China-linked hackers accounted for more than 58% of state-sponsored targeted intrusions against technology companies, according to CrowdStrike’s 2026 Technology Threat Landscape Report released June 9. The cybersecurity firm said the activity appears driven by interest in technology development, intellectual property, downstream customer access, and AI capabilities that are likely high-value targets.
The report lands after an April White House memo accused foreign entities, principally based in China, of industrial-scale efforts to distill US frontier AI systems. For enterprise technology leaders, the operational concern is direct: AI systems, model APIs, cloud identities, and developer tools now sit inside the attack surface adversaries are probing.
Why China-nexus hackers are targeting tech companies
CrowdStrike’s 2026 Technology Threat Landscape Report named five China-nexus adversaries — MURKY PANDA, MUSTANG PANDA, OVERCAST PANDA, SUNRISE PANDA, and WARP PANDA — that targeted the technology sector more than any other industry. MURKY PANDA also conducted password-spraying attacks against more than 340 primarily US-based organizations across sectors, with technology firms among the most affected.
Crowdstrike tied the targeting to China’s push for technological self-sufficiency, technology development, intellectual property, and intelligence collection goals. The risk extends beyond the first victim because China-nexus adversaries also seek access to downstream customer environments.
Separate findings in the 2026 Global Threat Report showed that China-nexus activity increased 38% in 2025, reinforcing the need to treat AI models, training infrastructure, model-serving endpoints, developer environments, and software supply chains as high-value systems.
How AI widens the attack surface
Direct network intrusion is only one path into AI-related systems. Model APIs can be abused for extraction attempts, developer repositories can be compromised through malicious dependencies, and stolen credentials can open access to cloud systems that support AI workloads.
The report also described non-China activity against developer ecosystems. An unknown actor operating Glassworm malware compromised 350 GitHub repositories, and DPRK-nexus STARDUST CHOLLIMA compromised the Axios npm package, which is downloaded about 100 million times per week. Initial access brokers advertised access to 277 technology companies, a nearly 30% year-over-year increase.
AI tools are also becoming targets. The Global Threat Report said adversaries injected malicious prompts into legitimate GenAI tools at more than 90 organizations, abused AI development platforms to establish persistence and deploy ransomware, and used AI-generated scripts to accelerate credential dumping and erase forensic traces. Attackers have also abused real ChatGPT shared links to make fake software downloads look more credible.
Cloud exposure adds another pressure point. Cloud-conscious intrusions rose 37% overall in 2025, including a 266% increase from state-nexus actors targeting cloud environments for intelligence collection. Average eCrime breakout time fell to 29 minutes, and the fastest observed breakout occurred in 27 seconds. Research into an AI-powered worm points to the same pressure on response windows.
Security teams should focus on model-access controls, API monitoring for unusual extraction patterns, and audits of open-source dependencies, especially as AI systems surface large volumes of open-source security flaws faster than maintainers can patch them.
In an April 23 memo, White House science adviser Michael Kratsios accused foreign entities “principally based in China” of industrial-scale efforts to distill US AI systems. China rejected the allegations, and AP reported that the House Foreign Affairs Committee backed legislation to identify and punish foreign actors that extract key technical features from closed-source US AI models.
AI infrastructure, developer tooling, cloud identity, and software supply chains are now core enterprise assets. Treating them as secondary systems leaves technology companies exposed to espionage and downstream compromise.
Also read: OpenAI’s Daybreak initiative shows how AI agents are moving deeper into vulnerability detection, threat modeling, and secure code review.