Bank of America: Meta sells computing power, aiming to tell a compelling AI investment return story

Meta is plotting to monetize its vast AI computing assets. This initiative is not only the embryonic form of a new business line, but also a strategic signal responding to investors' doubts about high capital expenditure returns. According to reports, Meta is formulating plans to launch a cloud infrastructure business, offering external clients access to AI computing power and model services. After the announcement, Meta’s stock price surged nearly 10% in a single day, far outpacing the S&P 500’s rise of about 0.25%, with the market reacting positively to this potential new business line. According to Chasing Trends Trading Desk, BofA Securities analysts Justin Post and Nitin Bansal stated in a July 1st research report that the advancement of the cloud business will help highlight the potential value of Meta’s computing assets and model development, thereby alleviating investor concerns about the company’s continued investment in AI infrastructure without seeing returns. Bank of America maintains a Buy rating on Meta, with a target price of $835. Meta’s Cloud Plan Emerges: Two Paths Progressing in Parallel According to Bloomberg citing sources, Meta's cloud business currently has two directions: first, providing AI model hosting services, allowing developers to access various models running on Meta's existing AI infrastructure, including its Muse Spark series, and charging based on access, similar to Amazon AWS’s Bedrock product; second, directly selling raw computing power, resembling new cloud computing providers like CoreWeave. This initiative belongs to a Meta internal strategy called "Meta Compute," focusing on the construction and operational management of AI infrastructure. Meta’s CEO has publicly hinted at business opportunities in the corporate market and stated that the company could sell computing power externally at prices above the cost of construction. BofA points out in its report that, from a macro perspective, if Meta’s 2026 capital expenditure can support up to 3GW of computing power (estimated at $4–4.5 billion per GW), then establishing a cloud business platform soon would give the company greater strategic flexibility—once there is excess computing power, it could be leased externally at $1–1.5 billion per GW per year, providing a positive impact for the company. Competitive Positioning in Question, Strategy Debate Unavoidable Despite the enthusiastic market response, BofA candidly points out potential concerns. Meta’s progress in self-developed chips seems behind mature large-scale cloud providers such as Amazon, Microsoft, and Google. Meanwhile, the company is still actively purchasing computing power via third-party agreements—including a recent 1.6GW procurement deal with Crusoe. This phenomenon prompts the market to question Meta’s strategic logic: Can a company still needing to purchase computing power externally, build a convincing computing power resale business? How will it position its competitiveness in the hyperscale cloud market? BofA believes that whether Meta can gain stronger market recognition in this field partly depends on the advancement of its large language model (LLM) capabilities—the higher the model capabilities, the greater the external demand for Meta’s computing power, and the more solid the business logic for the cloud business. AI Unit Economics Improvement—A Double-Edged Sword for Cloud Providers Beyond Meta’s cloud plan, the cost aspect of AI computing power is also showing noteworthy signals. According to The Information, OpenAI has reportedly found a system-level optimization solution that reduces the inference cost for certain models by about half. This optimization achieves more efficient utilization of existing server infrastructure, without the need for new hardware or model architecture. OpenAI has reportedly applied this optimization to ChatGPT traffic from non-logged-in users, which can now be run on just a few hundred Nvidia GPUs. The exact mechanism is unclear, and it’s uncertain whether it can be extended to logged-in users, API workloads, or compute-intensive inference products. BofA believes that improvements in computing cost efficiency are positive for the overall direction of large internet companies: if this technology spreads across the industry, it would expand the effective output of existing computing power without additional hardware investment, reduce the urgency for additional capital expenditure, and improve the unit economics of AI businesses. As agentic application scenarios drive massive token consumption, the strategic value of computing power optimization will become increasingly prominent. However, for hyperscale cloud providers, lower inference costs could also bring some price pressure risks; but at the same time, better gross margin structures and a broader addressable market are expected to drive continued growth in AI workload demand, which is seen as overall positive. ~~~~~~~~~~~~~~~~~~~~~~~~ The above excellent content is from Chasing Trends Trading Desk. For more in-depth interpretation, including real-time analysis and frontline research, please join [Chasing Trends Trading Desk▪Annual Membership]. Risk warning and disclaimer: The market carries risks and investment should be cautious. This article does not constitute personal investment advice and does not take into account individual users’ specific investment objectives, financial status, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their specific situation. Invest accordingly and take responsibility for your own actions.