TL;DR

  • DeepSeek’s January 2025 release of R1 shook markets - a frontier-grade reasoning model trained for a reported $6M, a fraction of US lab budgets
  • The app shot to #1 on Apple’s App Store inside days, and the open weights forced an industry-wide rethink of what training really costs
  • Subsequent releases (V3 and beyond) cemented DeepSeek as a serious competitor in the open-source and cost-efficient AI category
  • The story is less “China caught up” and more “the cost floor moved” - implications for closed-model pricing, GPU demand, and open-weight strategy
  • Worth understanding as the moment that made cheap, capable, open models a credible default rather than a curiosity

Overview

Wow, crazy times, the best technology in the world is now becoming incredibly cheap and accessible to everyone! 🤯

DeepSeek’s AI chatbot has become the top free app on Apple’s store since its January launch. This rapid success, fuelled by DeepSeek’s cost-effectiveness compared to US competitors, has sent shockwaves through financial markets.

Venture capitalist Marc Andreessen has lauded DeepSeek’s AI as a “groundbreaking achievement”, while the company claims its models rival industry leaders like ChatGPT at a fraction of the cost. DeepSeek reportedly invested only $6 million in its development, a stark contrast to the billions spent by US AI companies.

DeepSeek had already shipped its V3 base model in late December 2024, but it was the R1 reasoning release in January 2025 that broke into the mainstream. The lab has continued iterating since - V3.1, V3.2, and further reasoning-tuned variants - and has cemented its place as one of the most credible open-weight competitors to the US frontier labs.

YouTube

[2025-01-27] OpenAI is Done, China Won (Deepseek Explained)

[2025-01-26] I Did 5 DeepSeek-R1 Experiments | Better Than OpenAI o1?