AI开始缺电 这家基金募了16亿

This article covers the key event of Zhaoxi Capital, a firm focused exclusively on the new energy vertical, closing RMB 1.6 billion for its second fund amid the overwhelming AI investment boom, with practical, on-the-ground insights into primary market dynamics and industry trends.

1. It offers a clear view of real primary market conditions: Capital and attention are heavily concentrated in the AI sector this year, with the number of newly established funds in the first half rising 143% year-over-year, even as total fundraising fell nearly 5% and average fund size shrank. State-backed capital accounts for nearly 90% of total primary market commitments, while purely market-driven financial investments without mandatory local reinvestment requirements make up only 20-30% of the market. Most limited partners (LPs) have not recovered losses from prior investment cycles, and while market attention has returned, actual deployment remains cautious.

2. It outlines practical criteria for evaluating credible investment institutions: First, look at repeat LP commitment rates—half of Zhaoxi’s second fund came from returning LPs, and over 70% of its capital is from market-driven sources without mandatory reinvestment rules, a direct indicator of institutional capability. Second, assess return orientation: Zhaoxi deployed all management fees from its first fund into portfolio investments, reaching a 96% deployment rate above the industry average of 90%, signaling it does not rely on management fees for revenue and has strong confidence in its ability to generate returns. Third, examine investment and exit discipline: Zhaoxi does not fixate on IPOs, prioritizes returning principal to LPs first, and sets clear exit rules for projects with different return expectations.

3. It highlights real bottlenecks in AI development: The bottleneck for AI deployment is shifting from chips to power supply, and AI is set to directly and indirectly expand investment opportunities in the new energy sector by roughly 50%.

The primary market capital flows and industry trends disclosed in this article provide strategic, R&D and brand building reference for brands operating in the technology and new energy sectors.

1. They help brands identify long-term industry and brand communication trends: While primary market capital is currently flooding into the AI track, new energy will not be sidelined—instead, it will see new development dividends driven by power demand from AI computing. Brands focused on energy supporting infrastructure, intelligent applications and domestic substitution will gain greater market visibility.

2. They clarify pragmatic directions for product R&D: The incremental market driven by AI covers three segments: energy infrastructure including intelligent computing center power supply, liquid cooling systems, server power supplies and distributed microgrids; scenario-based application products integrating AI with power inspection, fault prediction, dispatching and power trading; and hard technology products with domestic substitution and scenario expansion opportunities, such as heavy-duty gas turbines and perovskite photovoltaics. For new technology R&D, teams do not need to target large-scale scenarios requiring long validation cycles initially. For example, dry-process perovskite technology can first enter high-premium niche scenarios such as consumer electronics and aerospace before expanding to utility-scale power generation, for faster commercialization.

3. They offer transferable logic for brand trust building: Zhaoxi’s strategy of avoiding AI hype, staying focused on its vertical track and delivering strong "certainty" to partners also applies to brand building. Compared with tying brands to hot concepts, maintaining a clear positioning and delivering predictable returns to partners builds longer-term trust.

The industry trends and market opportunities outlined in this article point out growth directions, flag operational risks and provide actionable operational insights for all types of business operators.

1. Clear growth opportunities to capture: The explosion of AI computing power will directly and indirectly expand relevant market size by around 50%, with key opportunities in intelligent computing center energy supporting facilities, AI applications in power scenarios, and domestic heavy-duty gas turbine supply chains. In particular, solutions that reduce costs and improve efficiency in power grid inspection, fault prediction and dispatching, as well as new technology products that first enter high-premium niche segments before scaling, are positioned to gain orders and scale quickly.

2. Relevant operational risks to avoid: Operators should not blindly expand based on the one-sided narrative of a "primary market boom". Market-oriented LPs have generally not recovered from prior losses, and while attention has returned to the market, actual deployment remains cautious. The vast majority of state-backed capital comes with mandatory local reinvestment requirements, and purely financial, market-driven capital accounts for only 20-30% of total commitments. Entities with fundraising plans should make targeted preparations in advance.

3. Pragmatic operational approaches to adopt: Establish clear investment and exit rules, prioritize principal safety and positive cash flow, avoid blindly chasing hot trends, do not fixate on a single IPO exit path, and deploy capital in a timely manner around industry inflection points (such as corporate business spin-offs, or transitions of new businesses from 0-to-1 to 1-to-10 growth stages) to achieve higher return on investment.

The trend of AI driving upgrades in energy demand highlighted in the article points out new order opportunities, product design directions and digital upgrade pathways for manufacturing factories.

1. Clearly defined new business opportunities to pursue: Construction of AI computing infrastructure directly drives hardware procurement demand, including data center power supply equipment, intelligent computing center liquid cooling supporting facilities, server power supplies and distributed microgrid components. AI deployment in energy scenarios will spur demand for hardware related to intelligent inspection, fault prediction and power grid dispatching. Currently, leading overseas gas turbine manufacturers have order backlogs extending beyond 2030, creating opportunities for factories producing gas turbine whole machines, upstream blades, high-temperature alloys, coatings and processing services to enter the global supply chain; after maturing technically, these players can further expand into the aero-engine supply chain.

2. More pragmatic commercialization pathways for product design: New technology products do not need to initially target large-scale application scenarios that require long validation cycles. For perovskite photovoltaic technology, for example, adopting a dry process route to first enter segments with high unit value and lower lifespan requirements—such as consumer electronics, flexible scenarios and aerospace—before expanding to utility-scale power generation will lead to faster commercialization and higher returns on investment.

3. Practical entry points for digital upgrades: Factories can deploy AI technology in production dispatching, equipment fault prediction, site inspection and power trading to tangibly reduce costs and improve efficiency, building differentiated competitive advantages with customers.

The primary market and industry changes noted in the article clearly identify industry trends, core client pain points and corresponding solution directions for B2B service providers.

1. Clear industry development trends to track: While primary market capital is currently heavily concentrated in the AI track, structural opportunities remain in the new energy vertical. AI is set to expand investment opportunities in the new energy sector by roughly 50%, and investment and industrial services at the energy layer currently have very few participants, making it a scarce track. The industry broadly faces lengthened fundraising cycles: state-backed capital is the main source of funding, but market-driven capital without local reinvestment requirements accounts for only 20-30% of total commitments. Most LPs are dissatisfied with the return performance of funds they have invested in, making "certainty and security" the core consideration for LPs when selecting partner institutions.

2. Core pain points of different client segments to accurately identify: Investment institution clients generally face difficult fundraising, insufficient LP trust and high exit pressure. Energy manufacturing clients commonly struggle to identify viable AI落地 scenarios, face long technology commercialization cycles and encounter supply chain upgrade challenges. Power grid and data center clients suffer from low inspection efficiency, difficulty in fault prediction, high dispatching costs and mounting power supply pressure.

3. Targeted service solutions to design: Around the above pain points, providers can deliver precision services including fundraising endorsement, exit process optimization, AI scenario matching, commercialization pathway design, supply chain resource matching and intelligent system construction to fill existing market gaps.

The venture capital and industry trends disclosed in this article provide clear reference for investment attraction, operation management and risk mitigation for platforms offering venture capital services and industrial services.

1. Accurate insight into real demands of platform users: The core tension in the current venture capital space is that LPs are cautious about committing capital, setting high bars for actual return capability and capital security from partner institutions, while general partners (GPs) generally face lengthened fundraising cycles and difficulty finding high-quality projects at inflection points. On the industrial side, startups at the intersection of AI and energy need financing matchmaking and resource matching services, while traditional energy manufacturing enterprises seek support for AI upgrades and entry into new supply chains.

2. Clear priorities for investment attraction and operation: Platforms can focus on onboarding two categories of entities: first, vertical-sector investment institutions with clear positioning, strict investment and exit discipline, strong return performance and high repeat LP commitment rates, such as GPs that focus on full new energy industry chain layout rather than blindly chasing hot trends; second, high-quality projects positioned at the intersection of AI and energy, including AI applications in power scenarios, intelligent computing center energy supporting facilities, domestic gas turbine supply chains, and perovskite technology enterprises pursuing pragmatic commercialization routes.

3. Proactive trend risk mitigation in operations: Avoid blindly supporting pure concept-driven AI projects, stay alert to technology routes requiring 20-30 years of ultra-long validation cycles with vague commercialization prospects, and refrain from onboarding entities that only capitalize on hot trends without delivering actual returns to partners. Prioritize serving institutions and projects with strong落地 capability and tangible cash value creation.

The phenomenon of Zhaoxi Capital closing a new energy vertical fund against the prevailing AI investment boom reflects multiple new trends and emerging issues in China’s primary market and new energy industry, carrying high research value.

1. Notable new dynamics in the primary market to track: In the first half of this year, the number of newly established domestic funds rose 143% year-over-year, but total fundraising size shrank by nearly 5%, with average fund size declining. State-backed capital accounts for nearly 90% of primary market commitments, while purely market-driven financial investments without mandatory local reinvestment requirements make up only 20-30% of the market. The LP community has not recovered from the impact of below-expectation investment returns in previous cycles; while attention has heavily returned to the AI track, actual deployment remains cautious. Vertical-focused GPs that deliver predictable security to LPs, maintain clear investment and exit discipline, and demonstrate strong return performance can successfully complete fundraising against the trend even when investing in non-hot sectors.

2. Emerging industrial development trends to follow: The bottleneck for AI deployment is shifting from chips to power supply. AI will directly and indirectly expand investment opportunities in the new energy sector by roughly 50%, covering three directions: computing energy infrastructure construction, AI-enabled traditional energy scenarios, and domestic substitution of traditional manufacturing sectors such as high-end gas turbines driven by AI demand. Vertical capital with deep industry roots shows clear advantages in capturing industry inflection points.

3. Relevant new questions for in-depth research: These include the reasonable scale boundary of non-standardized investment approaches for vertical industry funds (for example, Zhaoxi, based on its team capacity and investment approach, considers RMB 2 billion the upper size limit for a single fund), the balance mechanism between state-backed capital’s local reinvestment requirements and market-driven financial return demands, and shifts in investment return logic as AI restructures the energy industry landscape.