Global AI model call volumes declined to 206 trillion tokens last week. China's calls dropped to 444.1 trillion, while U.S. volumes rose to 490.8 trillion, surpassing China's for the first time in two months. Four of the top nine models are Chinese, with DeepSeek V3.2 second in calls.

DeepSeek, a Chinese AI startup, is negotiating its first external funding round, aiming for at least $300 million at a $10 billion valuation. Despite previous rejections of investment offers, its fundraising discussions are now reportedly underway.

DeepSeek seeks to raise $300 million at a $10 billion valuation, significantly lower than leading U.S. AI firms like OpenAI and Anthropic. Their R1 model's low training cost disrupted the market, leading to investor concerns about valuation disparities and revenue potential.

Gate News message, April 17 — According to The Information, AI company DeepSeek is conducting its first external financing round since inception, with a target valuation exceeding $10 billion.

Bytedance denied reports of researcher Guo Daya joining its Seed team with a 100 million yuan salary. Guo confirmed his role as an Agent lead, but salary details are undisclosed, while compensation structures for the team are uniform.

ByteDance has denied reports of offering 100 million yuan to recruit ex-researcher Guo Daya. VP Li Liang confirmed that all technical staff have the same compensation structure and recent hires did not exceed 100 million yuan, although options might yield significant returns over time.

Gate News message, April 10, according to multiple people familiar with the matter, in recent internal communications DeepSeek’s founder, Liang Wenfeng, of AI large model company DeepSeek, revealed that DeepSeek’s next-generation flagship large model, DeepSeek V4, will be officially released in late April.

Gate News message, April 8, DeepSeek’s website and app launched two new modes early this morning. Among them, the Quick mode is for everyday conversation scenarios, providing instant responses; the Expert mode is designed for complex problems and supports deep-thinking functionality, but it requires waiting during peak hours and does not support attachment or voice upload.

DeepSeek experienced three service disruptions from March 29 to 31, with durations of 1 hour 48 minutes, 10 hours 13 minutes, and 1 hour 3 minutes, respectively. The issue has been resolved, and the service is now operating stably again. Over the past 30 days, the web chat service availability has been 98.61%. This incident has raised concerns about the stability of AI infrastructure.

DeepSeek is recruiting for 17 positions, with R&D focus shifting toward Agent productization. Positions include algorithm research, data evaluation, and infrastructure engineers, focusing on reinforcement learning and evaluation system design. The company also values experience with AI coding tools. Compared to the deep learning researcher positions from the beginning of the year, this recruitment clearly tilts toward Agent productization.