NEAR AI has restored its Cloud service after an attacker exploited a software vulnerability to access internal systems, exposing information belonging to approximately 100 users and data from 18 AI agents. Summary NEAR AI restored its Cloud service after fixing an SGLang vulnerability exploited on November 8. Attackers accessed account information belonging to approximately 100 users, including email addresses and names. Eighteen legacy AI agents had chat histories, settings, and logs accessed during the security incident. Two customers had Cloud API keys exposed through logs, and NEAR AI revoked both credentials. NEAR AI said its blockchain protocol remained unaffected, with no established connection to Binance’s hack. According to the company’s security report, published November 9, the attacker gained access through SGLang, an open-source inference engine used on its servers. The company said it had fixed the vulnerability, replaced affected credentials, and completed a forensic review. NEAR Protocol and other NEAR services were not affected. Update: On November 8, an attacker used a vulnerability in SGLang, the inference engine we run, to access part of our infrastructure. We’ve closed the access and fixed it.Affected users have been contacted. If you haven’t heard from us, no action is needed.…— NEAR AI (@near_ai) November 10, 2026 NEAR AI breach exposed account details and legacy agents The incident began on November 8, when an attacker exploited a remote code execution vulnerability in SGLang. The flaw allowed the unauthorized party to execute code on one inference server before accessing other parts of the company’s infrastructure. Investigators found that the attacker subsequently reached an internal system connected to NEAR AI’s previously discontinued Private Chat service. From there, the unauthorized party downloaded account information belonging to approximately 100 users. The exposed records included email addresses and names, alongside subscription plan and account balance details for some users. The company reported no evidence that stored Private Chat conversations or uploaded files were accessed. However, investigators confirmed unauthorized access to 18 agents operating on an older version of NEAR AI Agents. Those deployments were not running inside the company’s trusted execution environment, or TEE, which provides hardware-based isolation for confidential computing. Information retrieved from the affected agents included chat histories, configuration settings, and operational logs. Two NEAR AI Cloud customers had API keys exposed through those logs, and the company subsequently revoked both credentials. The investigation further established that the attacker could have observed requests processed on the compromised inference server on November 8. This included requests running in confidential mode, although the report did not specify how many requests were observed. NEAR AI Cloud returns after security fixes NEAR AI initially suspended Cloud inference operations after detecting unauthorized access to its systems. The company subsequently confirmed that it had closed the attacker’s access but kept Cloud operations paused while completing its security work. In a later update, the company confirmed that its Cloud infrastructure had returned online after engineers deployed fixes across its servers and replaced affected credentials. The response included revoking active sessions and administrative credentials, replacing agent tokens, and taking Private Chat offline. Engineers applied the SGLang security fix across all inference servers before completing the remaining credential rotations. NEAR AI said it had finished a full forensic review and contacted users whose accounts or agents were confirmed to have been accessed. “If you haven’t heard from us, no action is needed,” the company stated in its latest public update. Affected customers received instructions from [email protected], while owners of the 18 compromised agents received individual notifications. The company advised users to treat unsolicited messages referencing their account information or balances with caution. It said regulatory notifications would follow where required. Crypto.news reached out to NEAR AI for comments but received no response by publication time. SGLang vulnerability affected confidential AI processing The security report identified a limitation involving the software running inside NEAR AI’s confidential computing environment. The company uses trusted execution environments to isolate sensitive information during AI processing. Under normal operating conditions, the technology prevents infrastructure operators and underlying hosts from accessing protected data. However, the SGLang vulnerability allowed the attacker to execute code inside an inference server, potentially exposing requests while the software processed them. In its report, NEAR AI explained that confidential computing protects against access from outside the protected environment but cannot prevent vulnerabilities within the applications running inside it. The company has reported the vulnerability to the SGLang project and said its security reviews will now cover inference software, related services, dependencies, and deployment configurations. NEAR AI’s confidential computing operations have continued expanding through AI inference products and blockchain-based payment services. In October, the platform surpassed one million NEAR staked through a service that allows users to fund AI inference using staking rewards. Its documentation describes confidential computing as a feature of supported hosted models. NEAR co-founder Illia Polosukhin had warned about software vulnerabilities in October, citing weaknesses across internet infrastructure and calling for stronger AI-assisted security systems. NEAR Protocol unaffected amid separate Binance laundering activity NEAR AI maintained that the November incident did not affect the underlying NEAR Protocol blockchain or other services within its ecosystem. The incident occurred during the same week that investigators identified NEAR-based tokens being used to move funds associated with the Binance exchange hack. According to a November 9 Elliptic investigation, attackers created six temporary memecoins and used decentralized liquidity pools to transfer approximately $430,000 in stolen cryptocurrency. The researchers identified roughly 117,000 NEAR routed through pools involving tokens created between November 5 and November 7. Elliptic attributed the laundering activity to operators associated with the Binance theft. Neither its findings nor NEAR AI’s incident report established a connection between the two security events. NEAR AI said it had completed its credential replacement process and reported the SGLang vulnerability to the project’s maintainers.
