Goldman Sachs Maintains MiniMax Buy Rating, Target Price Adjusted to 860 HKD

Overview

Goldman Sachs released a research report conducting in-depth analysis on MiniMax Group's stock performance and market concerns following the M3 model launch on June 1. The report points out that although M3 price cuts raised concerns about pricing power and performance, this is actually an alternative path for the company to maximize Annual Recurring Revenue (ARR) by using highly attractive pricing strategies to boost Token usage. Goldman Sachs maintains a Buy rating for MiniMax but lowers the 12-month target price from 1000 HKD to 860 HKD, implying a valuation of approximately $35 billion.

Core views

On M3 Price Cut Strategy and Market Reaction: After MiniMax M3 launch, stock price fell 41%, mainly due to market concerns over lack of pricing power behind the price cut, as well as contrast with peer Knowledge Atlas (Zhipu AI) GLM5.2 launch followed by 34% stock rise. Goldman Sachs believes MiniMax's low-price strategy (approx $0.22 per Million Tokens) allowed it to maintain sustainable positive gross margins while performance matched trillion-parameter models. Data shows M3 Token usage on third-party API platforms like OpenRouter quickly climbed to number one, proving the strategy effectively boosted market share. Video Generation Model Hailuo 3 Outlook: Compared to highly competitive foundational text models, the competitive landscape in video generation is more favorable with higher profit margins. MiniMax's next-gen video model Hailuo 3 is mostly trained, expected to be released within upcoming weeks. Referencing peers Seedance 2.0's high 70% API gross margin and $2 billion ARR expectation, Goldman Sachs believes MiniMax's multimodal strategy will deliver better overall profit prospects, distinguishing from other AI standalone firms focusing only on foundational text models. Financial Forecast and Valuation Adjustment: Given M3's more aggressive pricing strategy, Goldman Sachs lowered API revenue gross margin forecasts for FY26E/FY27E/FY28E to 40%/30%/35% (previously 50%/50%/55%), where foundational text model gross margin expected only 10-20%. Correspondingly, adjusted net loss forecasts widened. Based on DCF valuation model (WACC 12%, terminal growth rate 2%), assuming global market share increases yearly and long-term adjusted EBIT profit margin 18%, target price set at 860 HKD.

Analysis framework

Goldman Sachs adopted an ARR Quadrant analysis framework, decomposing the path to maximizing ARR into two dimensions: Maximize Token Usage and Maximize Pricing Level. The report notes MiniMax is moving towards the upper-left quadrant via price cut strategy, exchanging partial unit price for explosive usage growth. Additionally, the report applied sum-of-parts valuation logic, differentiating foundational text model business which faces fierce competition and pressured margins from video generation multimodal business with better competitive landscape and higher margins, to more comprehensively assess the company's long-term profitability potential.

Methodology notes

• Industry/Industrial Analysis FrameworkSupply and Demand Framework

  ARR Quadrant Analysis (Volume vs Pricing)

  By breaking down Annual Recurring Revenue (ARR) into Token Usage and Unit Price variables, analyze corporate revenue maximization paths under different competitive strategies. MiniMax chooses low-price high-volume strategy, aiming to compensate for price reduction through economies of scale.

• Valuation MethodDCF Cash Flow Discounting

  DCF Valuation Model

  Use Weighted Average Cost of Capital (WACC) 12% and terminal growth rate 2% to discount company's future free cash flows to derive intrinsic value and target price, suitable for unprofitable tech companies with high growth potential.

• Industry/Industrial Analysis FrameworkUpstream-Midstream-Downstream Industry Chain Transmission

  Differentiated Competition between Multimodal and Foundational Text Models

  Analyze competitive landscape and margin differences across different sub-sectors in AI industry chain (e.g., foundational text vs video generation), pointing out video generation field has better pricing power and margin space due to technical barriers and relatively scarce supply.

Asset mapping & comparison

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

• MiniMax Group (0100.HK)

  Beneficiary target, improves market share via M3 low-price strategy, Hailuo 3 expected to improve profit structure

  Strengths

  Strong multimodal integration capability, strong commercial execution, low Token cost, high organizational efficiency

  Weaknesses

  Foundational text models face fierce price wars, gross margin under pressure, capital strength weaker than some giants

  Comparison

  Compared to Knowledge Atlas (Zhipu AI), MiniMax has deeper layout in multimodal and video generation fields, but short-term stock performance weaker; compared to DeepSeek, smaller capital scale

  Risks

  Model performance below expectations, slow commercialization process, tight capital chain, geopolitical risks

Key data

• 12-Month Target Price860 HKDReduced from 1000 HKD, implies valuation approx $35 billion
• M3 Model PriceApprox $0.22 / Million TokensHighly competitive low-price strategy, driving usage increase
• OpenRouter RankingNo. 1Weekly Token usage ranking climbed to first after M3 launch
• FY26E API Gross Margin40%Reduced from 50%, reflecting impact of foundational text model price war
• Expected EBIT Break-even Time2029Rely on positive gross margin to cover high training and operating costs

Impact & implications

For MiniMax, short-term stock price volatility reflects market worries about price wars, but long term, its strategy of acquiring market share through low prices helps build user stickiness and data flywheel. Upcoming launch of video generation business will become new growth engine, improving overall profit structure. For investors, need to monitor its capital burn rate and subsequent financing ability, as well as ability to maintain technological leadership amidst fierce competition.

Risks

• Global foundational model industry competition intensifies leading to model performance below expectations
• Profit visibility path slower than expected
• Commercialization ability weaker than expected
• IP and content generation related risks
• Cash burn and self-financing ability risks
• Geopolitical risks brought by intensified US-China tech war

What to watch

• Launch progress and market feedback of Hailuo 3 video generation model
• Sustained growth situation of M3 model Token usage
• Progress of company A-share listing plan
• Actual trend changes in API gross margins