Chinese AI company DeepSeek is developing its own custom AI chip, according to a report from Reuters.Citing three people familiar with the matter, the chip – which is being designed to support inference workloads – will allow the company to reduce its reliance on Nvidia and Huawei hardware.The report said that DeepSeek’s development efforts started about a year ago, but noted that the process was still in the early stages. Thus far, the company has entered into discussions with external partners across design, foundry, and memory.Reuters further stated that the AI company has hired multiple chip design engineers in recent months, but has not been publicly posting job listings as the project is still under wraps.The company did not respond to Reuters' request for comment.DeepSeek rose to prominence in early 2025 when the AI company sent waves through the US stock market with the release of its open-source LLM, which offers comparable performance to the systems produced by top US AI labs, but which was reportedly trained using a fraction of the compute power.Questions were immediately raised about the validity of DeepSeek's claims, with OpenAI and Microsoft looking into whether DeepSeek used OpenAI’s API to integrate OpenAI’s AI models into DeepSeek’s own models, and US officials reportedly investigating whether the company actually acquired more powerful Nvidia GPUs than it said it did via third-party companies in other countries.Despite this, major cloud companies in the US and China alike brought DeepSeek models to their cloud platforms.Should the company move ahead with the development of its own chip, it would become the latest in a growing line of companies opting to produce custom hardware.Last month, OpenAI unveiled its Jalapeño Intelligence Processor, co-designed with Broadcom, and, in April 2026, it was reported that Anthropic was also exploring the possibility of designing its own AI chips. Amazon, Google, Meta, and Microsoft have all developed custom chips.