Israel’s elite military intelligence unit has been using Microsoft Azure to build and operate a colossal surveillance system that hoovers up Palestinian phone calls on an industrial scale, a joint investigation by The Guardian, +972 Magazine, and Local Call has revealed. Since 2022, Unit 8200 has fed a continuous stream of intercepted telecommunications into Azure’s cloud infrastructure, amassing an archive of roughly 11,500 terabytes—equivalent to some 200 million hours of conversations—and using AI-driven tools to turn that data into intelligence that shapes military operations in Gaza and the occupied West Bank.
The invisible backbone of this operation is a direct partnership between Unit 8200’s commander, Yossi Sariel, and Microsoft CEO Satya Nadella. Their face-to-face meeting in 2021, until now undisclosed, set in motion a collaboration that would grant the Israeli signals intelligence agency access to Azure’s virtually unlimited storage and compute power. With that capacity, Unit 8200 engineers built a centralized repository capable of ingesting up to a million calls per hour, indexing and analyzing them in near real time. What emerges is a system that does not merely archive; it actively fuels lethal targeting decisions, blurring the boundary between bureaucratic surveillance and kinetic warfare.
The scale is difficult to overstate. Eleven thousand five hundred terabytes of voice data—everyday chatter, intimate conversations, business dealings—are stored inside Microsoft data centers, accessible to analysts at the click of a button. The system’s ingestion rate of one million calls each hour means that in less than nine days it can process a volume of calls equal to the entire population of Gaza having a brief conversation. This is not targeted wiretapping; it is bulk, indiscriminate collection that sweeps up combatants and civilians alike, men, women, and children, without distinction.
Unit 8200 did not stop at storage. Azure became the substrate for a suite of AI-powered analytical tools that automatically parse call metadata and content for patterns, keywords, and relational links. Language models trained on Arabic dialects scan transcripts for terms that might indicate militant activity, while graph algorithms map social networks, constructing organizational charts of persons of interest. These outputs feed directly into recommendation engines that prioritize targets for drone strikes or other kinetic actions. Three Unit 8200 sources told investigators that the cloud infrastructure “facilitated the preparation of deadly airstrikes” and “shaped military operations in Gaza and the West Bank.” The system, one source said, acted as a “force multiplier,” converting raw audio into actionable combat intelligence with unprecedented speed and scope.
The humanitarian ramifications are catastrophic. Since the Hamas-led attack on October 7, 2023, Israel’s military campaign in Gaza has killed more than 61,100 people according to local health authorities, reduced entire neighborhoods to rubble, and pushed the enclave to the verge of famine. The International Criminal Court issued arrest warrants for Prime Minister Benjamin Netanyahu and former Defense Minister Yoav Gallant, citing alleged war crimes and crimes against humanity. In this context, the role of a cloud platform that enables mass surveillance and accelerates targeting decisions raises urgent questions about corporate complicity in possible violations of international law.
Microsoft’s involvement is not passive. The company’s senior leadership, including Nadella, engaged directly with the intelligence unit’s commander to discuss the adoption of Azure. While Microsoft has long sold cloud services to governments, the extent of the collaboration described by sources—a tailored, intelligence-grade surveillance infrastructure built on Azure—implies a deeper entanglement than standard commercial licensing. The company’s public statements emphasize compliance with local laws and commitment to ethical principles, but they do not address the documented use of its technology for bulk surveillance of an entire population under military occupation. No public audit or transparency report accounts for Azure’s use by Israeli intelligence, and Microsoft has not disclosed whether it conducted a human rights impact assessment before or after the partnership began.
This opacity runs counter to a growing international consensus that technology companies bear responsibility for how their products are used, particularly when those uses may contribute to human rights abuses. The United Nations Guiding Principles on Business and Human Rights call on firms to carry out due diligence to identify, prevent, and mitigate adverse human rights impacts linked to their operations. The EU’s proposed AI Act and the Biden administration’s executive order on AI both signal tightening expectations for accountability. Yet the Azure-Gaza case shows a wide gap between principle and practice.
Legal scholars point out that facilitating the collection and analysis of personal data at this scale, without any form of judicial oversight or consent from the monitored population, may violate the right to privacy under international law. When that data is used to select targets for airstrikes in a densely populated urban area, the chain of consequences leads directly to civilian death and displacement. If International Criminal Court investigators were to trace targeting decisions back to the Azure-based system, Microsoft could find itself entangled in war crimes inquiries—a scenario that would carry profound reputational and financial consequences.
The Guardian’s investigation has landed like a thunderclap among human rights organizations and tech accountability advocates. Access Now, Privacy International, and Amnesty International have all condemned the use of cloud infrastructure for mass surveillance in conflict zones. They demand immediate transparency: Microsoft must disclose the full scope of its contracts with military and intelligence agencies, implement binding policies that forbid uses facilitating human rights abuses, and subject sensitive deployments to independent oversight. Some activists have called on the company’s institutional investors—who collectively manage trillions of dollars—to press for answers, arguing that the risk to long-term shareholder value from complicity in war crimes is too great to ignore.
For Microsoft’s enterprise customers, the revelations inject new urgency into debates about cloud sovereignty and ethical procurement. Governments, universities, and corporations that entrust their data to Azure may now question whether their information sits alongside a massive trove of intercepted calls, and whether Microsoft’s commercial interests could override their own privacy expectations. The case strengthens arguments for data localization laws and for rigorous, enforceable contractual clauses that restrict secondary use of cloud infrastructure.
The Israel–Azure episode is not an isolated anomaly. It is the most dramatic public example to date of a broader fusion between commercial cloud computing and state surveillance power. Intelligence agencies around the world are migrating from on-premises data centers to public clouds because of the cost, scalability, and advanced analytics they offer. The same tools that let a retailer personalize shopping recommendations can be flipped to identify adversaries, track dissent, and optimize kill chains. Without binding international rules, the line between legitimate national security and oppressive surveillance blurs, and the next mass atrocity could be aided by any of the major providers.
What happens next will test the tech industry’s capacity for self-regulation and the international community’s willingness to enforce human rights standards in the digital domain. Microsoft has the opportunity to lead by example—by coming clean about past uses, halting any ongoing assistance to mass surveillance in conflict zones, and championing a new framework for cloud governance that puts fundamental rights ahead of contract value. Failing that, the Azure revelations may become a benchmark for corporate culpability in the age of algorithmic warfare. For the millions of Palestinians whose voices have been archived without consent and used to target their loved ones, the question is not abstract. It is a matter of survival.