DeepSeek is reportedly raising $1.5B ahead of a planned IPO that would value the Chinese AI company at around $71B. The company is also developing its own AI inference chips to reduce dependence on Nvidia GPUs.
The dual-track strategy puts DeepSeek in an elite group of AI companies pursuing vertical integration. Its in-house chip project, started about a year ago, focuses on inference workloads rather than training. The company has quietly recruited chip engineers through targeted outreach, building a team covering architecture, verification, and software enablement.
DeepSeek gained global attention with its open-source DeepSeek-V3 model and R1 reasoning model. Compute costs represent more than half of operating expenses for many AI companies, and DeepSeek’s custom silicon aims to dramatically lower that ratio.
The $71B valuation would make DeepSeek one of the most valuable AI companies globally. Chinese filings have previously implied a valuation around $52B, suggesting the company’s value has risen sharply amid growing demand for its models and services.
