AiGpu

Industry·

Supply Chain Attack via Optimizely Infects Hundreds of Websites

Researchers uncovered a massive malware campaign that abused Optimizely’s legitimate JavaScript delivery to inject crypto theft, card skimming, and iOS exploit kits into visitor browsers.

Illustration of malicious JavaScript injection alert on a web page

Security researchers have exposed a large‑scale supply chain attack that leveraged the Optimizely website personalization platform to serve malicious code to visitors of more than 178 sites, including major brands such as CNN, Zoom, the BBC, and Salesforce.

Technical Overview

The attackers compromised Optimizely’s signed JavaScript snippets, inserting four malicious components that activated simultaneously. These payloads performed crypto‑wallet address swapping for nine currencies, skimmed payment card data via HTTP requests, harvested cookies and screenshots, and delivered iOS exploit kits (Coruna and DarkSword) to iPhones and iPads.

Because the malicious scripts were served from cdn.optimizely.com with a valid certificate, the attack bypassed many traditional CDN‑cache poisoning defenses. The exact method of gaining write access to multiple unrelated customer projects remains unknown, highlighting a critical gap in third‑party code governance.

Why It Matters for GPU / AI Infrastructure

Cloud‑based GPU and AI workloads often rely on similar JavaScript or container delivery pipelines. A compromised supply chain can inject malicious code into inference services, steal model credentials, or hijack compute resources for crypto mining. Securing the software stack—from code signing to runtime monitoring—is essential to protect expensive GPU resources and maintain trust in AI services.

Enterprises should enforce least‑privilege access for CDN credentials, implement strict Content Security Policies, monitor third‑party script integrity, and adopt hardware‑rooted trust mechanisms for AI inference pipelines.

Key Takeaways

  • Legitimate platform code can be weaponized, bypassing certificate checks.
  • Multi‑stage payloads target both web users and mobile devices.
  • Supply chain attacks pose a direct risk to cloud GPU/AI workloads.
  • Continuous verification and strict script governance are critical defenses.
  • aigpu
  • ai gpu
  • ai gpu cloud
  • aigpu dubai
  • supply chain security
  • cdn attack
  • web infrastructure

By AiGpu Editorial · Editorial rewrite based on public reporting (Ars Technica)

← All articles