DeepSeek reports 475 million yuan revenue in first half of 2026, seeks 5 billion yuan in second funding round

DeepSeek, a Chinese AI research lab, has reported revenue of approximately 475 million yuan (around 707 million US dollars) for the first seven months of 2026, a tenfold increase compared to its full-year revenue in 2025. The company also noted a reduction in its net loss, which dropped from 935 million yuan in 2025 to 715 million yuan in the same period. The financial performance has fueled confidence in the company's growth trajectory, with plans to pursue an initial public offering (IPO) in 2027. The company recently completed its first round of funding, raising 5 billion yuan, and is now preparing a second round aimed at securing an additional 5 billion yuan. The target valuation for the second funding round is set at 50 billion yuan. Investment banks have already been hired to support the IPO process. DeepSeek's strong financial performance is attributed to the widespread adoption of its open-source models, particularly the V4 series, with the lightweight version V4-Flash gaining popularity among global developers due to its cost-effectiveness and performance. The company's gross margin for the first seven months of 2026 reached 44.6%, with API call revenue contributing a gross margin of 82.9%, significantly higher than OpenAI's 39% and Anthropic's estimated 63%. In early July, DeepSeek raised its API pricing, with the V4-Pro model now priced at 3.96 dollars per million tokens during peak hours and 1.98 dollars during off-peak hours, a substantial increase from the previous rate of 0.87 dollars. This pricing strategy, combined with its growing revenue and gross profit margins, has positioned DeepSeek to scale its commercial narrative, leveraging its open-source foundation to deliver high-performance solutions at a lower cost.