AI Threat Watch

Case deepseek-ddos-registration-haltLLMjackingChina

DeepSeek halts new sign-ups after month-long DDoS and brute-force campaign peaks

DeepSeek said its services were under large-scale malicious attack on 28 January 2025 and limited sign-ups to Chinese mobile numbers; Qi An Xin XLab traced a month of escalating DDoS and brute-force traffic.

HighAttack on AIStatus: ResolvedVerified case file

Open in the interactive mapCite this case

Reported
Jan 28, 2025
Location
China
Target
DeepSeek chat and API services
Severity
High
Triggered by
Unknown
Actor type
Unknown
Reported loss
None reported
CVE
None

What happened

DeepSeek said its services were under large-scale malicious attack on 28 January 2025 and limited sign-ups to Chinese mobile numbers; Qi An Xin XLab traced a month of escalating DDoS and brute-force traffic.

Impact: Hours of outages and degraded API performance for users worldwide; new registrations blocked for non-Chinese phone numbers; existing users could still log in.

Who triggered it

Actor: Unknown

Attribution: Qi An Xin XLab said attack and brute-force IPs were located in the United States; no group was named and DeepSeek did not attribute the attack.

Motive: Unknown; analysts suggested disruption of the newly popular R1 service.

How it happened

Attack vector: Three phases: HTTP proxy floods from 3 January, SSDP and NTP reflection amplification from 20 January, then application-layer HTTP proxy attacks and password brute-forcing peaking 27 to 28 January.

Root cause: Sudden global demand for DeepSeek R1 outpaced the startup's DDoS mitigation and account-protection capacity.

Timeline

  1. XLab observes first suspected HTTP proxy attacks against DeepSeek
  2. Application-layer attacks surge; DeepSeek reports degraded web and API performance
  3. DeepSeek status page cites large-scale malicious attack and restricts registration to +86 numbers
  4. The Paper publishes Qi An Xin XLab analysis of the three attack phases

Response

DeepSeek switched IP addresses at 00:58 on 28 January, restricted sign-ups and restored service in stages; no user-data theft was confirmed.

Current status: Resolved

Sources (open in a new tab)

Cite this case

DeepSeek halts new sign-ups after month-long DDoS and brute-force campaign peaks. AI Threat Watch, case deepseek-ddos-registration-halt, Jan 28, 2025. https://aithreat.io/cases/deepseek-ddos-registration-halt/

Permanent link: https://aithreat.io/cases/deepseek-ddos-registration-halt/

This case file in other languages: EnglishFrançaisEspañolالعربية中文