major Chinese CSPs, will exceed $600 billion.
If this huge investment continues, the AI semiconductor market will continue to rise.
However, if there is even the slightest slowdown in investment or doubts about AI monetization, semiconductor stocks will quickly move into a selective market.
In this article, we will discuss the basis for the theory that the semiconductor bubble will burst in 2026, how it differs from the dot-com bubble, the impact it will have on Japanese semiconductor-related stocks, and the leading indicators that investors should look at.
Why is there a theory that the semiconductor bubble will burst?
Semiconductor market conditions appear to be extremely strong.
GPUs for AI servers, HBM, advanced packaging, and network semiconductors for data centers remain in high demand.
According to SIA, global semiconductor sales in the first quarter of 2026 were $298.5 billion, up 25% from the fourth quarter of 2025.
Still, the collapse theory is emerging because there are three distortions in the foundation of growth.
Risk 1: CAPEX limit theory for hyperscalers
The biggest fuel supporting the current semiconductor market is AI capital investment by large tech companies.
Microsoft, Amazon, Alphabet, Meta, Oracle and others are investing billions in AI data centers, GPUs, networking, and power and cooling infrastructure.
TrendForce predicts that major CSPs' 2026 CAPEX will exceed $600 billion.
While this is a strong tailwind for the AI hardware ecosystem, it is also a major cause for concern for investors.
What the market sees is
With this much investment, will monetization of AI really catch up?
That is one point.
If a giant tech company were to make the following statements in the future, the assumptions for semiconductor stocks would be shaken significantly.
- It takes time to recover AI investment
- Postpone some data center investments
- Level out the pace of GPU procurement
- Start of operation delayed due to power and construction constraints
- Depreciation burden puts pressure on profits
In other words, the biggest risk to the semiconductor market in 2026 lies not in the semiconductor companies, but in the investment stance of the hyperscalers, who are the buyers.
Risk 2: Oversupply and inventory cycle reversal
The semiconductor industry has long had a silicon cycle.
When demand is strong, companies increase investment all at once.
However, by the time fabs and equipment start operating, demand may have subsided and there may be an oversupply.
The same structural risks are present this time as well.
GPUs and HBM for AI are strong, but as each company makes large-scale investments in 2025-2026, supply capacity may increase rapidly from 2027 onwards.
SEMI predicts that global semiconductor manufacturing equipment sales will grow to $156 billion in 2027.
In Japan, based on SEAJ forecasts, it is predicted that sales of Japanese-made semiconductor and FPD manufacturing equipment will reach the 5 trillion yen level in fiscal 2025, and reach the 6 trillion yen level in 2027.
This is a tailwind for equipment companies, but from an investor's perspective,
Stock prices fall first the moment capital investment peaks out.
It is necessary to pay attention to this point.
Semiconductor equipment stocks tend to sell off significantly when the outlook for orders slows down, rather than when business performance is good.
Risk 3: Weak demand for semiconductors other than AI
In the current semiconductor market, only AI-related areas are particularly strong.
On the other hand, demand recovery has been patchy for some products for smartphones, PCs, general-purpose memory, consumer electronics, and automobiles.
This state is
AI is not enough, but other than AI is not strong
This is a polarization.
Rather than the semiconductor market as a whole experiencing healthy growth, funding and supply are being concentrated towards AI data centers.
Therefore, if investment in AI slows down even a little, the valuation of the entire sector is likely to plummet.
Similarities to the dotcom bubble
The current AI semiconductor market is often compared to the dot-com bubble around 2000.
At that time, with the spread of the Internet, funds were concentrated in communications infrastructure, server, and network equipment companies.
Currently, with the spread of generative AI and agent AI, funds are concentrated on GPUs, HBM, data centers, and network semiconductors.
There are three similarities:
| Similarities | Contents |
|---|---|
| Infrastructure ahead | Capital investment rapidly expands before profitability |
| Concentration on representative stocks | Cisco at the time, now NVIDIA, etc. |
| IPO/Fund Raising Fever | Huge amount of money flowing into AI startups and infrastructure companies |
Investors are especially wary of the fact that infrastructure investment alone is becoming huge, even before the final profitability of AI services.
Definitive difference from the dot-com bubble
However, it is dangerous to simply equate the current semiconductor market with the dot-com bubble.
The key difference is that major companies have actual sales and profits.
Companies such as NVIDIA, TSMC, SK Hynix, Broadcom, AMD, Tokyo Electron, Advantest, and Disco are actually increasing their sales and profits as a result of the wave of AI investment.
During the dot-com bubble, there were many cases in which loss-making companies were bought solely on the basis of ``number of accesses'' or ``future potential.''
On the other hand, current AI semiconductor companies have real cash flow and high profit margins, at least for the top players.
Therefore, the problem for 2026 is
Is it a bubble without substance?
rather than
To what extent have stock prices gone too far in anticipation of real growth?
It is.
Impact on Japanese semiconductor-related stocks
Japanese companies are no longer as large as they once were in the global market share of finished chips.
However, in terms of manufacturing equipment, inspection equipment, materials, post-processing, silicon wafers, and chemicals, it holds important parts of the global semiconductor supply chain.
Therefore, if the AI semiconductor market collapses, it will have a major impact on Japanese stocks.
Impact on Tokyo Electron Advantest
Tokyo Electron (8035) and Advantest (6857) are representative stocks that have strongly benefited from AI semiconductor investment.
Tokyo Electron has an important position in front-end equipment, and Advantest has an important position in semiconductor testers.
In a bull market, the following factors tend to buy stocks:
- Expansion of cutting-edge investment by TSMC and Samsung
- AI chip demand for NVIDIA/AMD
- HBM related investment *Advanced packaging investment
- Increase in demand for testers
On the other hand, in a bear market, stock prices can drop significantly simply due to concerns that orders will peak out.
This is because even if business performance is still strong, the market will start pricing in the next slowdown.
Impact on Disco Ibiden Lasertech
Companies like DISCO, IBIDEN, and Lasertec are closely connected to the advancement of AI semiconductors.
| Brand | Related areas | Features |
|---|---|---|
| DISCO | Cutting, grinding, polishing | Important for HBM and advanced packages |
| IBIDEN | IC package substrate | Linked with demand for AI servers |
| Lasertech | EUV inspection equipment | Reflecting expectations of advanced process investment |
These companies have higher exposure to AI/advanced semiconductors than general-purpose semiconductors.
Therefore, medium- to long-term technological trends are strong.
However, stock prices tend to incorporate very high expectations, and they tend to be pushed lower when sentiment in the semiconductor sector as a whole deteriorates.
Two major scenarios in the second half of 2026
Scenario A: Soft landing
In a soft landing, AI investment will remain at a high level, but overheated stock prices will cool down.
In this case, semiconductor stocks could rise again toward 2027, with an adjustment of around 5% to 15%.
The conditions are as follows.
*Big Tech's CAPEX is maintained *Monetization of AI services continues
- Tight supply and demand for HBM and advanced packages
- Strong guidance from NVIDIA and TSMC
- Interest rate increases are limited
In this scenario, rather than the semiconductor bubble bursting,
Healthy trend in overheating market
It can be understood as
Scenario B: Hard landing
In a hard landing, the very premise of AI investment will collapse.
For example, the following situation exists.
- Big Tech lowers CAPEX
- Delay in AI monetization revealed in financial results
- Valuation compressed due to high interest rates *GPU operation is delayed due to data center construction and power constraints
- Memory prices and GPU demand seem to be peaking out
In this case, semiconductor stocks may not just be in a bearish market, but may enter a full-fledged bear market.
In particular, equipment stocks and Japanese semiconductor stocks with high P/E ratios tend to have their stock prices adjusted significantly before performance changes.
Signals that investors should watch every day
The most important leading indicator for the semiconductor market in 2026 is not just the financial results of semiconductor companies.
Rather, what we should be looking at is the investment stance of those who buy semiconductors.
| Check items | Reasons to watch |
|---|---|
| CAPEX of Microsoft, Amazon, Alphabet, Meta, Oracle | Fuel for AI semiconductor demand |
| NVIDIA's data center sales guidance | Current state of GPU demand |
| TSMC's capital investment plan | Direction of advanced process/CoWoS investment |
| HBM price and inventory | Strength of AI server demand |
| SOX Index | Global Semiconductor Stock Sentiment |
| US long-term interest rates | Valuation pressure on high P/E stocks |
Of particular importance to investors is the
Are the quarterly results of the 4 and 5 giant tech companies showing whether their AI investment plans have been maintained or revised upwards?
It is.
As long as this fuel continues, the semiconductor market is unlikely to collapse.
On the other hand, when the market breaks, you need to change your view of the market.
Summary: From concentration to selection in 2026
The semiconductor market is in a historic expansion phase with global sales expected to reach approximately $1 trillion in 2026.
However, it also faces risks such as the sustainability of AI investments, oversupply, inventory cycles, interest rates, and weak consumer demand.
In conclusion, the semiconductor market price in 2026 will be
Prices of all semiconductors are rising
From,
A selective market where only companies that can truly continue to make profits remain
is moving to.
What investors need is neither blind bullishness nor excessive speculation about a crash.
Things to watch are the CAPEX of big tech companies, the guidance of NVIDIA and TSMC, the supply and demand of HBM, and the order trend of Japanese equipment stocks.
The theory that the semiconductor bubble will burst is not a statement meant to incite fear.
That is, the AI market after 2026 will be
From expectations to results
This is a warning that the situation is about to change.
Reference materials
- SIA Global Annual Semiconductor Sales Increase 25.6% to $791.7 Billion in 2025
- SIA Global Semiconductor Sales Increase 25% from Q4 2025 to Q1 2026
- TrendForce CSP CapEx Expected to Exceed US$600 Billion in 2026
- SEMI Market Intelligence
- SEAJ statistical data
- [TECH+ Japanese semiconductor manufacturing equipment will reach 5.5 trillion yen in 2026] (https://news.mynavi.jp/techplus/article/20260116-3983170/)