×
DeepSeek AI’s cost and privacy claims examined
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

The Chinese AI startup DeepSeek has released R1, an open-source AI model that reportedly outperforms GPT-4 while costing significantly less to develop.

Key developments: DeepSeek’s R1 model has emerged as a significant player in the AI landscape, climbing to third place on HuggingFace‘s Chatbot Arena rankings and gaining prominence through its AI assistant application.

Technical specifications and accessibility: The R1 model combines impressive performance metrics with cost-effective deployment options.

  • DeepSeek reports spending less than $6 million on training the model
  • API access is priced at $0.14 per million tokens, substantially lower than competitors like OpenAI
  • The open-source nature of R1 allows for unrestricted downloads and usage
  • Training dataset details remain undisclosed

Data privacy considerations: DeepSeek’s operations in China raise important security and privacy considerations.

  • User data, including IP addresses, device IDs, and chat histories, is stored on Chinese servers
  • Chinese cybersecurity laws could potentially grant government access to stored data
  • The open-source architecture provides some transparency and allows for local deployment of smaller versions
  • Unlike TikTok, DeepSeek maintains transparency about its data storage practices in China

Market implications and competitive landscape: R1’s emergence could reshape the AI industry’s power dynamics and investment patterns.

  • The success of a lower-cost, high-performing model could enable smaller laboratories to develop competitive AI solutions
  • This development occurs during a sensitive period in US-China technology relations
  • The competitive pricing structure could pressure established players to reconsider their pricing models

Future trajectory: While DeepSeek’s technological achievements are impressive, questions remain about data security and international market access, particularly given increasing scrutiny of Chinese technology companies in Western markets.

What to know about DeepSeek AI, from cost claims to data privacy

Recent News

SoftBank teams with Quantinuum to advance quantum computing applications

Early quantum-classical hybrid systems will be tested in Japan to tackle network optimization and materials science challenges that exceed classical computing limits.

Ex-Google, Apple engineers unveil Oumi AI, a truly open-source AI development platform

Former Google engineers release fully transparent AI development tools that allow unrestricted access to model code and training data.

Google thwarts hacker group using Gemini to breach accounts

State-sponsored hacking groups from Iran, China, and North Korea attempted to use Gemini for cyberattacks but were blocked by Google's security controls.