Can the AI sector continue to drive the tech stock rally?

   Since the start of 2023, AI-related stocks have led several rounds of gains in global tech stocks. From the “generative AI” boom sparked by ChatGPT, to NVIDIA's record-breaking performance, to the ongoing expansion of data centers, computing power, and model applications.

AI has evolved beyond a mere technological revolution to become the most prominent theme in capital markets. 

However, as we enter the second half of 2024, concerns about whether AI is “too high to sustain” are growing. Issues such as overvaluation, performance delivery, intensifying competition, and policy uncertainty are mounting.

In the context of high interest rates and slowing growth, can the AI sector continue to drive the tech stock rally? This question may well become one of the key variables shaping the direction of global capital markets.

From “concept speculation” to “performance delivery”: The first phase of the AI sector

Looking back over the past year, the rise of the AI sector was not without basis.

From underlying chip manufacturers (such as NVIDIA and AMD), to AI infrastructure (cloud services, data centers, and network equipment), to model development and application layers (Microsoft, Google, Meta, and OpenAI-related concept companies), the entire value chain has seen substantial growth from bottom to top.

The most direct example is NVIDIA's market capitalization surpassing $3 trillion in 2024, overtaking Apple to become the world's second-largest company by market capitalization, trailing only Microsoft. Behind this is the surging global demand for AI computing power and the market consensus that “GPUs are the new oil.”

Microsoft has achieved actual monetization of AI tools by deeply integrating with OpenAI and embedding Copilot into its full suite of Office products. Google, Amazon, and Meta have also successively announced their AI foundation model strategies and commercialization efforts. At this stage, the AI story is not only compelling but also beginning to generate revenue.

The enthusiasm of the capital market has also quickly reflected in valuation levels. According to FactSet data, as of June 2024, the median price-to-earnings ratio of global mainstream AI-related companies exceeded 40 times, while the median for non-AI tech stocks was less than 20 times.

The polarization in valuations reflects the market's dual bet on AI as both a “technology cycle” and an “industrial revolution.”

Valuation bubble or revaluation of value? Market confidence is diverging

However, entering 2025, the performance of the AI sector began to diverge. While NVIDIA remained strong, its stock price volatility began to increase, and the growth of AI-related “edge suppliers” such as Intel, TSMC, and Arm slowed. Some AI startups and AI SaaS companies even saw their valuations halved.

Market divergence is primarily concentrated in three areas:

1. Is AI's monetization capability sufficient?

Currently, AI companies that can clearly drive revenue growth are primarily concentrated in underlying hardware and certain platform enterprises. Most AI applications remain in the testing, deployment, and optimization phases, with a significant gap before achieving large-scale commercialization.

For example, many companies that have deployed large models have found that operational costs are prohibitively high, data security is constrained, and employee usage efficiency falls short of expectations.

2. Will competition compress profits?

While AI technology has high entry barriers, it is not an insurmountable moat. The open-sourcing of models (e.g., Meta's LLaMA 3 and Mistral), computing power leasing (e.g., AWS and Azure AI platforms), and standardization of development tools are gradually lowering participation barriers.

The result is increased homogenization, heightened challenges in achieving differentiated commercialization, and potential compression of profit margins.

3. Will macroeconomic factors undermine valuations?

The US remains in a high-interest-rate cycle, and while the Federal Reserve has signaled potential rate cuts, the pace is slow. In a high-interest-rate environment, the high valuations of tech stocks are under significant pressure. Additionally, stricter global policy regulations (such as the EU AI Act and US restrictions on AI chip exports) cast a shadow over industry development.

Under these intertwined factors, the AI sector's performance is shifting from “broad-based premium” to “structural evolution.” The strong will remain strong, such as Microsoft and NVIDIA; mid-tier companies will either be marginalized or acquired; and fringe concept stocks face the risk of capital outflows.

The “second curve” driven by AI: Who will take the lead?

Although core AI concept stocks face short-term valuation pressures, AI may still continue to support the tech stock market, provided it can drive a “second curve”—that is, substantive recovery and growth in other sectors and industries.

1. Cloud computing and edge computing are reborn

AI model training and inference have extremely high demands for computing power and data transmission, which is driving a reshuffle among cloud computing providers. Since 2024, AWS, Azure, and Google Cloud have all increased investments in AI infrastructure, driving the transformation of cloud services from traditional IT support to “AI-native platforms.”

Meanwhile, edge computing companies (such as Cisco, Dell, and Equinix) are also seeing new growth opportunities due to the demand for “on-premises AI model deployment.”

2. The infrastructure supply chain for data, storage, and networking is gaining momentum

AI is fundamentally “data-driven and computation-intensive,” so sectors like data centers, optical communications, server storage, cooling, and power supply—which were once considered niche—are now benefiting from the AI value chain.

Companies like Broadcom, Marvell, and Supermicro in the US stock market, and Unisoc and Cambricon in the A-share market are getting revalued in the capital market.

3. Software applications are moving from “AI-enhanced” to “AI-reconstructed.”

Applications like AI coding, AI customer service, AI translation, and AI video creation have moved from the demo stage to everyday business scenarios. Companies like Adobe, SAP, and Salesforce are also gradually integrating AI as a core feature into their products.

Although the short-term contribution to revenue is limited, it may lead to business model upgrades and revenue structure optimization in the long term.

Therefore, the AI trend is not driven by a single factor but rather by the “technological ecosystem reconstruction” across the entire industry. Just as the survivors of the 2000 internet bubble truly built the infrastructure of the Web era, AI may now follow a similar path.

How should investors reassess the AI sector?

For investors, the AI sector should be reassessed from the following perspectives:

1. Return to fundamentals and focus on companies that can “deliver.”

Stock prices ultimately return to profits and cash flow. Companies that can truly implement AI into products, services, and customer renewals have long-term investment value. Although valuations are high, if growth is delivered, they are still worth holding long-term.

2. Broaden horizons and focus on “AI beneficiary stocks” rather than just “AI concept stocks.”

Sectors such as advanced manufacturing (AI chip equipment, precision machine tools), new energy (data center power consumption), advanced materials, and intelligent transportation can all unlock new demand through AI enablement. Expand布局 from the “AI origin” to the “AI periphery.”

3. Beware of valuation traps and establish exit mechanisms.

The AI narrative has been hyped too aggressively, leading to speculative bubbles in some small-cap concept stocks. Especially for AI startups or companies that have not yet established a closed-loop business model, risks far outweigh opportunities. It is recommended to allocate to AI ETFs or leading companies for greater defensive resilience.

4. Macroeconomic forecasts must not be ignored.

AI market trends are inseparable from interest rate environments, technology policies, and geopolitical factors. If the Federal Reserve delays interest rate cuts or US-China tech conflicts escalate further, the AI sector may face valuation corrections. Therefore, maintaining dynamic caution is key.

The ultimate outcome of the AI sector: a new-era “platform war”

At its core, AI is not a fleeting tech trend but an “infrastructure war” for the next-generation productivity platform. Whoever controls AI's foundational models, data resources, and computing power platforms will hold the reins of the future information society.

From this perspective, the AI trend is far from over; it has entered a deeper phase. It is no longer a stage for “bandwagon speculation” but a battlefield for “selecting winners.”

Within the next five years, the global technology market may give rise to several “Microsofts of the AI era” or “platform-level operating system companies.” These may not be the hottest companies today but rather long-termists quietly building ecosystems, integrating resources, and accumulating data.

Can the AI sector sustain the tech stock rally? The answer is: yes, but not through speculation, but through ecosystem construction and industrial realization.

This is a marathon, not a sprint. Value persists after the bubble, and calmness is needed after the noise. For investors, embracing AI should involve both passion and resolve.