Quick Summary
Covering the latest research from top Wall Street investment banks

M3 Model Launch Boosts MiniMax's Leading Position in China’s LLM Market; Maintain Overweight Rating

Institution
Morgan Stanley
Date
20260601
Authors
Lydia Lin, Gary Yu
Company
MiniMax
Ticker
0100
Industry
AI, AR
Rating
Overweight
BullishHigh confidenceReiterateMedium-termMaintain overweight rating, target price of HK$1,100 implies 31% upside
AuthorsLydia Lin, Gary Yu
Target priceHK$1,100.00
CoverageChina
Research firm divisions/subsidiariesMorgan Stanley Asia Limited(Division/Team)

AI summary card

M3 Model Launch Boosts MiniMax's Leading Position in China’s LLM Market; Maintain Overweight Rating

Morgan Stanley believes MiniMax's newly launched M3 large model has competitive advantages in performance and pricing strategy, driving rapid ARR growth, maintaining an overweight rating and optimistic outlook for its A-share IPO before 2027.

Overweight | Target Price HK$1,100
Artificial IntelligenceLarge Model UpgradeARR GrowthA-share IPOPricing StrategyTechnical Advantage
  • M3 is the world's only open-weight model supporting a 1 million token context window
  • API prices have doubled compared to M2.7, still at 11-14% of U.S. frontier model prices
  • ARR increased from $150 million to $300 million in two months, domestic revenue share rose to 40%
  • Maintain overweight rating, target price of HK$1,100 implies 36x 2027 sales multiple
  • Expected to complete A-share listing in 2027, Hailuo 03 product to be launched in June

Report interpretation

Overview

Morgan Stanley releases a management interview summary on MiniMax, focusing on its flagship large model M3 launched on June 1. The report views M3 as achieving major breakthroughs in coding capability, multimodal understanding, and long-context processing, supported by pricing adjustments that accelerate commercialization. It maintains an overweight rating and sees strong medium-term growth potential.

Core views

Model Technology: As the world’s only open-weight model, M3 integrates top-tier coding and agent performance, native multimodal understanding (text/image/video), and a 1 million token context window. MiniMax Sparse Attention (MSA) technology reduces computational cost to 1/20 at 1 million context length, prefill speed increases 9x, and decoding speed increases 15x. Pricing Strategy Adjustment: M3 API prices have doubled compared to M2.7, placing it in the mid-high range among China's leading models (higher than DeepSeek V4 Pro and MiMo V2.5 Pro, comparable to GLM5.1, lower than K2.6), but only 11-14% of U.S. frontier model prices. Subscription plans transitioned from request-based to token-based, offering 15 times more tokens per price than U.S. peers. Significant Business Progress: Over the past two months, ARR grew from $150 million to over $300 million, domestic MaaS revenue share rose from below 30% to 40%, ToC/ToB revenue ratio reached 50%/50%. The company has initiated A-share IPO guidance and expects completion by 2027.

Analysis framework

The institution uses DCF (Discounted Cash Flow) valuation methodology, assuming a 15% WACC and 3% perpetual growth rate, with the target price implying a 36x 2027 sales multiple. The analysis focuses on three drivers: technical breakthroughs (MSA architecture), pricing strategy (cost-performance advantage), and commercial progress (ARR growth). Competitive positioning is validated through comparisons of parameters, prices, and ARR data between U.S. and Chinese large models.

Methodology notes

  • Valuation MethodDCF Discounted Cash Flow

    DCF Discounted Cash Flow

    Estimates company value by forecasting future free cash flows and discounting them to present value. This article assumes a 15% Weighted Average Cost of Capital (WACC) and 3% perpetual growth rate, reflecting the high-growth characteristics and risk premium of the AI industry.

  • Industry/Industrial Analysis FrameworkSupply and Demand Framework

    Supply and Demand Framework

    Analyzes the match between supply-side technological breakthroughs (e.g., MSA reducing computational costs) and demand-side commercial progress (ARR growth) to assess corporate competitiveness.

Asset mapping & comparison

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

  • MiniMax(0100.HK)
    Direct beneficiary, M3 model launch drives tech leadership and revenue growth
    Strengths
    MSA technology lowers compute cost, significant ARR growth, A-share IPO expectation
    Weaknesses
    Subscription plan change may cause loss of price-sensitive users
    Comparison
    API price lower than K2.6 but higher than DeepSeek V4 Pro, performance comparable to U.S. frontier models
    Risks
    Geopolitical risk, intensified industry competition

Key data

  • ARR (Annual Recurring Revenue)$300 million+Data as of April 2026, double from February's $150 million
  • M3 API PriceRMB 6.30 per million tokensDoubled from M2.7, at 11-14% of U.S. model prices
  • Target Price Implied Valuation36x 2027 Sales MultipleCalculated using DCF model
  • Domestic Revenue Share40%Rose from below 30% over the past two months

Impact & implications

The report believes that the M3 launch will consolidate MiniMax’s leading position in China’s frontier large model market. Its technological advantages and pricing strategy are expected to accelerate commercialization. The A-share IPO expectation offers capital operation space, though risks include sustainability of ARR growth and industry price wars.

Risks

  • Geopolitical risk
  • Intensified industry competition and price wars
  • Model performance lagging behind peers
  • User resistance due to subscription plan changes

What to watch

  • Latest ARR data post-M3 launch
  • Progress on Hailuo 03 product launch in June
  • Larger parameter M3 series expected in Q3 2026
  • Progress of A-share IPO review process
Zhejiang ICP No. 2022035445-5
Disclaimer: Market data, charts, indicators, research views, and other information provided on this website are intended solely for information display, research communication, and educational reference. They should not be regarded as personalized investment advice, securities recommendations, trading instructions, solicitations, or guarantees of return. While we strive to improve the reliability of our data and content, such information may still be subject to delays, errors, incompleteness, or untimely updates due to source differences, methodological limitations, system processing, or market volatility. Users should exercise independent judgment based on their own circumstances and bear all risks and responsibilities arising from the use of this website.

Settings

Sign in to view recent logins