Goldman Sachs maintains Buy on MiniMax, but lowers target price to HK$860 due to M3 price cuts

Overview

This is a Goldman Sachs company research report on MiniMax Group, focusing on the key investor debates after the M3 model launch, MiniMax’s path to achieving its ARR targets, the potential launch of Hailuo 3, financing pressure on independent AI model companies, and revisions to valuation and earnings forecasts. The report maintains a Buy rating, but lowers the target price from HK$1,000 to HK$860 due to more aggressive M3 pricing leading to lower gross margins for the base text API.

Core views

The report’s core view is that the negative market reaction after the M3 launch does not fully indicate a deterioration in MiniMax’s fundamentals, but rather reflects investor concerns over pricing power, frontier model capability, and financial strength. Goldman Sachs believes that besides the high-end model path of maximum intelligence and maximum pricing power, MiniMax can also achieve ARR scale through low pricing, sufficiently good model performance, global reach, and higher token share. If Hailuo 3 launches smoothly in the coming weeks, the company will benefit from the video generation segment, which has less competition and more favorable pricing and margins. Financing capability remains a key variable for independent AI model companies, as MiniMax is not expected to get closer to EBIT breakeven until around 2029E.

Analysis framework

The report evaluates MiniMax using event-driven analysis, peer stock-price comparisons after model launches, a token usage and pricing framework, revisions to API gross margin and net loss forecasts, and updated DCF valuation. Goldman Sachs views M3 price cuts as an alternative path to ARR expansion outside the premium high-price/high-performance route, and derives the target price using target market share, WACC, terminal growth rate, and long-term adjusted EBIT margin.

Methodology notes

• Valuation methodsDCF valuation

  Discounted cash flow

  The target price is based on DCF valuation, using a 12% WACC and 2% terminal growth rate, and assumes the global subscription plus API revenue market share rises by 0.2-0.7 percentage points annually from 2027E-2030E, reaching 2.5% in 2030E.

• Commercialization analysisToken usage and pricing quadrant chart

  ARR quadrant

  The report divides model commercialization paths into two favorable quadrants: high intelligence with high pricing power, and low price with high adoption, and believes M3 can drive ARR through lower pricing, agentic use cases, and higher token share.

• Scenario analysisBull/bear valuation scenarios

  bull/base/bear case

  Goldman Sachs lowered the base-case target price to HK$860, while reducing the bull-case valuation from HK$1,600 to HK$1,350 and the bear-case valuation from HK$420 to HK$330.

Asset mapping & comparison

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

• MiniMax Group (00100.HK)

  Covered company in the report

  Strengths

  Comprehensive multimodal product offering, strong global reach, advantage in unit token cost, rapid growth in M3 token usage, and Hailuo 3 is expected to enter the more favorable video generation segment.

  Weaknesses

  M3 price cuts weaken gross margins for the base text API and raise concerns over pricing power; the company remains loss-making, making capital and self-financing ability important.

  Comparison

  Share price performance was weaker than that of Knowledge Atlas after the GLM5.2 launch; financing scale is also weaker than that of DeepSeek mentioned in the report.

  Risks

  Model performance below expectations, slower-than-expected commercialization, cash burn, IP and content generation risks, and geopolitical risks.

• Knowledge Atlas (Zhipu, 2513.HK)

  Hong Kong-listed peer comparison

  Strengths

  Strong market reaction after the GLM5.2 launch; the report says its share price significantly outperformed the Hang Seng Index.

  Weaknesses

  The report does not cover the company and does not provide detailed fundamental valuation.

  Comparison

  MiniMax shares fell after the M3 launch, while Knowledge Atlas rose after the GLM5.2 launch, creating a clear contrast.

  Risks

  The report does not elaborate on company-level risks.

• DeepSeek

  AI model peer and financing capability benchmark

  Strengths

  The report says its funding exceeds US$7bn and its valuation exceeds US$50bn, indicating stronger financial strength.

  Weaknesses

  The report does not elaborate on its commercialization and profitability details.

  Comparison

  Its funding scale is significantly higher than the US$0.6bn raised by MiniMax in its January 2026 IPO.

  Risks

  The report does not elaborate on company-level risks.

• Seedance 2.0

  Video generation competition and Hailuo 3 reference

  Strengths

  The report says its ARR has further increased to US$2bn, with API margins of around 70%.

  Weaknesses

  Its SOTA performance and first-mover advantage have intensified market concerns over the launch timing of Hailuo 3.

  Comparison

  Goldman Sachs believes competition in the video generation segment is generally weaker than in base text models; if Hailuo 3 launches smoothly, MiniMax’s multimodal path will still have an advantage.

  Risks

  If Hailuo 3 underperforms expectations in capability or launch timing, it will weaken MiniMax’s multimodal narrative.

Key data

• Report date2026-06-17Equity Research 17 June 2026.
• RatingBuyGoldman Sachs maintains a Buy rating on MiniMax.
• 12-month target priceHK$860.00The previous target price was HK$1,000.
• Current share priceHK$416.60The current price shown in the report.
• Implied upside106.4%Based on the HK$860 target price and the current price of HK$416.60.
• Share price performance after M3 launch-41%The Hang Seng Index was -4% over the same period.
• Share price performance after Knowledge Atlas launched GLM5.2+34%The Hang Seng Index was -1% over the same period, used for peer comparison.
• M3 parameter scale428 billion parametersThe report says M3 has advantages in dataset and cost efficiency.
• Low-price agentic model price rangeUS$0.1-0.2 per 1M tokensGoldman Sachs believes this is a path to ARR scale through higher token share.
• Premium model price rangeUS$0.9-1.2 per 1M tokensCorresponding to premium coding scenarios with maximum intelligence and maximum pricing power.
• M3 token usage rankingOpenRouter weekly #1The report says M3 recently ranked first by token usage on the third-party API platform OpenRouter.
• Token speed30TPS increased to 70TPS, with a further 30-40% increase plannedImprovements in computing resources and inference speed disclosed by MiniMax.
• API revenue gross margin assumptionFY26E/FY27E/FY28E at 40%/30%/35%Previously 50%/50%/55%; lowered due to more aggressive M3 pricing.
• Adjusted net loss forecast2026E/2027E/2028E at -US$425mn/-US$485mn/-US$376mnPreviously -US$409mn/-US$420mn/-US$203mn.
• Key DCF assumptions12% WACC; 2% terminal growth; 2035E long-term adjusted EBIT margin of 18%The long-term adjusted EBIT margin was previously 21%.

Impact & implications

In terms of investment implications, although the report lowers the target price and earnings forecasts, it still emphasizes that MiniMax has advantages among Chinese AI model companies in terms of a complete multimodal product suite, global commercialization, unit token cost, and organizational efficiency. In the short term, the share price may continue to be affected by M3 pricing power, Hailuo 3 launch timing, financial strength, and industry price wars; in the medium term, the key question is whether the low-price, high-usage strategy can translate into ARR growth, and whether video generation and multimodal APIs can maintain healthier margins.

Risks

• In global foundation model competition, model performance may come in below expectations.
• Improvement in earnings visibility may be slower than expected.
• Commercialization capability may be weaker than expected.
• IP and content generation-related risks.
• Cash burn and self-financing capability risks.
• Geopolitical risks arising from escalating US-China technology competition.
• M3’s low-price strategy may continue to pressure gross margins for the base text API and market confidence in pricing power.

What to watch

• The launch timing, rollout stability, and market feedback for Hailuo 3 in the coming weeks.
• Whether M3 can sustain its token usage share on third-party API platforms such as OpenRouter.
• Whether gross margins for the base text API stabilize near Goldman Sachs’ revised assumptions after M3 price cuts.
• Whether MiniMax can deliver on its ARR targets through the low-price, high-usage strategy.
• Whether video generation and multimodal APIs can continue to maintain relatively high margins.
• Changes in funding and shareholder structure related to the A-share listing plan and the expiry of the six-month Hong Kong listing lock-up period.
• The gap versus peers such as GLM/Qwen, DeepSeek, and Seedance in model performance, pricing, and financing capability.