Chinese AI start-up DeepSeek, which had released its R1 reasoning model the previous week, overtook ChatGPT in downloads from Apple’s US App Store, prompting a sharp global sell-off in AI-related shares. Nvidia shares fell almost 17% and the company lost close to US$600bn in market value in a single day, a record for a US company.
What happened
- DeepSeek said R1 performed comparably to leading US models on several benchmarks and made it openly available under a permissive licence.
- Its claim that its earlier V3 model had been trained for less than US$6m in computing costs challenged assumptions about the spending needed for frontier AI, although analysts questioned the full cost.
- On 30 January Italy’s data protection authority, the Garante, ordered DeepSeek to block its app in Italy with immediate effect over how it handled personal data, and other regulators raised concerns.
- In early February Australia, Taiwan, and South Korea restricted its use on government devices.
Why it mattered
The episode challenged assumptions about US leadership in AI and showed how quickly a new model can raise data protection and national security questions.
Lessons for organisations
Before staff use a new AI tool, check where data is stored and processed, what the provider’s terms allow, and whether it meets your data protection obligations. An approved-tools list and clear AI use policy reduce the risk of sensitive data leaking.
Sources: Reuters (via Yahoo Finance), The Hacker News
Part of our Top stories archive of headline-making events in information security, privacy, and AI. If you would like help applying the lessons to your organisation, contact us.
