{
  "id": 19321274,
  "title": "Lava Security’s Yakir Kadkoda on the AI Data Center Boom: Growth, Security, and the Risks in Between",
  "url": "https://urgent.news/2026/11/10/lava-securitys-yakir-kadkoda-on-the-ai-data-center-boom-growth",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-10T11:00:45.000Z",
  "source": {
    "name": "HackerNoon",
    "slug": "hackernoon",
    "url": "https://hackernoon.com/lava-securitys-yakir-kadkoda-on-the-ai-data-center-boom-growth-security-and-the-risks-in-between?source=rss"
  },
  "original_language": "en",
  "account": "The rapid growth of artificial intelligence is fueling a surge in demand for computing power, leading to a debate over data center expansion in the United States. While new facilities offer increased capacity, concerns over electricity consumption, water usage, and environmental effects have sparked demands for stricter regulation, with some suggesting a temporary halt to new constructions. However, another critical issue arises: securing the infrastructure that powers AI.\n\nLava Security, a cybersecurity firm specializing in protecting neoclouds and AI data centers, provides visibility across computing, networking, storage, and hardware management systems. This aids operators in identifying vulnerabilities, ensuring firmware integrity, and discovering security gaps between tenants. Yakir Kadkoda, co-founder and CTO of Lava Security, previously worked as Director of Security Research at Aqua Security and currently leads the OWASP Top 10 for Neocloud and AI Data Center Security Risks project. He discusses the ongoing debate about data center growth, vulnerabilities in GPU cloud infrastructure, and the roles of operators, customers, and policymakers.\n\nThe U.S. requires more AI data centers because AI is now considered essential infrastructure, akin to power or transportation networks. Hospitals, banks, universities, and manufacturers all require AI models, and most cannot afford to purchase GPUs in bulk. They must rent them. This need for more data centers reduces prices and wait times for those who aren't large cloud providers. Three examples illustrate this beyond Big Tech: a hospital group running models on patient data needs local or rented GPUs instead of public APIs, a university lab faces delays due to shared clusters, and startups on neoclouds are directly affected by price fluctuations.\n\nIf Congress were to pause new data center construction and expansion, startups, researchers, and smaller companies would bear the brunt initially. Larger tech companies already possess significant computing resources. Restricting new data centers would result in higher prices, reduced availability, and greater concentration of power among the biggest players. In response, AI development might shift to countries still investing in infrastructure.\n\nLava Security analyzed monitoring data from over 12,000 GPUs and discovered that a significant number reported zero utilization between March and May 2026. This finding highlights the inefficient use of existing capacity. While some GPUs may be temporarily unused, awaiting jobs, maintenance, or reservations, a concerning 60% had zero utilization at the time of observation. This points to a larger issue: thousands of expensive AI systems are leaking sensitive operational information to anyone on the internet. Addressing this requires better visibility, security, and utilization of current infrastructure alongside further capacity expansion.\n\nProponents of a data center moratorium highlight the need for responsible infrastructure development, including concerns about electricity, water usage, environmental impact, and security. However, a construction pause is not the solution. Instead of halting data center growth, the focus should be on improving how these facilities are built and operated. Clear requirements, strict enforcement, and accountability measures for operators are necessary. While infrastructure growth must be managed responsibly, stopping construction would only exacerbate the problem by driving prices up, reducing availability, and concentrating power among established companies. Moreover, AI development will continue to move to regions investing in infrastructure, regardless of U.S. actions.\n\nWhen advising a town considering a large data center, Yakir Kadkoda would focus on three key aspects: measurable economic benefits, responsible resource consumption, and operational security. The developer should provide detailed information on electricity and water consumption, who bears the costs of infrastructure upgrades, and the number of permanent jobs created by the project. These commitments should be quantifiable and enforceable. Additionally, a credible security plan is essential, as modern AI data centers involve complex systems with numerous interconnected components, such as servers, management controllers, networking equipment, and industrial control systems. While towns should welcome investment, they must insist on transparency and accountability to ensure that data centers benefit both the local community and the broader economy.",
  "summary": "Lava Security CTO Yakir Kadkoda discusses AI data center expansion, GPU cloud vulnerabilities, tenant isolation, exposed infrastructure and accountability.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}