What happened
NEAR AI reported containing an unauthorized access incident on its internal systems on November 10, 2026. The company temporarily suspended its NEAR AI Cloud service to investigate and replace compromised credentials. It subsequently confirmed the vulnerability was fixed across all servers and the service was restored. NEAR AI stated that the NEAR Protocol and other services were unaffected, though it has not yet confirmed whether customer data was accessed.
On November 10, 2026, NEAR AI disclosed via X that it had identified unauthorized access to its internal systems. In response, the company paused its NEAR AI Cloud service as a precautionary measure while conducting an investigation.
Subsequent updates indicated that NEAR AI had closed the unauthorized access and was completing its investigation. The company stated that it had fixed the vulnerability across its servers and replaced the affected credentials.
By 09:02 AM IST on the same day, NEAR AI confirmed that the NEAR AI Cloud service was back online. The company maintained that the NEAR Protocol and its other services were not affected by the incident.
NEAR AI has not yet published a detailed account of the incident, including how the access occurred, who was responsible, or which specific internal systems were reached. The company stated that a full incident report would follow.
Source details: cryptotimes.io
Why it matters
This incident highlights the security challenges facing AI infrastructure providers, particularly those relying on confidential computing. While NEAR AI uses hardware-isolated Trusted Execution Environments (TEEs) to protect customer prompts during , the breach of internal systems raises questions about the broader security perimeter. The distinction between protecting data in use within an enclave and securing the surrounding operational infrastructure is critical for enterprise adoption. Until a full incident report is published, the scope of the breach and the integrity of customer data remain unverified, creating uncertainty for users of the platform.
NEAR AI Cloud utilizes confidential technology, running AI models inside hardware-protected environments such as Intel Trust Domain Extensions (TDX) and NVIDIA confidential-computing tech. These Trusted Execution Environments (TEEs) are designed to prevent operators and infrastructure providers from accessing prompts and outputs.
However, the breach of internal systems raises questions about the limits of this privacy model. Hardware isolation protects data processed inside the enclave but does not inherently secure other internal systems, credentials, or operational data outside that specific environment.
The incident underscores the importance of distinguishing between data protection during and the security of the broader infrastructure. Until NEAR AI provides a detailed incident report, it is unclear whether the unauthorized access was limited to systems outside the protected inference environment or if it impacted customer data.
This event serves as a reminder for AI infrastructure providers to maintain robust security practices across all layers of their stack, not just within the confidential computing enclaves.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Why can ethical evaluation not be reduced to one model score?
What to watch next
Users should monitor for NEAR AI's forthcoming detailed incident report, which is expected to clarify the specific systems accessed, the nature of the vulnerability, and whether any customer data was exposed. Additionally, customers should review their own API key activity and consider rotating credentials as a precautionary measure, especially if they handle sensitive workloads.
The forthcoming incident report from NEAR AI is the most critical next step. It should detail the affected systems, the specific vulnerability exploited, the credentials involved, and whether any customer data was accessed or exposed.
Customers using NEAR AI Cloud should review their API-key activity and access logs. Rotating credentials is a recommended precaution, particularly for users handling sensitive workloads, although NEAR AI has not reported that customer API keys were compromised.
For NEAR holders and blockchain users, the company's assertion that the NEAR Protocol was unaffected remains its public assessment. Continued monitoring of official communications will be necessary to confirm the long-term stability and security of the protocol.