Analysts from Everbright Securities Company Limited have released a report stating that Meta Platforms, Inc.'s trial of external compute capacity rental is aimed at opening new monetization channels for its massive AI infrastructure investments. This move confirms the critical role of cloud computing in quickly recouping upfront capital expenditures and improving current-period returns. The ability to monetize externally also provides a cash flow guarantee to support the ongoing expansion of AI investments, and Meta's overall medium to long-term plan to continue increasing its AI compute capacity remains unchanged.

ETF investment tools to consider include: 159363, 515000.

Market Trend Focus

News that Meta Platforms, Inc. plans to rent out surplus computing capacity triggered a sharp correction in global AI hardware and tech stocks, with South Korea's Kospi index plunging as much as 7% intraday. However, market panic may have led to an overreaction. Industry insiders and analysts believe this move is not a signal of compute oversupply or a slowdown in capital expenditure. Instead, it represents an exploration into cloud services by Meta, driven by its consumer-facing business model to enhance shareholder returns and cash flow.

In reality, Meta's compute strategy exhibits a characteristic of "using the old to fund the new." This means generating cash flow from recycling existing, older computing capacity while continuing its pursuit of cutting-edge, high-end compute power. Meta has already signed a contract with Crusoe this week to secure approximately 1.6 GW of new AI computing capacity. Furthermore, it has maintained its high capital expenditure guidance for Q1 2026, indicating that its demand for high-end GPUs needed for frontier model training remains urgent. Current hardware demand primarily stems from the inference side rather than the training side, and the rental market's order backlog remains substantial.

Although short-term market sentiment has been impacted by profit-taking due to concerns about the pace of AI investment, the medium to long-term narrative for AI capital expenditure remains intact. Sectors related to computing power, power supply, and the broader technology theme still hold allocation value. Investors should rationally assess Meta's strategic adjustment, focusing on the structural opportunities arising from the expansion of AI demand rather than succumbing to blind panic.

In-Depth Analysis

Regarding Meta Platforms, Inc.'s trial of external compute rental, Everbright Securities Company Limited points out that the underlying rationale is to open a new monetization channel for its massive AI infrastructure investment. This indirectly confirms that cloud computing is the fastest way to recoup initial investments and improve current-period returns. This move demonstrates that external monetization capability has become a cash flow guarantee supporting Meta's continued expansion of AI investment, and its overall medium to long-term plan to ramp up AI compute has not shifted.

Recently, North American tech giants have seen a surge in capital expenditures. Alphabet, Amazon, Microsoft, and Meta Platforms, Inc. have announced that their total capital expenditure for 2026 will reach $660 billion, a 60% increase from 2025. AI has transitioned from a cost center to a revenue engine. For example, Amazon's AI shopping assistant Rufus has generated $12 billion in incremental annual sales, and Azure's annual revenue exceeds $75 billion. Although there are short-term market concerns that high capital expenditure may consume cash flow, we believe that within this AI investment supercycle, the growth trajectory of AI tokens has not slowed, and commercial returns are robust.

Furthermore, Amazon AWS announced a price increase of approximately 15% for its EC2 machine learning capacity blocks, breaking a long-standing tradition of "only decreasing, never increasing" prices. NVIDIA made an additional $2 billion investment in CoreWeave to accelerate AI compute construction, and Meta signed a $6 billion fiber optic agreement with Corning. These events collectively validate the high demand for AI cloud infrastructure. It is recommended to pay attention to investment opportunities in the AI cloud IaaS sectors encompassing "AIDC + compute rental" and the fiber optic segment.

Investment Rationale

Meta Platforms, Inc.'s trial of external compute rental is not a signal of compute oversupply but rather a move to open new monetization channels for its massive AI infrastructure investment. This confirms the crucial role of cloud computing in rapidly recouping upfront investments and improving current-period returns. This initiative not only strengthens the cash flow guarantee for continued AI investment expansion but also reflects the strategic wisdom of "using the old to fund the new." This involves generating cash flow from recycling existing computing capacity while maintaining an urgent demand for cutting-edge, high-end compute power, as evidenced by the recent contract securing 1.6 GW of new AI capacity and the maintained high capital expenditure guidance. Concurrently, Amazon AWS's 15% price increase, breaking a two-decade tradition of "only decreasing, never increasing," and NVIDIA's additional $2 billion investment in CoreWeave both corroborate the high global demand for AI compute and the scarcity of infrastructure. Additionally, Meta's planned $6 billion procurement of fiber optics, combined with the long expansion cycle and tightening supply of optical fiber preforms, is expected to continue driving demand and prices for fiber optic cables upward. Overall, the AI capital expenditure supercycle is far from its peak. The explosion of AI agents is pushing the growth ceiling for tokens to limitless heights. Sectors such as compute rental, AIDC data centers, and fiber optics possess strong growth logic and allocation value. It is recommended to focus on the earnings elasticity of leading companies within these sectors under the current price increase trend.

The ChiNext Artificial Intelligence ETF Huabao (159363) invests in popular sectors such as communications, computers, and media. Its top ten holdings are New E-Strong, Zhongji Innolight, Tianfu Communication, Jiangbolong, Xiecang Data, Beijing Junzheng, Tonghuashun, Runze Technology, Changxin Boke, and BlueFocus, with a combined weighting of 67.99%.

The Technology ETF Huabao (515000) invests in popular sectors such as electronics, communications, and computers. Its top ten holdings are Zhongji Innolight, GigaDevice, Naura Technology, Montage Technology, Advanced Micro-Fabrication Equipment Inc., Hygon Information, Hengrui Medicine, Shengyi Technology, Sanhuan Group, and Wus Printed Circuit, with a combined weighting of 51.54%.