DeepSeek raising capital at ~$74B valuation, founder says AGI over profit

DeepSeek is simultaneously raising money, teasing a public listing, and making a philosophical stand. Reuters reported the firm is raising fresh capital at roughly a $74B valuation — up sharply from $50B in June — targeting up to 50 billion yuan ahead of a possible Shanghai STAR Market IPO filing this year. That would make it one of China's most valuable AI companies and a rare frontier lab pursuing public markets.

Founder Liang Wenfeng, per Yicai/US News reporting, said DeepSeek's focus is long-term AGI rather than near-term profit maximization, and signaled the company will likely keep its top models open-source, arguing open-source development and commercial monetization are not mutually exclusive. He identified compute as the primary constraint against US competitors and said DeepSeek plans to build its own large AI computing clusters.

Operationally, there's a hard edge: a July 24 15:59 UTC cutoff retires the legacy deepseek-chat and deepseek-reasoner model IDs, forcing an API migration — a small but real disruption for integrators today.

The broader context is the week's 'Chinese open-weight resurgence' subplot: Moonshot's Kimi K3 is being framed as another 'Sputnik moment,' US startup founders are lobbying the Trump administration not to shut off Chinese open-weight models (759 HN points), and China is reportedly weighing tighter export controls on model weights for firms like Alibaba and ByteDance. DeepSeek's open-source-plus-IPO posture sits right at that fault line. What to watch: whether the STAR Market filing actually materializes, and how export-control politics on both sides affect DeepSeek's open-weight commitments.