Microsoft CEO Satya Nadella personally met with the head of Israel’s elite cyber-intelligence unit in 2021 to greenlight a cloud-powered mass surveillance system that now captures and stores virtually every mobile phone call made by Palestinians in Gaza and the West Bank, a joint investigation by The Guardian, +972 Magazine, and Local Call has revealed.

The partnership between Microsoft Azure and Unit 8200—Israel’s equivalent of the NSA—enabled an unprecedented dragnet of civilian communications, with the harvested data stored on Azure servers located in Ireland and the Netherlands. The system became operational in 2022 and grants Israeli intelligence the ability to play back and analyze millions of conversations at will.

This exposé, published on August 7, 2025, draws from a trove of leaked Microsoft documents and interviews with 11 sources inside the tech giant and Israeli military intelligence. It lays bare a controversial alliance that deepens questions about Big Tech’s role in state surveillance and the erosion of civilian privacy in conflict zones.

How Azure Became the Backbone of Unit 8200’s Surveillance

Unit 8200, long celebrated as a hothouse of cybersecurity talent, ran up against hard physical limits when it sought to expand signal intelligence operations in the occupied Palestinian territories. In-house servers could not handle the volume of voice data the agency wanted to vacuum up. By 2021, its commander, Yossi Sariel, turned to Microsoft for a solution.

Azure offered exactly what the military needed: elastic storage, high-performance computing for real-time interception and analysis, and the ability to carve out a physically and logically isolated cloud environment. According to the investigation, Microsoft agreed to set up a dedicated, air-gapped section of Azure exclusively for Unit 8200. This bespoke infrastructure allowed Israeli intelligence to bypass domestic hardware constraints and rapidly scale a system that now sweeps in virtually every mobile call made by Palestinians under occupation.

The surveillance is indiscriminate. Rather than targeting specific persons of interest, the system hoovers up the communications of an entire population—children, medical workers, journalists, and aid workers included. Once captured, calls are retained indefinitely, creating a searchable voice archive that can be replayed for intelligence leads, profiling, or evidence gathering.

The Technology Underpinning the Operation

Microsoft Azure is one of the world’s leading public clouds, trusted by governments and enterprises for its security and scalability. For intelligence agencies, it delivers:

  • Elastic storage that grows on demand, eliminating the need for expensive data center expansions.
  • High-performance compute to run complex voice recognition, transcription, and sentiment analysis at scale.
  • Air-gapped regions that keep sensitive workloads physically disconnected from the public internet.
  • Redundancy and resilience to ensure continuity during cyberattacks or kinetic conflict.

Unit 8200’s deployment exploited these features to build a massive eavesdropping machine. The system reportedly uses automated speech-to-text and keyword spotting to sift through the daily deluge of calls, flagging individuals based on unknown criteria. With rapid advances in AI, the stored data could later be mined for deeper behavioral profiling—a chilling prospect for a population already living under severe restrictions.

The Nadella-Sariel Meeting: A Pivotal Encounter

Central to the scandal is a high-level meeting at Microsoft’s Redmond headquarters in 2021. Satya Nadella, Microsoft’s CEO, sat down with Yossi Sariel to discuss the project. Sources say Sariel explicitly sought Nadella’s backing to leverage Azure’s storage and processing might for the intelligence system.

Microsoft has responded by claiming Nadella was never briefed on the specific type of data Unit 8200 planned to store. “At no time during this engagement has Microsoft been aware of the surveillance of civilians or collection of their cellphone conversations using Microsoft’s services,” a spokesperson stated, adding that an external review commissioned by the company found no evidence to the contrary.

Yet the leaked documents and insider accounts contradict that defense. Former employees and intelligence sources allege that the civilian surveillance context was widely understood internally, and that Microsoft carved out a special Azure enclave with full knowledge of its intended use. The company’s own staff reportedly raised red flags about the ethical and legal implications, but leadership pushed ahead.

The meeting’s significance goes beyond technology procurement. It symbolizes the direct involvement of a global CEO in enabling a military operation that independent United Nations experts and rights groups like Amnesty International have described as possibly violating international humanitarian law. Nadella’s swift public condemnation of the October 7 Hamas attacks—but silence on Israeli actions in Gaza—has only intensified criticism.

The Azure-Unit 8200 partnership sits at a volatile intersection of law, ethics, and corporate policy.

  • Jurisdictional chaos: Data from Gaza and the West Bank flows to Azure servers in Ireland and the Netherlands, raising questions about which country’s privacy and surveillance laws apply. European regulators have so far taken no public action.
  • Wilful blindness: Microsoft claims it did not know its technology would be used for civilian surveillance. But critics argue that a company with Microsoft’s resources and access to Israeli military briefings should have foreseen—and actively interrogated—the end use. The “don’t ask, don’t tell” posture, they say, amounts to wilful blindness.
  • International humanitarian law: Mass, indiscriminate surveillance of a protected civilian population may violate the principles of distinction and proportionality under the Geneva Conventions. If Azure-stored intelligence directly supported airstrikes or military operations, as some sources allege, the legal liability could extend to Microsoft.

Human rights groups warn that the system creates a permanent chilling effect on speech. Palestinians who know their private calls can be stored and replayed by Israeli intelligence are less likely to express dissent, seek medical aid, or communicate openly with journalists. The indiscriminate sweep also endangers humanitarian workers and medical teams, whose communications can be intercepted and potentially used to target relief operations.

Microsoft’s Denials and Internal Revolt

Microsoft has issued forceful denials. A company statement says: “Based on our review, we have found no evidence that Microsoft’s Azure and AI technologies, or any of our other software, have been used to harm people.” It stresses that external audits back its position.

Inside the company, those assurances ring hollow for many. Joe Lopez, a firmware engineer working on Azure, disrupted Nadella’s Build developer conference keynote in May 2025 with pro-Palestinian chants. He later explained his protest in a Medium post: “As one of the largest companies in the world, Microsoft has immeasurable power to do the right thing: demand an end to this senseless tragedy, or we will cease our technological support for Israel.”

Lopez’s action is the most visible sign of bubbling internal dissent. Multiple Microsoft employees have reportedly objected in internal channels, only to see the company reportedly ban the word “Palestine” from company-wide emails and chats—a move that fueled perceptions of institutional bias. The tension mirrors earlier tech industry fights over contracts with U.S. immigration authorities and China’s surveillance apparatus.

The Bigger Picture: Cloud Militarization and Corporate Complicity

Microsoft is not alone in courting defense and intelligence contracts. Amazon Web Services, Google Cloud, and Oracle all host sensitive military workloads. The trend of cloud-driven warfare and surveillance is accelerating, often without meaningful public oversight.

Key risks extend beyond this single case:

  • Normalization of mass surveillance: Once a major cloud provider enables a population-wide dragnet, other states will seek similar capabilities, lowering the barrier for authoritarian regimes.
  • Vendor lock-in: Intelligence agencies that build their operations on a proprietary cloud are beholden to the provider’s pricing, policies, and political whims.
  • Escalation of data-driven conflict: AI-powered analysis of voice, video, and location data can accelerate targeting cycles, making wars deadlier and faster.

Calls for reform are growing. Rights groups and tech ethicists demand pre-deal human rights impact assessments, far greater transparency about isolated government cloud regions, and robust whistleblower protections for employees who raise alarms. Without such guardrails, the Azure-Unit 8200 blueprint may become a template for future militarized cloud deployments worldwide.

The Reckoning Ahead

The allegations about Microsoft Azure’s role in the mass surveillance of Palestinians are a stark warning. They show how rapidly commercial technology can be weaponized against civilian populations when corporate governance fails to keep pace with geopolitical realities.

For Microsoft, the immediate choices are stark. The company can continue to deny, deflect, and stonewall—a path likely to deepen employee unrest, consumer boycotts, and regulatory scrutiny. Or it can launch a genuinely independent investigation, sever ties with military units using its infrastructure for illegal surveillance, and champion binding industry-wide standards.

The stakes go beyond one company or one conflict. As cloud and AI become ever more embedded in the machinery of state power, the world must decide whether the same tools that connect and empower can be allowed to oppress and harm without accountability. The Azure-Unit 8200 revelation may prove to be the tipping point that forces that decision.