Morgan Stanley is bullish on multiple catalysts for MiniMax over the next 2-3 months

Overview

This is a Morgan Stanley company research report on MiniMax (0100.HK). The core view is that although MiniMax has recently underperformed Zhipu/Z.ai, the market is not fully pricing in its ARR growth, model capabilities, multimodal commercialization, and infrastructure unit economics advantages. The report believes product, model, pricing, and index-related events over the next 2-3 months could act as stock catalysts.

Core views

The report's core bullish thesis has three points: first, MiniMax's ARR growth could be in line with or even stronger than peers, while the market underestimates its revenue scale by the end of 2026; second, MiniMax has model capability and commercialization potential in agent workflows, multimodal, and AI coding/agent token consumption scenarios; third, global cloud partners, access to advanced chips, and optimized inference costs give it a favorable unit economics model.

Analysis framework

The report uses an analytical framework combining company fundamentals and catalysts: it first compares MiniMax's stock performance and expectation gap versus Z.ai, then breaks down variables such as ARR, TPM, model capabilities, multimodal, inference costs, pricing adjustments, and index inclusion, and finally uses DCF valuation to derive a target price while listing upside and downside risks.

Methodology notes

• valuation methodDCF

  discounted cash flow valuation

  The report uses the DCF method for valuation, assuming a 15% WACC and a 3% terminal growth rate; the target price implies about 56x 2027e P/S.

• financial forecastingMorgan Stanley ModelWare

  internal forecasting framework

  Unless otherwise specified, the financial metrics in the report are based on the Morgan Stanley ModelWare framework; consensus data comes from Refinitiv Estimates.

• relative ratingOverweight / Equal-weight / Underweight

  rating system relative to the industry coverage universe

  Morgan Stanley's Overweight means the expected risk-adjusted total return over the next 12-18 months is above the average of the analyst's industry coverage universe.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• MiniMax (0100.HK)

  Core covered name, a Hong Kong-listed AI foundation model and multimodal applications company.

  Strengths

  It has strengths in agent workflow capabilities, multimodal commercialization opportunities, global compute supply, access to advanced chips, and optimized inference costs; TPM is growing rapidly, and product and model upgrades are future catalysts.

  Weaknesses

  It has recently underperformed Z.ai; the market is concerned about the lock-up expiry on July 8 for IPO cornerstone investors and some pre-IPO investors.

  Comparison

  The report says MiniMax has lagged Zhipu/Z.ai by 40 percentage points since April, as the market prefers Z.ai's near-term ARR growth in coding and agentic directions and its more aggressive price increases.

  Risks

  Key risks include geopolitics, intensified competition and price wars, model performance lagging peers, and supply pressure from lock-up expiry.

Key data

• RatingOverweightThe disclosure table shows MiniMax is rated O, dated 02/20/2026.
• Target priceHK$990.00The report table lists the price target as HK$990.00.
• Closing priceHK$898.50Corresponds to the closing price on April 21, 2026.
• Upside to target price10%The report table lists upside/downside to price target as 10%.
• ARR target/peer referenceUS$1bn by end-2026The report says MiniMax's ARR growth could be in line with or stronger than peers.
• TPM growth10%-20% WoWManagement said tokens per minute is growing 10%-20% week over week.
• Inference cost<US$0.3/minuteThe report says MiniMax's inference cost is below the industry average of about US$0.5/minute.
• Inference revenueabout US$1/minuteThe report says it can generate about US$1 of revenue per minute on an 8×H800 inference server.
• 2026e revenueUS$240mnIn the financial table, Revenue, net 12/26e is US$240mn.
• 2027e revenueUS$712mnIn the financial table, Revenue, net 12/27e is US$712mn.
• 2028e revenueUS$1,531mnIn the financial table, Revenue, net 12/28e is US$1,531mn.

Impact & implications

If the report's judgment is correct, MiniMax's share price may be driven by catalysts in the short term to recover from relative underperformance; medium-term valuation support depends on ARR delivery, the boost to token consumption from model upgrades, the pace of multimodal commercialization, and whether demand remains resilient after price increases. Given the roughly 10% upside to the target price, this report is more oriented toward catalyst-driven expectation repair rather than a deep-value re-rating with substantial upside.

Risks

• Geopolitical risks may affect computing power, advanced chips, and global business deployment.
• Intensifying competition and price wars in AI foundation models may compress revenue growth and gross margin.
• If model performance lags peers, ARR growth, TPM growth, and commercialization expectations may fail to materialize.
• The July 8 lock-up expiry for IPO cornerstone investors and some pre-IPO investors may create short-term share price pressure.
• If potential price increases are not accepted by customers, call volume and revenue growth quality may be affected.

What to watch

• Whether a new agent product will be launched over the next 2-3 months.
• The launch timing, parameter scale, and performance of the major M3 upgrade and Hailuo-03 model upgrade.
• Whether API or cache-related pricing will be raised alongside model releases.
• Progress on potential inclusion in HSCI and HSTECH.
• Whether TPM can continue to maintain 10%-20% week-over-week growth.
• Whether ARR can approach the peer reference level of US$1bn by the end of 2026.
• Trading volume and stock price performance before and after the July 8 lock-up expiry.