Goldman Sachs, in its latest research report, identifies MiniMax as one of the most strategically positioned artificial intelligence model companies in China, with significant potential for growth in the global Total Addressable Market (TAM) for text and code generation.
On February 23rd, Goldman Sachs initiated coverage on the Chinese full-modal AI company MiniMax with a "Neutral" rating. Using a Discounted Cash Flow (DCF) model, the firm set a 12-month target price of HKD 10.18, implying an enterprise valuation of approximately $38.9 billion.
Goldman Sachs highlighted that MiniMax, as one of the world's leading AI model companies, is in a phase of rapid expansion. Its key strengths include a highly globalized revenue structure, with 70% of income originating overseas, and a comprehensive multi-modal product portfolio encompassing text, video, audio, music, and image capabilities. The report emphasizes that MiniMax holds an excellent competitive position within the vast global TAM for text/code, multi-modal applications, and AI agents/digital labor.
Key implications for investors include:
**Valuation Linked to Market Share:** Goldman Sachs provided bull, base, and bear case valuations of $66 billion, $42 billion, and $16 billion, respectively. These scenarios are primarily dependent on the company's expansion of its global market share in AI foundation model subscriptions and API revenue, with a base case projection of reaching 2.5% by 2030 and a bull case of 5.0%.
**Commercialization and Path to Profitability:** Revenue is forecasted to surge from $75 million in 2025 to $980 million by 2027. Although the company is expected to remain loss-making on an adjusted net basis through 2027, gross margins are projected to improve significantly from 21% in 2024 to 38% in 2027. MiniMax is well-capitalized with over $1 billion in cash reserves, deemed sufficient to fund operations until an anticipated breakeven point in 2029.
**Dense Pipeline of Technical Catalysts:** The release of the M2.5 model has notably narrowed the performance gap with leading US State-of-the-Art (SOTA) models, while operating at just one-tenth of the cost. Future catalysts include the launch of the Hailuo 3.0 video model and the potential widespread adoption of AI agents at a cost as low as $1 per hour.
**Multi-modal Advantage and Global Footprint** MiniMax's core moat is built on its full-modal strategy and its high overseas revenue contribution of 70%. The company has adhered to a multi-modal fusion approach since inception, boasting leading models like M2.5 (foundation text), Hailuo 2.3 (video), Speech 2.6 (audio), Music-2.5, and Image-01. This native integration of modalities provides distinct advantages in areas such as instruction following, semantic understanding, and physical interaction.
Commercially, MiniMax has targeted the global market from day one. Its AI-native products have cumulatively served over 212 million individual users across more than 200 countries and regions, while its foundation models serve over 130,000 enterprises and developers. The company has established diverse monetization channels, including subscriptions, token-based billing, online marketing, and API calls.
**Extreme Cost Efficiency and Architectural Innovation** Leveraging a Mixture of Experts (MoE) architecture and linear attention mechanisms, MiniMax achieves performance near the global顶尖 level at extremely low costs. Goldman Sachs estimates the company's total training cost from 2022 to 2025 to be only about $420 million. This cost advantage translates directly into competitive product pricing; for instance, its API pricing is approximately 8% of leading overseas models like Claude 4.5 Sonnet. This allows its To-B API business to maintain healthy gross margins even after significant price reductions. A flat organizational structure, with an R&D team of around 300 people, also contributes to high innovation efficiency.
**Agent Proliferation and the "$1 Digital Labor"** The recently released M2.5 model shows performance approaching that of Claude Opus in programming and office scenarios, but at one-tenth the price. Based on this cost-effectiveness, MiniMax's AI agent applications demonstrate significant disruptive potential. Goldman Sachs calculates that running a MiniMax agent could cost as little as $1 per hour. This economic attractiveness is expected to drive widespread adoption of AI in enterprise scenarios such as HR screening and code debugging.
**Financial Projections and Valuation Rationale** Goldman Sachs forecasts revenue growth to $980 million by 2027, driven primarily by Hailuo AI (video), the open API platform, and C-end entertainment products like Talkie/Xingye. While the company is expected to report adjusted net losses until 2027 due to high R&D and inference costs, gross margins are projected to climb to 38%, supported by an increasing contribution from high-margin multi-modal API revenue. The base case valuation assumes a 2.5% global market share by 2030, with bull and bear cases corresponding to 5.0% and 1.2% shares, respectively.
Comments