--- title: "Morgan Stanley: MiniMax's Short-Term ARR Potential Underestimated; Multiple Catalysts to Materialize in Coming Three Months" type: "News" locale: "en" url: "https://longbridge.com/en/news/283626876.md" description: "Morgan Stanley maintains an Overweight rating on MiniMax with a target price of HKD 990, implying approximately 10% upside from current levels. The report highlights that the market systematically underestimates its multimodal commercial value—generating roughly USD 1 per minute in revenue per inference server at a cost of less than USD 0.3, far exceeding industry averages. Multiple catalysts, including the M3 model launch, API price hikes, and potential inclusion in stock indices, are expected to materialize densely over the next two to three months, potentially offsetting selling pressure as the July lock-up period expires" datetime: "2026-04-22T08:49:12.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/283626876.md) - [en](https://longbridge.com/en/news/283626876.md) - [zh-HK](https://longbridge.com/zh-HK/news/283626876.md) --- # Morgan Stanley: MiniMax's Short-Term ARR Potential Underestimated; Multiple Catalysts to Materialize in Coming Three Months Market expectations for MiniMax's recent annual recurring revenue (ARR) growth are overly conservative. In its latest research report, Morgan Stanley notes that MiniMax's ARR growth potential is on par with or even stronger than peers, supported by its multimodal capabilities, global footprint, and infrastructure advantages—a reality not yet fully reflected in current market pricing. According to the Morgan Stanley report, the bank maintains an Overweight rating on MiniMax with a target price of HKD 990, representing approximately 10% upside from the April 21 closing price of HKD 898.5. **The report suggests MiniMax could achieve an ARR target of USD 1 billion by the end of 2026, matching or surpassing market expectations for competitors.** Several near-term catalysts are expected within the next two to three months, including the launch of new agent products, flagship model upgrades, API price increases, and potential index inclusion. These findings hold direct reference value for investors. Since April, MiniMax's stock performance has lagged behind competitor Z.ai by approximately 40 percentage points. This gap stems primarily from market preference for Z.ai's programming and agent-focused direction, and concerns regarding the lock-up period expiring for MiniMax's IPO cornerstone investors and certain pre-IPO investors on July 8, rather than any substantive fundamental differences. ## Market Systematically Undervalues Multimodal Commercial Potential Current market sentiment systematically underestimates the commercial potential of MiniMax's multimodal capabilities. The report states that **large language models (LLMs) primarily serve monetization in intelligent productivity scenarios, while multimodal models cover creative productivity scenarios; these two markets reinforce each other, collectively forming a substantial commercial space.** In terms of model capabilities, MiniMax's LLM has demonstrated outstanding performance in various agent workflows (such as OpenClaw and Hermes Agent), while maintaining significant exposure to core token consumption drivers like programming and AI agents. Management disclosed that MiniMax's tokens processed per minute (TPM) currently maintains a month-over-month growth rate of 10% to 20%, which Morgan Stanley expects to continue. ## Infrastructure Advantages Translate into Superior Unit Economics Infrastructure capability remains a key constraint on ARR growth for AI foundation model companies, where MiniMax holds distinct advantages. **The report indicates that MiniMax secures global computing power supply through partnerships with major cloud service providers, gains access to advanced chip resources, and achieves superior unit economics through infrastructure optimization.** Specifically, MiniMax generates approximately USD 1 in revenue per minute on a single 8xH800 inference server, with corresponding costs below USD 0.3, significantly outperforming the industry average of around USD 0.5 per minute. This cost structure advantage provides a solid foundation for sustained gross margin improvement. ## Gradual Price Hikes; M3 Launch May Trigger New Round of Increases Regarding pricing strategy, MiniMax has not yet increased input or output API prices but has raised prompt cache read prices (per million tokens) by 50% from RMB 0.21 during the M2.5 era to RMB 0.42 in the M2.7 era. Morgan Stanley anticipates that with the anticipated mid-2026 launch of the M3 model, the company will drive another round of price increases. M3 is positioned as a major model upgrade with significantly expanded parameter scale, accompanied by iterations of the video generation model Hailuo-03. Morgan Stanley believes that the synergistic advancement of model upgrades and price increases will become a key driver for accelerating ARR growth. Additionally, Morgan Stanley identified several catalysts over the next two to three months that could support the stock price: first, potential launches of new agent products; second, implementation of M3 and Hailuo-03 model upgrades; third, potential API price increases accompanying model releases; and fourth, the possibility of MiniMax being added to the Hang Seng Composite Index and Hang Seng Tech Index. The timing window for these catalysts overlaps with the expiration of the July 8 lock-up period, and positive progress at the fundamental level should help offset potential selling pressure resulting from the lock-up release. ### Related Stocks - [00100.HK](https://longbridge.com/en/quote/00100.HK.md) - [MS.US](https://longbridge.com/en/quote/MS.US.md) - [02513.HK](https://longbridge.com/en/quote/02513.HK.md) - [0HSCI.HK](https://longbridge.com/en/quote/0HSCI.HK.md) - [STECH.HK](https://longbridge.com/en/quote/STECH.HK.md) - [MS-O.US](https://longbridge.com/en/quote/MS-O.US.md) - [MS-Q.US](https://longbridge.com/en/quote/MS-Q.US.md) - [MS-E.US](https://longbridge.com/en/quote/MS-E.US.md) - [MS-I.US](https://longbridge.com/en/quote/MS-I.US.md) - [MS-L.US](https://longbridge.com/en/quote/MS-L.US.md) - [MS-P.US](https://longbridge.com/en/quote/MS-P.US.md) - [MS-A.US](https://longbridge.com/en/quote/MS-A.US.md) - [MS-F.US](https://longbridge.com/en/quote/MS-F.US.md) - [MS-K.US](https://longbridge.com/en/quote/MS-K.US.md) ## Related News & Research - [After tech giants, a new cohort of “AI tigers” finds footing in China](https://longbridge.com/en/news/283389393.md) - [ZAWYA: WSO2 launches API platform to make enterprise APIs agent-ready without vendor lock-in](https://longbridge.com/en/news/283299447.md) - [AI quota inflation is no token effort. It's baked in](https://longbridge.com/en/news/283318062.md) - [07:30 ETBallerine Introduces Agentic Detection to Stop Merchant Fraud Early as AI Scams Rise](https://longbridge.com/en/news/283491045.md) - [Scenarios, contradictions, and the ultimate outcome as seen by 16 AI payment practitioners](https://longbridge.com/en/news/283308885.md)