In response, China’s DeepSeek has released a new advanced open source artificial intelligence model. alphabet(GOOGL) Google rolls out its own Gemini 3. Nvidia (NVDA) and the cloud hyperscaler were sold in January 2025 after DeepSeek introduced a low-cost open-source AI model.
Chinese companies continue to promote open source AI models, aiming to gain global support. Open source model functionality is provided free of charge to developers.
“DeepSeek and its Chinese peers are no longer the underdogs,” Gradient Flow newsletter editor Ben Lorica told IBD. “Chinese developers now account for the majority of new open-model downloads, and some VC partners estimate that around 80% of new AI startup pitches are built on Chinese open source stacks. As developer mindshare shifts east toward 2026, US players are playing catch-up.”
Lorica added: “America’s counterattack is shaped by transparency and architecture, not scale. America’s counterattack strategy relies on the full-stack transparency of AI2 (Allen Institute for AI Research)’s Olmo 3 and NVIDIA’s Nemotron, as well as the architectural novelty of companies like Arcee, aiming to provide a quality alternative to the full proliferation of China’s open-weight model.”
Anyone can reuse and build on open source large-scale language models. As a result, more efficient open source models are emerging that reduce the computing power required to “train” the LLM, essentially feeding it with data.
Other Chinese companies building on open source models include tech giants Baidu (BIDU) and alibaba (BABA), as well as Zhipu AI, MiniMax, and Moonshot AI.
Meta loses ground in open source AI battle
meta platform (META) is a leading proponent of the open source model. But observers say Chinese companies have overtaken Meta’s open source model.
January’s “Deep Seek Moment” firmly demonstrated China’s AI technology competitiveness.
The new Deepseek AI model is reported to have superior scores in mathematical reasoning. The new DeepSeek model is open sourced on Hugging Face and GitHub.
Google released its latest artificial intelligence system “Gemini 3” on November 16th. In the performance review, it will be pitted against OpenAI’s GPT-5 and Anthropic’s Claude family. With Gemini 3, Google has improved the chatbot’s coding, search, and image creation capabilities.
Both the proprietary AI model camp (OpenAI, Google, Anthropic) and their open source rivals are focused on building AI agents that are autonomous and perform actions.
Investor concerns about the potential for disruption to Google’s core internet search business began when OpenAI rolled out ChatGPT in late 2022. While ChatGPT provides answers to search queries, Google’s business model is based on providing web links.
Google’s stock price will soar in 2025
But Google’s stock price rose 68% in 2025 as the company’s artificial intelligence strategy took off. Alphabet is fully leveraging its artificial intelligence capabilities, including internet search, cloud computing, digital advertising, self-driving car developer Waymo, YouTube, and apps like Gmail, Workspace, and Maps.
Meanwhile, Nvidia stock plummeted 17% after DeepSeek released its R-1 AI model in late January. However, NVIDIA stock eventually rebounded, rising 30% in 2025. But NVIDIA stock and other AI businesses like cloud hyperscalers are volatile.
cloud hyperscaler Amazon.com (AMZN), microsoft (MSFT) and Google have spent tens of billions of dollars on new AI data centers. DeepSeek’s progress raises questions about the computing power needed to develop AI systems, a key driver of AI stocks.
Pricing for Chinese open source models
“China’s AI API prices are among the lowest in the world,” Jefferies analyst Edison Li said in a Nov. 11 report. “DeepSeek recently reduced the price of its API by 63% citing improved efficiency. The lower price of AI in China could more easily drive application development and user adoption. We see Chinese AI as having the potential for a better return on investment, even if China’s inference is supported by suboptimal semiconductors.”
meanwhile, airbnb(ABNB) CEO Brian Chesky made a splash in October when he said the company had chosen Alibaba’s Qwen open source model over OpenAI’s ChatGPT.
In November, Databricks co-founder Andy Konwinski said the United States needs to promote open source AI models to compete with China globally.
Follow Reinhard Klaus on X (formerly Twitter). @reinhardtk_tech Get the latest information on artificial intelligence, cybersecurity, and cloud computing.
You may also like:
Want to trade options? Try these strategies
Watch companies reaching new buy points on IBD’s Breakouts of the Day list
IBD Digital: Get access to IBD’s premium stock lists, tools and analysis now
Build a custom stock price screen with MarketSurge
Futures prices rise as AI giant Nvidia’s performance soars
