Alibaba Cloud

Jan 29, 2025

Why Alibaba’s New “Better than DeepSeek” AI Model Did Not Impact US Stocks Recent releases of AI models from Chinese companies DeepSeek and Alibaba demonstrate growing competition in AI development while highlighting challenges in evaluating technical claims. Published on January 29, 2025 06:39 AM EST Neither the author, Tim Fries, nor this website, The Tokenist, provide financial advice. Please consult our  website policy  prior to making financial decisions. The artificial intelligence landscape saw major developments from Chinese companies this week, with Alibaba’s (NYSE: BABA) release of its Qwen2.5-Max model following DeepSeek’s emergence as a leading AI provider , marking increased competition in a field traditionally dominated by U.S. firms. Alibaba Announces New Model Claimed to be “Better than DeepSeek” Alibaba announced its Qwen2.5-Max model on Sunday, claiming the large-scale AI system has been trained on over 20 trillion tokens. The company said the model is available through Alibaba Cloud’s API services but provided limited technical details about its development process or training methods. While Alibaba published benchmark results comparing Qwen2.5-Max to other publicly available models, it notably excluded comparisons with proprietary systems from major U.S. companies like OpenAI and Anthropic, making it difficult to assess its relative capabilities. The announcement comes days after DeepSeek’s AI model gained prominence by becoming the top-rated free app in Apple’s U.S. App Store, triggering significant market reactions. DeepSeek’s open-source approach and claims of superior computational efficiency – requiring only about 2,048 GPU chips compared to ChatGPT’s estimated 20,000 – led to widespread reassessment of AI infrastructure investments, which also lead to a selloff in energy stocks like Vistra (NYSE: VST). Join our Telegram group and never miss a breaking digital asset story. Why Alibaba’s Announcement Did Not Impact US Stocks DeepSeek’s public release of its training methods and model architecture has allowed independent verification of its claims, with researchers noting its innovative reinforcement learning algorithms and optimization techniques. The company’s transparency has earned praise from industry experts, including Meta’s AI chief Yann LeCun. The contrasting approaches between Alibaba and DeepSeek highlight the challenges in evaluating AI advancement claims. While Alibaba maintains traditional corporate secrecy around its development process, DeepSeek’s open-source strategy allows for public scrutiny but risks revealing competitive advantages. The rapid pace of innovation, combined with varying levels of technical disclosure and the complexity of AI systems, makes it challenging to provide accurate comparisons or investment guidance. This was evident in recent market volatility, where speculation about AI efficiency gains led to significant price swings in technology and energy stocks before technical claims could be fully verified. At the time of writing, US-listed shares of Alibaba (NYSE: BABA) are up 3.13% in premarket trading at the time of writing. Nvidia has recorded a 1.04% decline in premarket today after gaining 8.93% over the last trading day. Disclaimer: The author does not hold or have a position in any securities discussed in the article. All stock prices were quoted at the time of writing.