Experts warn hackers could shut down entire power grids by hijacking cloud accounts

Zhejiang University researchers warned that GPU workloads could destabilize local networks and cause blackouts
Attackers could exploit ~1000 GPUs to drain power and generate excess heat in systems
Theoretical attack called Bit2Watt; Mitigations include malicious pattern detection and power buffering systems.

Whenever an AI data center thinks very, very carefully, it can increase its power consumption so much that it causes interruptions and possibly even blackouts and equipment malfunctions. So is it possible for a malicious actor to deliberately trigger this scenario to cause physical harm?

Several researchers from Zhejiang University in Hangzhou, China, wrote a research paper titled “Bit2Watt: A Cyber-Physical Vulnerability Exploiting GPU Workloads in Computing and Power Infrastructures.”

In it, they claim that a malicious cloud tenant is, in theory, capable of launching GPU workloads so intensive that they cause physical damage.

Suggest mitigations

“Our results indicate that GPU loads can reach modulation frequencies greater than 6,000 Hz, compared to only a few hertz observed in conventional household loads such as air conditioners,” the team wrote in their research paper.

“These high-frequency modulations can substantially induce voltage excursions, harmonic distortion, and damping degradation.”

An attacker could use around 1,000 GPUs to attack a one-megawatt local power grid consisting primarily of distributed energy sources (such as solar panels), causing it to lose nearly half of its electrical current and generating about 20% more heat than usual.

“This not only threatens the availability of IT equipment, but also produces a negative damping ratio of -0.27, introducing an unstable mode into the system,” the document adds.

“Once protections are activated and computing loads are removed, cascading failures can be triggered, potentially resulting in blackouts greater than 80 percent on large-scale power systems.”

AI data centers creating huge swings in power consumption are nothing new, and it’s a challenge that some of today’s brightest minds are trying to solve.

Fortunately, the attack is (still) purely theoretical and the researchers published the paper to warn of possible misuse. They also suggested mitigations: Defenders could look for malicious computing patterns, while operators should create energy storage systems for unusual spikes in demand.

Through The Registry