DeepSeek reportedly did not share its next AI model with US engineers and instead granted early access to Chinese companies, further escalating the US-China tech war, starting February 26, 2026.
Nuphoto | Nuphoto | Getty Images
Chinese artificial intelligence startup DeepSeek released a preliminary version of its highly anticipated V4 big language model on Friday, allowing users to test its new capabilities and features.
This release comes more than a year after the Hangzhou-based company presented its R1 reasoning modelwhich shook the global technology markets due to its surprising performance and profitability.
Similar to DeepSeek’s previous V3 model, the latest upgrade is open source, allowing developers to download the code, run it locally, and modify it.
The model is available in ‘pro’ and ‘flash’ versions, depending on size, with DeepSeek claiming that V4 performs strongly against domestic competitors, particularly in agent-based tasks, knowledge processing and inference.
“DeepSeek’s V4 preview is a serious solution,” offering lower inference costs than previous models, Neil Shah, vice president of research at Counterpoint Research, told CNBC.
Inference costs refer to the IT and financial expenses of running a trained AI model to generate results.
DeepSeek also said that V4 has been optimized for use with popular agent tools such as Claude Code and Anthropic’s OpenClaw.
According to Wei Sun, Counterpoint’s principal AI analyst, V4’s baseline profile suggests it could offer “excellent agent capabilities at a significantly lower cost.”
Will DeepSeek shock the world again?Founded in 2023, DeepSeek gained attention in late 2024 with its free and open source V3 model, which it said was trained with less powerful chips and at a fraction of the cost of models built by OpenAI and Google.
A few weeks later, in January 2025, the company released a reasoning model, R1, which met similar benchmarks or outperformed many of the world’s leading LLMs.
The R1 model had alarmed investors when DeepSeek revealed that it only took two months, and not even 6 million dollarsto build the model using lower capacity Nvidia chips. This calls into question America’s lead in AI as well as big tech’s massive spending on AI infrastructure.
Since then, DeepSeek has released a series of model upgrades, but none have matched the impact of the R1.
The V4’s debut is unlikely to have the same market impact as the R1 because traders have already internalized the fact that Chinese AI is competitive and cheaper to use, Ivan Su, senior equity analyst at Morningstar, told CNBC.
However, DeepSeek’s latest positioning places other Chinese open source models as direct competitors, Su said.
“This is a framing that did not exist with R1, and this alone shows how national competition has intensified,” he added.
Since the release of R1, DeepSeek has faced increased competition in China’s booming AI sector, with players like Alibaba and ByteDance is also launching new models this year.
Shares of several other Chinese AI players were lower in Hong Kong on Friday. MiniMax and Knowledge Atlas Technology, also known as Zhipu, each fell about 8%, while Hangzhou-based developer Manycore Tech plunged 9%.
What chips drove the V4?A major question surrounding the release of DeepSeek’s V4 model is what chips were used to train and support it.
Chinese tech giant Huawei confirmed on Friday that its latest AI computing cluster, powered by its Ascend AI processors, can support DeepSeek’s V4 model.
However, it is unclear to what extent Huawei’s chips were used in training, compared to Nvidia’s.
Chinese developers are barred from directly purchasing Nvidia’s most advanced AI chips due to U.S. export controls.
Meanwhile, Beijing has stepped up efforts to develop its domestic chip industry and would have has pushed Chinese technology companies to adopt domestic alternatives from chipmakers such as Huawei, rather than foreign alternatives.
Counterpoint’s Wei Sun said the V4’s ability to run natively on local chips could have massive implications, helping Beijing achieve greater AI sovereignty and further reduce its reliance on Nvidia.
“This will also ultimately accelerate global developments in AI,” she added.
After DeepSeek announced the release of its V4, shares of Chinese contract chipmakers rose in Hong Kong, with SMIC And Hua Hong Semiconductor up 8.9% and 15.2%, respectively.