discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.
Featured

MiniMax expands Alibaba cloud pact as compute needs surge for training and inference

Chinese AI firm MiniMax has dramatically increased its cloud computing deal ceiling with Alibaba Group Holding to US$1.2 billion, driven by surging demand for AI model training and inference.

By Ann Cao·Aug 28·scmp.com·4 min read

Intelligence analysis by Gemini 2.5 Flash

MiniMax expands Alibaba cloud pact as compute needs surge for training and inference
Image: scmp.com

MiniMax, a prominent Chinese AI developer, has expanded its three-year cloud services agreement with Alibaba Cloud by 220 percent, raising the total purchase ceiling to US$1.2 billion. This significant increase comes after MiniMax exhausted two-thirds of its original US$115 million budget for 2026 by June, highlighting the escalating computational demands within the AI sector.

Why it matters

This development underscores the immense and rapidly growing demand for computing power among leading AI developers, particularly in China, indicating a significant investment trend in AI infrastructure and the escalating costs associated with advanced model development and deployment.

Imagine a super-smart student named MiniMax who needs lots of special computers to learn new things and answer questions really fast. MiniMax is learning so much that they've used up their computer budget much quicker than expected! So, they've asked their computer supplier, Alibaba, for a much bigger and more expensive computer plan, like upgrading from a small library card to a pass for a giant, super-fast research center, so they can keep learning and growing even faster.

Analysis

The substantial expansion of MiniMax's cloud computing agreement with Alibaba Cloud reflects a critical trend in the artificial intelligence industry: the insatiable demand for computational resources. As AI models, particularly large language models (LLMs) and video-generation models, become more sophisticated, the processing power required for both their initial training and subsequent real-time inference grows exponentially. This deal highlights how even well-funded AI firms are rapidly consuming their allocated budgets, necessitating frequent and significant upward revisions to their infrastructure spending.

US$1.2 billion

MiniMax's decision to raise its three-year purchase ceiling with Alibaba Group Holding to US$1.2 billion marks a dramatic increase of 220 percent from previous limits. This figure is not merely an accounting adjustment but a clear indicator of the scale of investment required to stay competitive in the rapidly evolving AI landscape. The company's original plan to spend US$115 million on Alibaba Cloud services in 2026 was nearly tripled to US$300 million after two-thirds of the budget was utilized by June, demonstrating an underestimation of its own growth and resource needs.

This revised financial commitment extends through 2028, with annual spending limits escalating from US$125 million to US$400 million for 2027, and from US$135 million to US$500 million for 2028. Such aggressive revisions underscore the long-term strategic importance MiniMax places on robust cloud infrastructure. Furthermore, the company also expanded its application programming interface (API) service budget with Alibaba, raising the 2026 cap from US$650,000 to US$7.5 million, an over tenfold increase, bringing its three-year API spending ceiling to US$62.5 million.

MiniMax

MiniMax, a Shanghai-based artificial intelligence firm, is a key player in China's burgeoning AI sector, known for its M-series large language models, the H3 video-generation model, and the popular consumer app Hailuo AI. The company's aggressive expansion of its cloud pact is directly linked to its rapid growth and increasing operational scale. This week, MiniMax reported a remarkable 283 percent surge in first-half revenue, reaching US$116.6 million, primarily fueled by a 700 percent leap in enterprise sales.

This revenue growth provides the financial backing for such substantial infrastructure investments, yet it also highlights the capital-intensive nature of advanced AI development. The company's ability to burn through its cloud budget so quickly suggests a high rate of innovation and deployment, pushing the boundaries of what its existing infrastructure can support. The expansion of its API service budget also indicates a growing reliance on external services and a broader integration strategy for its AI capabilities.

Alibaba Cloud

Alibaba Cloud, a subsidiary of Alibaba Group Holding, is positioned as a critical enabler for China's leading AI developers like MiniMax. This expanded pact signifies a major win for Alibaba Cloud, reinforcing its position as a dominant provider of high-performance computing resources essential for AI training and inference. The deal provides a substantial and predictable revenue stream for Alibaba's cloud division, which is a key growth area for the tech giant.

For Alibaba Cloud, securing such a large and expanding contract with a rapidly growing AI firm like MiniMax demonstrates its capability to meet the demanding computational requirements of cutting-edge AI development. It also serves as a testament to the scalability and reliability of its infrastructure, which is crucial for supporting the intensive workloads associated with large language models and other generative AI applications. This partnership solidifies Alibaba Cloud's role in powering the next generation of AI innovation within China.

Key points

  • Chinese AI firm MiniMax increased its three-year cloud computing deal with Alibaba Cloud by 220% to US$1.2 billion.
  • MiniMax burned through two-thirds of its original US$115 million 2026 budget by June, necessitating the expansion.
  • Annual spending limits on Alibaba Cloud services will jump to US$400 million for 2027 and US$500 million for 2028.
  • MiniMax also expanded its API service budget with Alibaba, raising the 2026 cap tenfold to US$7.5 million.
  • The company reported a 283% surge in first-half revenue to US$116.6 million, driven by a 700% leap in enterprise sales.
The Upside

The significant investment by MiniMax in cloud computing infrastructure signals robust growth and innovation within the Chinese AI sector, potentially leading to the development of more advanced models and applications. For Alibaba Cloud, this pact secures a substantial revenue stream and reinforces its market leadership in providing essential AI computing resources.

The Downside

The rapid consumption of MiniMax's cloud budget highlights the extremely high operational costs associated with developing and deploying advanced AI models, which could strain profitability or necessitate continuous capital raises. This intense demand for compute power also points to potential bottlenecks in infrastructure availability or escalating pricing pressures in the cloud services market.

Market signals

BABA· NYSE
  • BABA The expanded US$1.2 billion cloud computing pact with MiniMax represents a significant and growing revenue stream for Alibaba Cloud, boosting Alibaba Group Holding's financial outlook.

AI-generated analysis of potential market relevance. Not financial advice.

Originally reported at

scmp.com

Discernion covers the story. Read the full piece at the source.

Tagsaicloud-computingchinabusinesstechllmsstartups

Author

Ann Cao

Intelligence analysis by

Gemini 2.5 Flash

Published

Aug 28, 2026

Source

scmp.com

Share

Topics

aicloud-computingchinabusinesstechllmsstartups

Related

More from this desk

Aug 28·scmp.com

Shares in Chinese MLCC makers rise as AI lifts sales and profits

Chinese multilayer ceramic capacitor (MLCC) manufacturers CCTC and Guangdong Fenghua reported significant first-half sales and profit growth, leading to a rise in their share prices, driven by surging demand from AI data centers and a global pricing recovery.

Aug 28·producthunt.com

Revalvo: Run prompts on every model at once. Score. Version. Ship.

Revalvo is a new local-first workbench for prompt engineering and LLM evaluation. It allows users to run prompts on multiple models simultaneously, score responses with built-in evaluators, version prompts, and batch-test datasets before production deployment.

Aug 28·technode.com

Chinese Spatial Intelligence Firm Manycore Launches Lux3D and Starts World Model Testing

Manycore Tech has launched Lux3D, an AI tool that generates 3D assets from text or images, and is testing a new world-model product. The company reported significant revenue and profit growth in the first half of the year.

Aug 28·technode.com

Embodied-AI startup PsiBot raises over $100 million with industrial investors joining

Chinese embodied-AI startup PsiBot, focused on dexterous manipulation, has secured over US$100 million in a new funding round, attracting investments from industrial players and existing shareholders. The capital will fuel research into embodied-world models, human-operat…