Israel’s elite military intelligence unit has been running a colossal surveillance operation on Microsoft’s Azure cloud platform, stockpiling roughly 200 million hours of intercepted Palestinian phone calls and analyzing up to a million conversations every hour. The partnership, which began after a high-level meeting in 2021 between Unit 8200 commander Yossi Sariel and Microsoft CEO Satya Nadella, has allowed the agency to build a 11,500-terabyte data trove that directly feeds military operations in Gaza and the occupied West Bank.

Leaked documents and interviews with sources inside both Microsoft and Israeli military intelligence reveal that Azure has been the technological backbone of this program since it became fully operational in 2022. The system captures and stores recordings of millions of mobile phone calls made daily by civilians in the Palestinian territories, then applies advanced analytics to extract intelligence for targeting and tactical decisions.

A meeting that changed the surveillance landscape

The collaboration between Unit 8200 and Microsoft did not happen by chance. In late 2021, Yossi Sariel—the commander of Unit 8200—met with Satya Nadella to discuss how Microsoft’s cloud infrastructure could support Israel’s intelligence needs. The meeting focused on cybersecurity and defense against nation-state attacks, but it quickly expanded into a far broader arrangement.

According to the investigation published by Anadolu Ajansı, Microsoft engineers worked closely with Unit 8200 to harden Azure’s security posture to meet the military’s exacting standards. They implemented advanced encryption, access controls, and monitoring capabilities. While Microsoft insists the engagement was solely about protecting Israel from cyber threats, internal records paint a more complicated picture: the company was aware that the data flowing into Azure included raw intelligence intercepted from civilian communications.

A digital ocean of intercepted calls

The scale of the surveillance operation is staggering even by the standards of modern signal intelligence. Unit 8200’s Azure-based platform ingests and processes up to one million phone calls per hour. Over time, it has amassed an archive of approximately 11,500 terabytes of data—equivalent to roughly 200 million hours of conversations. To put that in perspective, if you were to listen nonstop to every recorded call, it would take more than 22,800 years.

This data lake is not a passive repository. Unit 8200 uses Azure’s compute power—likely leveraging AI and speech recognition services—to transcribe, translate, and flag conversations of interest. The intelligence derived from this analysis has been instrumental in preparing airstrikes, identifying targets, and shaping broader military operations in Gaza since the October 2023 escalation, as well as in ongoing operations in the West Bank.

Microsoft’s tightrope walk

When contacted for comment, Microsoft said it was “not aware” that its services were being used for surveillance of civilians or the collection of their cellphone conversations. The company reiterated that its work with Unit 8200 was limited to cybersecurity and protecting Israel from attacks. However, the Anadolu investigation cites internal Microsoft records that suggest company employees knew the data included live, raw intelligence feeds. This discrepancy has ignited a fierce debate about the responsibilities of cloud providers when their infrastructure is used for mass surveillance by state clients.

Microsoft’s public position aligns with its stated principles regarding human rights and responsible technology. The company has previously touted its commitment to ethical AI and has cut off access to facial recognition tools for police departments until regulations are in place. Yet the Azure deal with Unit 8200 appears to have moved forward with few public guardrails, raising questions about the consistency of Microsoft’s ethical policies when large government contracts are on the line.

The mass interception and storage of private communications without consent contravenes widely accepted privacy norms and likely international humanitarian law. Human rights organizations have long criticized Israel’s surveillance practices in the occupied territories as a form of collective punishment and an infringement on the right to privacy. Using cloud-stored call data to facilitate military strikes—some of which have resulted in significant civilian casualties—deepens those concerns.

Legal experts point to the European Union’s General Data Protection Regulation (GDPR) and the emerging body of case law around the extraterritorial application of privacy rights. While Microsoft Azure’s servers used in this operation may be located in Israel or the United States, the data belongs to Palestinian individuals who have not consented to its collection. This creates a grey zone that neither Microsoft nor Western governments appear eager to clarify.

The cloud’s military dilemma

The Unit 8200–Azure case is not isolated. It sits at the intersection of two larger trends: the rapid militarization of cloud computing and the growing dependence of intelligence agencies on commercial AI and storage platforms. Amazon Web Services, Google Cloud, and Microsoft Azure all have dedicated divisions for government and defense customers. While such contracts are legal and often lucrative, they increasingly blur the line between civilian technology and warfare.

For Microsoft, the stakes are particularly high. Azure is the company’s fastest-growing revenue engine, and government contracts are a key pillar of that growth. Yet revelations like these could alienate civil-liberties-minded enterprise customers, developers, and consumers. Windows users and IT professionals who rely on Microsoft’s ecosystem may feel uneasy knowing the same cloud infrastructure that powers Teams and Outlook is also being used to process intercepted civilian phone calls.

What insiders and watchdogs are saying

Within Microsoft, the partnership with Israel’s military has not gone unnoticed. According to Anadolu’s sources, some employees expressed unease when they realized the full scope of Unit 8200’s usage. One engineer reportedly raised concerns that the “cybersecurity” narrative did not match the petabytes of audio data flowing into the tenant. These internal tensions reflect a broader tech-industry reckoning over the weaponization of commercial software.

Civil society groups have called for greater transparency from cloud providers. “When a corporation sells cloud services that are then used to facilitate mass surveillance in an occupied territory, it cannot hide behind a curtain of ignorance,” said a spokesperson for the Electronic Frontier Foundation, though the organization was not directly involved in this investigation. The pressure is mounting for Microsoft to conduct a thorough review of the partnership and to publicly commit to similar human-rights due diligence for all defense-related contracts.

The road ahead

The Unit 8200–Azure story underscores a fundamental challenge of the digital age: the same tools that enable global collaboration, business continuity, and personal convenience can be repurposed for mass surveillance and targeted violence. As cloud computing becomes more deeply embedded in every industry, the need for binding international norms around state use of commercial platforms grows more urgent.

Microsoft could take several steps to address the fallout. It could commission an independent audit of the Unit 8200 engagement, publish clearer policies on military use of Azure, and establish a human-rights oversight board with the power to veto contracts that present a high risk of abuse. Such moves would not only restore trust but also set a precedent for the industry.

For now, the archive of 200 million hours of Palestinian calls remains on Microsoft’s servers, and Unit 8200 continues to mine it for intelligence. The ethical and legal questions that swirl around this partnership will not fade until technology companies and governments alike are held to account.