The 2025 AI Profit Wave Is Coming: A Complete Analysis from Industry Chain to Investment Opportunities

Since the emergence of ChatGPT, generative AI has shifted from a technological concept to a core driver of the capital markets. Companies related to AI technology not only see valuation boosts but also demonstrate strong momentum in actual performance growth. So, what is the current investment landscape of the AI industry? Which fields and individual stocks are worth关注? This article will delve into the core investment logic of AI concept stocks.

How Artificial Intelligence Penetrates the Capital Market

AI (Artificial Intelligence) essentially endows machines and systems with human-like cognitive abilities—learning new knowledge, performing logical reasoning, solving complex problems, understanding and generating natural language and visual content. From voice assistants, autonomous driving to medical diagnostics, AI applications have deeply integrated into daily life.

AI concept stocks refer to listed companies whose business is closely linked to artificial intelligence technology. These companies may be involved in chips manufacturing, server supply, cloud platforms, or software services. Fundamentally, investing in AI concept stocks means investing in the infrastructure and ecological applications of the entire AI industry chain.

The Scale of the AI Industry Drives Investment Boom

According to the latest IDC data, global enterprise investments in AI solutions and technologies are accelerating. By 2025, global enterprise AI expenditure is projected to reach ### 307 billion USD, and by 2028, the total AI expenditure including applications, infrastructure, and related services is expected to surpass ### 632 billion USD, with a compound annual growth rate of 29%.

Among these, expenditure on accelerated servers will surpass 75% by 2028, becoming the core hardware supporting AI commercialization. This indicates enormous investment potential at the infrastructure level, with upstream industry chain companies benefiting most.

This investment boom has attracted active allocations from institutional investors and hedge funds. For example, Bridgewater Fund significantly increased holdings in NVIDIA, Alphabet, and Microsoft in Q2 2025. Meanwhile, thematic ETFs and AI-related funds have become new choices for investors. As of the end of Q1 2025, global assets in AI and big data funds exceeded ### 30 billion USD.

Leading Companies in Taiwan’s AI Supply Chain

Quanta Computer (2382): A Successful Transition to Server Business

As a global leader in notebook OEM, Quanta has successfully entered the AI server field in recent years. Its subsidiary, Quanta Cloud Technology (QCT), specializes in servers and cloud solutions, has penetrated the US hyperscale data center supply chain, with major clients including chip manufacturers and international cloud service providers.

In 2024, Quanta’s revenue reached NT$1.3 trillion, with the proportion of AI servers continuously rising. Entering 2025, Q2 revenue broke NT$300 billion, up over 20% year-on-year, setting a new high for the same period. Foreign investors generally remain optimistic about its long-term growth, with an average target price around NT$350–NT$370.

Silicon Motion (3661): Explosive Orders for AI Accelerators

Silicon Motion focuses on customized chips (ASIC), serving clients including US cloud giants and leaders in high-performance computing. In 2024, full-year revenue reached NT$68.2 billion, with a growth rate over 50%, demonstrating strong momentum driven by AI demand.

In Q2 2025, quarterly revenue exceeded NT$20 billion, doubling from the same period last year, with gross margin and net margin both improving. As large AI client projects enter mass production, market expectations for its growth potential remain high, with an average target price between NT$2,200 and NT$2,400.

Delta Electronics (2308): Key Supplier of Cooling Systems

Delta Electronics, a global leader in power management and power solutions, has actively entered the AI server supply chain in recent years, mainly providing high-efficiency power supplies, cooling, and cabinet solutions. In 2024, revenue was about NT$420 billion, with the share from data centers and AI applications steadily increasing.

In Q2 2025, revenue was approximately NT$110 billion, up over 15% year-on-year, with gross margin remaining high, reflecting strong demand for AI infrastructure construction.

MediaTek (2454): Player in Mobile and Edge AI Deployment

MediaTek, a global fabless chip designer, covers mobile chips, smart home, and automotive electronics. With the rise of edge computing and generative AI, MediaTek is advancing its AI chip deployment. Its Dimensity series has integrated enhanced AI computing units, and it collaborates with NVIDIA on developing automotive and edge AI solutions.

In 2024, revenue reached NT$490 billion, benefiting from increased AI chip shipments, with gross margin improving quarter by quarter. In Q2 2025, revenue was about NT$120 billion, up approximately 20% year-on-year, with an average target price between NT$1,300 and NT$1,400.

Sunway (3324): Core Manufacturer of Liquid Cooling

As AI server chips’ power consumption surpasses the kilowatt level, traditional air cooling has reached a bottleneck. Sunway specializes in high-performance liquid cooling modules, successfully positioning itself in the global AI server supply chain. In 2024, full-year revenue was NT$24.5 billion, with over 30% annual growth.

In 2025, progress accelerates as major cloud service providers adopt liquid cooling solutions, leading to a significant increase in shipments of water-cooled AI server modules. With new high-power AI chips entering the market, the penetration rate of liquid cooling will rise rapidly, with many foreign institutions setting target prices above NT$600.

Leading Companies in the US AI Industry

NVIDIA (NVDA): The Computing Empire’s Core

NVIDIA has become the industry standard for AI training and inference with its GPUs and CUDA software platform. In 2024, revenue reached $60.9 billion, with an annual growth rate over 120%, demonstrating explosive growth driven by AI demand.

In Q2 2025, revenue hit a new high of about $28 billion, with net profit increasing over 200% year-on-year. The strong demand from cloud service providers and enterprises for Blackwell architecture GPUs continues to boost its data center business. As AI applications shift from training to inference and penetrate enterprise and edge scenarios, NVIDIA’s high-performance computing solutions will continue to grow exponentially.

Broadcom (AVGO): The Network Hub of AI Data Centers

Broadcom plays a key role in AI chips and network connectivity. Its advantages in custom ASIC chips, network switches, and optical communication chips position it well in the AI data center supply chain. In FY2024, revenue was $31.9 billion, with AI-related product revenue rapidly increasing to 25%.

In Q2 2025, revenue grew 19% year-on-year, benefiting from major cloud providers accelerating AI data center deployments, with demand for Jericho3-AI chips and Tomahawk5 switches continuing to rise. As AI model sizes expand, the demand for high-performance network connectivity will grow quickly. Most foreign reports rate it as a “Buy,” with target prices above $2,000.

AMD (AMD): Challenger in Second Sourcing

AMD acts as an innovator in the AI accelerator market. Its Instinct MI300 series and CDNA 3 architecture successfully penetrate the NVIDIA-dominated market, providing important alternatives for customers. In 2024, revenue was $22.9 billion, with data center business growing 27% annually.

In Q2 2025, revenue increased 18% year-on-year, benefiting from the adoption of MI300X accelerators by major cloud providers and the launch of MI350 series, with AI-related revenue multiplying. As AI workloads become more diverse, customer demand for alternative solutions grows, and AMD leverages its CPU+GPU integration to gradually expand its market share.

Microsoft (MSFT): Platform for Enterprise AI Transformation

Microsoft, through its exclusive partnership with OpenAI, Azure AI cloud platform, and Copilot enterprise assistant, leads enterprise AI transformation. In FY2024, revenue reached $211.2 billion, with Azure and cloud services growing 28%, and AI services contributing over half of the growth.

In FY2025, AI commercialization accelerates as large-scale deployment of Copilot and Azure OpenAI services occurs. Its intelligent cloud revenue first surpasses $30 billion. As Copilot is deeply integrated into products used by over 1 billion users worldwide, its monetization potential will continue to unfold. Many institutions see it as the most certain beneficiary of the enterprise AI wave.

Long-term Investment Outlook for AI Concept Stocks

The investment value of AI concept stocks ultimately depends on the development direction of AI technology. It is certain that AI will profoundly change production and lifestyles, similar to the internet, generating enormous economic effects.

Industry cycle insights: In the early stage, infrastructure and hardware suppliers will benefit first. However, historical experience shows that such high growth is difficult to sustain long-term. Referring to Cisco Systems (CSCO), a pioneer in internet equipment, which reached a high of $82 in the 2000 bubble but then fell over 90% to $8.12. After 20 years of operation, its stock price still has not returned to the peak. Stocks like these are suitable for phased investment strategies.

Challenges for downstream application companies: Companies developing AI or using AI to improve operations are generally considered to have more sustainable growth. However, the historical trajectories of giants like Microsoft and Google show that their stock prices also peaked during bull markets and then fell sharply, remaining difficult to recover to previous highs for years. Yahoo, once a leading internet company, was eventually overtaken by Google and others, illustrating that industry dominance is not eternal.

In theory, investors who can timely adjust holdings may achieve long-term returns. But for ordinary investors, accurately grasping market shifts is challenging. Therefore, phased investment should focus on AI development speed, technology monetization ability, and whether individual stocks’ profit growth shows signs of slowdown.

Diversified Paths to Investing in AI

Besides directly buying individual stocks, investors can also allocate through funds and ETFs:

Direct stock investment: Convenient trading, low costs, but risk concentrated in single stocks. Suitable for investors with deep industry understanding. Examples include TSMC (2330), NVIDIA (NVDA).

Stock mutual funds: Managed by fund managers selecting a mix of stocks, balancing risk and return. Moderate trading costs, suitable for those seeking diversification.

Index ETFs: Track AI-themed indices, with the lowest trading costs and management fees. Prone to premiums or discounts. Related products include Taishin Global AI ETF (00851), Yuan Global AI ETF (00762).

Investors can combine dollar-cost averaging strategies to buy stocks, funds, or ETFs, reducing volatility and average costs. Although AI is still in a rapid growth phase, the positive effects may not last long in the same company, as stock prices may already reflect AI benefits. Continuous adaptation is key to maximizing investment performance.

Investment Outlook for 2025–2030

With the rapid development of large language models and multi-modal AI, demand for computing power, data centers, cloud platforms, and dedicated chips will continue to grow. In the short term, chip and hardware suppliers like NVIDIA, AMD, and TSMC will remain the biggest beneficiaries, with infrastructure-related companies continuing to enjoy order dividends.

In the medium to long term, AI applications in healthcare, finance, manufacturing, autonomous vehicles, and retail will gradually land, translating into actual revenue for more enterprises and driving overall growth of AI concept stocks.

Macro constraints: Capital markets still focus on AI, but stock trends are inevitably influenced by central bank policies. Loose interest rate environments favor high-valuation tech stocks, while high rates may compress valuations. Additionally, AI concept stocks are sensitive to news, prone to short-term volatility, and capital may flow into new themes like renewable energy. Short-term turbulence is unavoidable, but the long-term trend remains upward.

Policy and regulatory variables: Governments worldwide view AI as a strategic industry, likely increasing subsidies and infrastructure investments, providing positive support. However, issues like data privacy, algorithm bias, copyright, and ethics may lead to stricter regulations, challenging some company valuations and business models.

Overall investment characteristics: At this stage, AI concept stocks will show a “long-term bullish, short-term volatile” pattern. Investors should prioritize chip, accelerated server, and infrastructure suppliers, or select companies with tangible applications such as cloud services, medical AI, and fintech. Diversified investment via AI-themed ETFs is also an effective strategy to reduce individual stock risk.

For ordinary investors, a more prudent approach is long-term allocation and phased entry, avoiding chasing highs in the short term to reduce market turbulence impact.

Core Risks Investors Must Understand in AI Investment

Industry Development Uncertainty

Although AI has existed for decades, practical applications at scale have only become feasible recently. Rapid technological iteration and industry changes make it difficult even for seasoned investors to keep pace. After purchasing a stock, investors may easily fall into hype cycles around that company, experiencing significant volatility.

Unproven New Enterprises

While major tech firms are involved in AI, many emerging AI startups lack history and proven track records. These companies carry higher operational risks than established, time-tested firms. Investors should evaluate carefully.

Potential Risks from AI Development

Experts in computer science have warned of possible dangers from AI development. As the field expands, public opinion, regulation, and policy environments may change unexpectedly, impacting AI concept stocks’ performance. Investors should fully understand these risks before investing.