🤯 AI Revolution: Alibaba's M890 Chip 🔥

May 20, 2026 |

Tech

🎧 Audio Summaries
English flag
French flag
German flag
Japanese flag
Korean flag
Mandarin flag
Spanish flag
đź›’ Shop on Amazon

đź§ Quick Intel


  • Alibaba launched the Zhenwu M890 AI processor, developed by T-Head, achieving three times the performance of the Zhenwu 810E.
  • The M890 is designed for AI agents requiring long-context retention, real-time coordination, and complex multi-step task execution.
  • Alibaba has shipped over 560,000 Zhenwu units to date, with deployments across 20 industries including automakers and financial services.
  • The V900 processor is expected in Q3 2027, delivering approximately a threefold performance gain compared to the M890.
  • The J900 processor is slated for release in Q3 2028, anticipating another threefold performance increase.
  • Alibaba’s Qwen 3.7-Max large language model can operate continuously for up to 35 hours without performance degradation.
  • The M890 will be available through Alibaba Cloud’s Bailian platform, utilizing the Panjiu AL128 server system with 128 M890 accelerators.
  • Production footprint data is being gathered at scale ahead of the M890’s rollout, taking place across Amsterdam, California, and London.
  • 📝Summary


    Alibaba has introduced the Zhenwu M890, a new AI processor developed by its T-Head subsidiary, boasting three times the performance of its previous model, the Zhenwu 810E. Designed for AI agents requiring extensive context and complex task execution, the M890 is intended to shape the future of enterprise AI. Alibaba reports shipping over 560,000 Zhenwu units to over 400 customers across various industries, including automotive and financial services. The company’s silicon roadmap includes the V900 in 2027 and the J900 in 2028, each anticipating significant performance gains. Concurrently, Alibaba unveiled the Qwen 3.7-Max large language model, capable of continuous operation for up to 35 hours. This multifaceted initiative, showcased at TechEx in Amsterdam, California, and London, signals Alibaba’s strategic investment in AI infrastructure and its anticipated leadership in the evolving technological landscape.

    đź’ˇInsights

    â–Ľ


    THE RISE OF ALIBABA’S AI CHIPS: A NEW COMPETITOR
    Alibaba has announced the Zhenwu M890, a new AI processor specifically designed for AI agents, marking a significant shift in the global AI hardware landscape. This launch is coupled with a multi-year silicon roadmap and a new large language model, signaling Alibaba’s ambition to create a fully integrated AI stack, strategically responding to potential export control limitations. The company’s focus is on building for the future demands of enterprise AI, rather than simply reacting to current trends.

    THE ZHENWU M890: ARCHITECTURAL INTENT
    The Zhenwu M890, developed by Alibaba’s semiconductor subsidiary T-Head, represents a substantial performance leap compared to its predecessor, the Zhenwu 810E, boasting three times the performance according to Reuters. However, the most important aspect of the M890 is its specialized architecture, tailored for the unique requirements of AI agents. These agents necessitate prolonged context retention, real-time coordination with other models, and the execution of complex, multi-step tasks with minimal human oversight. This demands significantly higher memory bandwidth and inter-model communication capabilities than standard inference chips are optimized for.

    A MULTI-YEAR SILICON ROADMAP: SUSTAINED INNOVATION
    Alongside the M890, Alibaba unveiled a comprehensive roadmap, including the V900 slated for release in the third quarter of 2027 and the J900 in the third quarter of 2028. This deliberate, sustained cadence of in-house silicon upgrades mirrors Nvidia’s “tick-tock” product cycles, demonstrating Alibaba’s commitment to maintaining a competitive edge in AI accelerator technology. The strategic planning highlights a long-term vision for technological leadership.

    CHINA’S SEMICONDUCTOR STRATEGY: MITIGATING RISKS
    The parallel between Alibaba’s approach and Huawei’s Ascend chip roadmap is notable. Both announcements reflect a shared conclusion among Chinese technology companies: reliance on foreign silicon, even with potential easing of export restrictions, represents an unacceptable structural risk. This has led to a strategic shift towards treating semiconductor development as a long-term capability-building exercise, rather than a simple procurement problem. Alibaba’s significant investment – over 380 billion yuan ($53 billion) – further underscores this commitment.

    PRODUCTION FOOTPRINT AND EARLY ADOPTION
    Alibaba has already demonstrated significant traction with the Zhenwu M890. Over 560,000 units have been shipped to date, with more than 400 external customers across 20 industries, including automotive and financial services. This substantial production footprint, coupled with real-world deployment data gathered at scale, provides Alibaba with invaluable insights before the M890’s full rollout. The deployment is facilitated through Alibaba Cloud’s Bailian platform and the Panjiu AL128 server system.

    QWEN 3.7-MAX: A MATCHING AI MODEL
    Complementing the M890 is Qwen 3.7-Max, Alibaba’s latest flagship large language model, specifically engineered for advanced coding and extended agent tasks. The model’s ability to operate continuously for up to 35 hours without performance degradation is a key differentiator, reflecting a design focused on autonomous operation and long-term agent functionality. This simultaneous release of the chip and model is a strategic platform play.

    A CLOSED-LOOP AI ECOSYSTEM
    Alibaba’s strategy is to create a closed-loop AI ecosystem, integrating its own silicon (T-Head), its own language model (Qwen), and its own cloud delivery platform (Bailian). This interconnected system is designed to reduce enterprise customers’ reliance on external vendors, fostering greater control and efficiency.

    GLOBAL EVENT AND MEDIA PARTNERSHIP
    The launch of the Zhenwu M890 and Qwen 3.7-Max coincides with the TechEx event in Amsterdam, California, and London, co-located with other leading technology events. AI News is powered by TechForge Media, offering additional resources and information on enterprise technology events and webinars.