Latest Cover Story | Who Can Harness DeepSeek's Massive Traffic? A New Round of Fierce Competition in Internet Computing Power, Applications, and Services (AI Translation)

最新封面报道｜谁能接住DeepSeek泼天流量？互联网算力、应用、服务新一轮激战

[para. 1] The emergence of China's DeepSeek application has sparked a wave of competition in the AI industry, with a focus on AI computing power, applications, large-scale models, and cloud services. Internet companies quickly recognized the potential of DeepSeek's R1 inference model, prompting foundational computing service providers to launch compatible products and major model manufacturers in the U.S. and China to release benchmark models against DeepSeek-R1. WeChat's AI search feature, incorporating DeepSeek-R1, has created significant market attention, highlighting the challenges of supporting massive traffic demands in large user-base platforms like WeChat.

[para. 2][para. 3] WeChat's AI Search faced traffic overload due to massive computing power consumption, prompting Tencent to redirect queries to another app, "Yuanbao," impacting user experience but satisfying demand. Tencent released its self-developed Hunyuan large model and incorporated it into apps like WeChat Reading and WeChat Input, yet these AI features went unnoticed by users. DeepSeek's R1 inference model stood out by enhancing intention deconstruction and reasoning tasks, introducing "deep thinking."

[para. 4] OpenAI's release of the o1 series inference models marked a turning point in the AI sector, showcasing strong reasoning capabilities, prompting Chinese companies like Moon's Dark Side to follow suit. DeepSeek-R1, open-sourced with an MIT license, led to intense competition for AI application entry points. Tencent's announcement of integrating its self-developed T1 inference model with Yuanbao reinforced its strategy of upgrading and competing for new entry points.

[para. 5][para. 6] In response to DeepSeek-R1, Baidu quickly introduced "Deep Search," leveraging DeepSeek-R1 and the Wenxin large model to attract over ten million users within an hour. Collaborative office software from Alibaba and ByteDance integrated DeepSeek, although platforms like Taobao or Douyin haven't yet. AI-generated content models were reassessed, reshaping business strategies and promoting model capability iteration over user growth.

[para. 7][para. 8][para. 9] DeepSeek's open-source nature and rapid acceptance altered business models, enticing more exploratory usage, accelerating a new wave in AI. Major companies began iterating their models after seeing how DeepSeek emphasized performance advancements over user base expansion. Tencent quickly integrated DeepSeek-R1 across applications, splitting teams and bolstering AI integration, demonstrating competition's rapid pace.

[para. 10][para. 11] Baidu grappled with keeping performance high across mobile search apps and AI capabilities, as they saw user declines while aiming to follow OpenAI's commercial strategies. Simultaneously, ByteDance's Liang Rubo emphasized AI intelligence improvement, reflecting sector-wide priorities.

[para. 12] Largely absent from this consumer battle, Alibaba engaged in internal development, facing structural challenges. Its applications like Tongyi Qianwen showed underwhelming performance, and strategic shifts were indicated toward hardware and cloud computing investments.

[para. 13][para. 14][para. 15] Global model competition saw boosts in infrastructure investments, with companies racing to enhance computational capacity. Baidu's funding allocation following an acquisition stands as a commitment to this race. The rapid model iteration emphasized the speed of technological advancements, with major releases globally driving forward the fierce AI competition.

[para. 16] Open-sourcing efforts by companies like Baidu accelerated to match market trends, reflecting DeepSeek’s influence on reshaping model development strategies. Alibaba Cloud's open-source ventures paralleled those of DeepSeek, indicating a sharing path.

[para. 17] DeepSeek's open-source launch has generated unprecedented application demand, notably in government services like Shenzhen's AI administrative deployments. Infrastructure investments are rapidly adjusting to support this demand, heralding a shift toward inference computing resource deployment.

[para. 18] DeepSeek reignited investor interest in China's tech sector, emphasizing its significant impact on valuation logic. Its open-source approach presents challenges and opportunities for existing models, potentially reshaping investment strategies and market narratives.

[para. 19] The narrative of technology-driven growth over user base expansion becomes paramount, affecting startups' valuation and strategic directions. AI infrastructure investments, accelerated by DeepSeek's demand, gained traction as a more widespread sector opportunity.

[para. 20] Termed by some as China’s “Sputnik moment,” DeepSeek's emergence emphasized China's expanding technological capabilities, reshaping both primary and secondary market sentiments. This refocused investment narratives, bridging technological development and financial growth prospects in AI.