Historically, productivity revolutions take over a decade or even two decades to significantly impact the economy, and there is currently no evidence suggesting AI progress has reached a peak.
AI data centers have high power demands, requiring energy storage performance that exceeds traditional commercial and industrial storage, which could present a dual opportunity for volume and price growth in the energy storage sector.
This technology bull market may not be driven by leverage or liquidity, but rather by industrial trends. As long as companies see growth in revenue and profits with room for market cap expansion, capital inflows are likely.
At the intersection of the current technology and energy revolutions, short-term volatility is inevitable, but the true driver of long-term returns is the direction of industrial trends.
These are the key insights shared by Zhang Jinqi, Deputy General Manager of Red Phoenix Fund, during the "Navigating the Fog: Understanding the Times and Identifying Strategies" live series on HaoMai TV. As Deputy General Manager, General Manager of the Private Investment Department, and Investment Director, Zhang has specialized in technology sector research and investment since beginning his career. He currently leads the Technology and Resources Group at Red Phoenix Fund, offering in-depth research and unique perspectives on AI computing power, electricity, and applications.
This year, the AI industry has experienced high景气度, while the new energy sector has seized new development opportunities amid the global energy transition wave. During the HaoMai TV "Navigating the Fog" live series, Zhang discussed the opportunities and challenges brought by the technology and energy revolutions, sharing his analysis of the AI industry cycle and new energy sector prospects. The following is a summary of his live broadcast.
AI Technology Revolution Still in Early Stages, Infrastructure Investment Reaches Mid-to-Late Phase
Regarding the current stage of the AI infrastructure cycle, Zhang suggests that, from a financial perspective, AI infrastructure investment is in the mid-to-late phase. Over the past three years, AI investment opportunities have primarily focused on capital expenditure. Capital expenditure began rising sharply in 2023, experiencing two years of high growth by 2025. Market expectations at the start of the year forecasted about 40% growth in 2026, but data from major overseas tech companies indicates growth may exceed 50% or even 60% this year.
A notable warning sign is that by the end of this year, capital expenditure from major overseas tech companies will nearly consume all of their free cash flow. This is historically rare. US companies typically prefer using free cash flow for stock buybacks, even borrowing for that purpose. Once the ratio of capital expenditure to free cash flow reaches a critical point, continued growth requires financing, which raises the question of returns on such massive capital expenditure.
Early signs of returns from asset spending are emerging, as AI investments by major tech companies have accelerated growth in traditional businesses like advertising, video streaming, and cloud services. However, no company has yet been able to isolate and analyze the specific return on investment for AI. Recently, two major US AI model companies have filed for IPOs, which may provide more clarity on the relationship between capital expenditure and returns after listing.
From a technical standpoint, AI is comparable to any previous technology revolution, and the scale of this industrial revolution may be even larger. Compared to the mobile internet, which significantly reduced the cost of information transmission, AI is replacing human intelligence itself, representing a major revolution in production methods. Before the arrival of AGI or even ASI, the pursuit of technological progress will continue to drive capital expenditure growth in the industry.
Historically, productivity revolutions take over a decade or even two decades to have a significant economic impact. AI technology is still advancing, with scaling laws remaining effective and optimization ongoing. There is currently no evidence that AI progress has reached a bottleneck.
Energy Storage Enters a New Cycle, AI Drives Volume and Price Growth
In analyzing the new energy sector, Zhang argues that it should not be discussed broadly, but rather broken down into specific sub-sectors. The photovoltaic industry is still in a phase of capacity removal, with the main production chain facing oversupply. The industry's turning point may be delayed further.
The wind power sector is generally stable with slight growth. Onshore wind is approaching saturation, but offshore wind is growing strongly, with both domestic and international demand on the rise.
Energy storage and lithium batteries are the areas Zhang is most optimistic about. Energy storage is at the start of a new cycle. After geopolitical conflicts, energy storage has become a widely accepted energy solution globally. The technological foundation of energy storage is still improving, and the penetration rate of storage-equipped markets is still at a relatively low level, leaving significant room for future growth.
Lithium batteries have generally exceeded expectations this year. Although the growth rate of new energy vehicle production and sales has slowed, the battery capacity per vehicle is steadily increasing, with 1,000 km range expected to become a common feature in high-end models. The demand-driving effect of energy storage on lithium batteries is clear. Investment strategies can focus on three main themes: lithium ore resources, battery leaders, and new technology directions.
AI's demand for energy storage is a significant incremental factor. Previously, energy storage primarily addressed energy security and backup issues. However, overseas AI data centers, with their aging power infrastructure and insufficient supply capacity, are facing energy bottlenecks. As a result, energy storage is now taking on the role of backup power or even alternative power sources. AI data centers have high power requirements and demand higher energy storage performance than traditional commercial and industrial storage, creating a dual opportunity for volume and price growth in the energy storage sector.
Focus on Industrial Trends, Primarily Technology with Diversified Allocation
Regarding allocation strategy, the core approach is to focus on industrial trends to identify high-certainty directions. Since 2023, AI has become a key industrial trend, and this overall assessment remains unchanged. However, the current phase demands higher certainty. Improvements in technical specifications are certain, while cyclical price increases are likely temporary. Therefore, priority should be given to directions with technological advancement.
Diversification should be implemented in two aspects: first, diversification across technology routes. AI technology development is divergent rather than convergent, so it is necessary to assess probabilities across different routes and invest accordingly. Second, position diversification. It is advisable to moderately reduce AI holdings and increase supplementary allocations to sectors like non-ferrous metals, while still maintaining a primary focus on technology.
Regarding the concentration of tech stock holdings, the recent correction has entered the mid-to-late stage of deleveraging. A rotation between high and low positions requires two conditions: a sufficiently large gap between the high and low sectors, and signs of a reversal in the low sectors. As of now, low sectors have not yet shown systemic reversal signals, so any rotation is more likely a rebound than a reversal.
This technology bull market may not be driven by leverage or liquidity, but rather by industrial trends. As long as companies see growth in revenue and profits with room for market cap expansion, capital inflows are likely. At the intersection of the current technology and energy revolutions, short-term volatility is inevitable, but the true driver of long-term returns may be the direction of industrial trends. During periods of position clearance and low market sentiment, it may instead be an opportune time to seek quality assets.
The technology and energy revolutions are advancing and intertwining at an unprecedented pace. The AI industry is moving from infrastructure to application outbreaks, the new energy sector is shifting from policy-driven to economic-driven growth, and energy storage is opening new growth space catalyzed by AI computing power demand. Only by standing at the intersection of these three main themes can one seize the investment opportunities presented by the times.
Comments