Are there still investment opportunities in AI stocks in 2025? A comprehensive guide to popular AI concept stocks and investment strategies

The AI wave is no longer just a future prospect but a reality in progress. Since the explosive popularity of ChatGPT, AI concept stocks have become darlings of the capital market. However, after more than two years of frenzy, is this sector still worth entering? This article will help you clarify the investment logic of AI stocks from three dimensions: market status, specific targets, and risk warnings.

AI Concept Stocks Market: From Concept to Reality Turning Point

According to the latest data from IDC, global enterprise spending on AI solutions and technologies is expected to reach $307 billion by 2025, surpassing $632 billion by 2028, with a compound annual growth rate of about 29%. This not only indicates that AI is not a bubble but also reveals a deeper change—AI is shifting from a compute power race to application deployment stage.

In this transition, capital flows have already shown differentiation. Take Bridgewater Fund as an example: in Q2 2025, they significantly increased holdings in key targets like NVIDIA, Alphabet, and Microsoft. This means smart money is no longer blindly buying AI concept stocks but selectively choosing leading companies that truly control core technologies and market share. Meanwhile, industry allocation through thematic funds and ETFs has also become mainstream. As of Q1 2025, global AI and big data fund assets exceeded $30 billion.

Key change: If 2023-2024 was a phase where “anything bought would go up,” then starting from 2025, investors need to carefully select sectors and individual stocks.

How to Classify the AI Industry Chain? Where Are the Investment Opportunities?

The AI industry chain is generally divided into three levels:

Level 1: Compute Power and Chip Suppliers

This is the upstream of AI infrastructure, including GPU manufacturers, server vendors, etc. These companies benefit most directly but are also most susceptible to market fluctuations. NVIDIA achieved revenue of $60.9 billion in 2024, an increase of over 120% year-over-year. In Q2 2025, revenue hit a new high of $28 billion, making it an absolute winner of the AI wave.

However, this level also faces fierce competition. AMD’s MI300 series accelerators are gradually gaining recognition from cloud service providers, with Q2 2025 revenue up 18% YoY, gradually eroding NVIDIA’s market share. This reminds investors that even dominant companies’ control is not eternal.

Level 2: Infrastructure and Support Services

Including companies providing cooling solutions, power management, server OEMs, network chips, etc. These companies tend to have stable growth and lower risks but are less aggressive in growth compared to upstream chip manufacturers.

For example, in Taiwan stocks, Quanta Computer focuses on AI servers, with revenue of NT$1.3 trillion in 2024, surpassing NT$300 billion in Q2, up over 20% YoY. Delta Electronics provides power and cooling solutions, with revenue around NT$420 billion in 2024 and a 15% YoY increase in Q2 2025. The performance of such companies is highly predictable.

Level 3: Application and Cloud Platforms

Large tech companies like Microsoft and Google monetize through cloud platforms and enterprise AI services. Microsoft’s fiscal year 2024 revenue was $211.2 billion, with Azure cloud services growing 28%, and AI services contributing more than half of that growth. These companies have strong monetization capabilities, but their stock valuations are already relatively high.

Differentiated Opportunities in Taiwan Stocks and US Stocks AI Concept Stocks

Taiwan Stocks’ Role

Taiwan AI concept stocks are mostly in the mid-to-upstream of the industry chain, mainly chip design, manufacturing, and supporting services. TSMC (2330), as the world’s largest wafer foundry, is the main OEM for all AI chips, with revenue of NT$3.2 trillion in 2024, up 18.8%. MediaTek (2454) continues to push in AI mobile and edge computing, with a 20% YoY increase in revenue in Q2 2025.

Advantages of these companies include: supply scarcity under capacity constraints, deep technological moat, and relatively stable profitability. But the disadvantages are also obvious—stock valuations are already high, and growth potential may not be as explosive as in early stages.

E.g., Vanguard-KY (3661) represents another opportunity in Taiwan stocks—AI chip design service providers. In 2024, revenue was NT$68.2 billion, up over 50% YoY, with Q2 2025 revenue doubling to NT$20 billion. These companies benefit from explosive AI orders but also face customer concentration risks.

US Stocks’ Advantages and Traps

US AI concept stocks cover all parts of the industry chain, with broader market participation. NVIDIA (NVDA), although its stock price has risen 176% (as of September 2025), still has a market cap of $4.28 trillion, making it the largest chip company globally, and difficult to be replaced in the short term. Broadcom (AVGO), in network chips and custom ASICs, has positioned itself in AI data centers, with AI products accounting for 25% of revenue in 2024, and target prices mostly above $2,000.

But a special reminder: valuation risks in US tech stocks should not be ignored. Microsoft, as the biggest beneficiary of enterprise AI transformation, with a market cap of $3.78 trillion, faces increased difficulty in doubling its stock price.

Which AI Concept Stocks Can Still Be Laid Out in 2025?

Three Priority Target Types

1. Chip and Accelerator Manufacturers: In the short term, these are still the biggest beneficiaries. But be clear—not all chip companies can make money. Only those mastering AI-specific chip design (like NVIDIA, AMD, Vanguard-KY) have sustained competitiveness. Traditional chip companies’ success in AI transformation is limited.

2. Infrastructure and Supporting Suppliers: Companies like Delta Electronics, Sunway (3324) focusing on cooling and power solutions have stable but modest growth. But that’s their advantage—when the entire AI sector faces adjustments, these companies tend to be more resilient.

3. Cloud Platforms and Application Companies: Microsoft, Google, and other platform companies are moving from pilot projects to large-scale AI applications. Their long-term growth certainty is higher, but short-term stock fluctuations can be larger.

Portfolio Suggestions

Avoid over-concentrating on a single AI concept stock. A more prudent approach is: allocate 30% to chip and server companies, 40% to infrastructure and supporting firms, and 30% to cloud platform and application companies. Use dollar-cost averaging to smooth out costs rather than chasing high prices in one shot.

Where Are the Deadly Risks of AI Concept Stocks?

Historical Lessons on Valuation Bubbles

Back in 2000, Cisco reached a peak of $82 during the internet bubble, then fell over 90% to $8.12, and even 20 years later, it has not returned to its peak. Yahoo, once an internet leader, was overtaken by Google. Microsoft and Google’s stocks also experienced long adjustments after market peaks. What does this tell us? Even high-quality companies’ stock prices cannot rise forever.

Currently, AI concept stocks face similar risks—the market has already assigned relatively high valuations, meaning future growth must be supported by stronger earnings. If AI’s expected performance is not validated by results, a significant correction could occur.

Black Swans in Policy and Regulation

Countries are strengthening AI-related regulations. The EU’s AI Act and the US’s AI executive orders are being implemented gradually. If data privacy, algorithm bias, or copyright regulations tighten, it could impact certain AI companies’ business models and valuations. Especially for companies relying on large-scale data training, regulatory changes could directly threaten their core competitiveness.

Uncertainty in Technical Routes

AI chip architectures are evolving rapidly. While NVIDIA’s CUDA ecosystem is deep, competitors like AMD and Intel are catching up. If new technological architectures emerge, current leaders could be overturned. For example, if quantum computing or other new computing architectures mature ahead of schedule, the entire GPU market might face restructuring.

Rapid Reversal of Market Sentiment

AI concept stocks are highly sensitive to news. A negative report (such as a major customer canceling an order or a competitor launching a new product) can trigger chain reactions of selling. In 2024, there were multiple instances of AI sector daily drops exceeding 5%. This poses the greatest risk for investors chasing high.

How to Invest Efficiently in AI Concept Stocks?

Direct Stock Purchase vs Funds/ETFs

Buying stocks directly carries the highest risk but also the greatest potential. Suitable for investors with research ability and risk tolerance. If choosing AI-themed funds or ETFs (like Taishin Global AI ETF, Yuan Global AI ETF), risks are more diversified, but management fees and premium/discount issues should be considered.

Dollar-Cost Averaging Over Lump Sum

Given the volatility of AI concept stocks, systematic investment through periodic purchases is more rational. This helps to average costs and avoid buying at peaks. Monthly or quarterly investments are recommended rather than lump-sum chasing after a stock’s rapid rise.

Timing vs Stock Selection

Rather than trying to guess the best entry point, focus on selecting better companies. Leading companies tend to outperform over the long term, so the emphasis should be on identifying firms with genuine competitive advantages rather than whether the entry is at the lowest point.

Beware of Platform and Costs

Investing in US stocks can be done via Taiwanese brokers’ cross-trading or opening accounts with overseas brokers. Different platforms have different fees and interest costs, which need to be compared. Over the long run, these costs, though seemingly small, can add up significantly.

Overall Outlook: AI Investment Landscape 2025-2030

Short-term Volatility, Long-term Uptrend

From late 2025 to 2026, AI concept stocks may face a period of performance verification. If corporate earnings growth slows, stock prices could see significant corrections. But this does not change the long-term growth trend of the AI industry.

Industry Differentiation Intensifies

Not all AI-related companies will ultimately succeed. Some seemingly promising firms may fall behind due to wrong technical routes, management errors, or market shifts. Therefore, continuous attention to company dynamics and industry changes is crucial.

Shift in Investment Mindset

From “buy and hold” to “regular evaluation.” The AI field changes rapidly—missing three months could mean missing new industry patterns. Regularly review your portfolio to ensure it still aligns with long-term logic and to identify risks of being overtaken by new competitors.

In summary, AI concept stocks still have opportunities, but the chance now lies not in blindly chasing gains but in rational understanding of the industry chain, in-depth analysis of corporate competitiveness, and full awareness of risks.