One of the world's largest companies. Its technologies are used by the Israeli military and police and by US government agencies to surveil immigrant communities and manage prisons.
Microsoft’s Azure cloud platform is the world’s second-largest provider of cloud storage services— after Amazon—controlling a quarter of the global market. It participates in the Joint Warfighting Cloud Capability contract, the main cloud platform for the U.S. military, and before that was the sole provider of cloud services for the U.S. Department of Defense (DOD), through the failed JEDI project. Microsoft has been supporting the U.S. military since at least 1980 and its DOD contracts account for most of its contracts with the U.S. federal government.
Azure has played an instrumental role in facilitating Israel’s 2023-2025 genocidal war on Gaza. It has provided the Israeli military with computing power and artificial intelligence (AI) tools that Israel publicly admitted it needed for the war, as detailed below. This potentially implicates Microsoft in war crimes through the mass killing of tens of thousands of Palestinians, many of them unarmed civilians.
War Crimes against Palestinians
For decades, Microsoft was Israel’s main cloud provider. Most Israeli government and military agencies, including units that handle classified information, reportedly “invested a lot” in developing and creating systems based on Azure.
This changed in 2021, when Microsoft lost its bid for Project Nimbus to Amazon and Google. Under the $1.2 billion Nimbus contract, Israel’s entire public sector is gradually migrating to new cloud infrastructure set up by Amazon and Google inside of Israel. Despite losing this bid, Microsoft’s business with Israel continues to grow.
In the immediate aftermath of October 7, 2023, preparations for the large-scale ground invasion of the Gaza Strip required unprecedented computing power, and the Israeli military’s internal Operational Cloud became overloaded. The military took several steps to address this problem, including using Amazon’s AWS, Google Cloud, and Microsoft Azure, as the commander of the military IT unit MAMRAM later revealed, adding that doing so gave Israel “very significant operational effectiveness” in Gaza.
Israel’s usage of Microsoft’s cloud services and artificial intelligence (AI) has spiked dramatically since Israel began its genocidal attacks on Gaza in October 2023, as reported by the Guardian, DropSite News, and +972 Magazine. The military’s usage of Azure rose by more than 155% between June 2023 and April 2024.
Leaked documents reviewed by the Guardian reveal that the Israeli Ministry of Defense (IMOD), which handles contracting for the Israeli military, “agreed to buy 19,000 hours of engineering support and consultancy services from Microsoft,” between October 2023 and June 2024, for a total $10 million. Some of these are visible on a IMOD procurement database, which discloses unclassified contracts that were awarded without a tender process. It shows that between November 2023 and June 2024, IMOD awarded Microsoft fifteen contracts for Premier Microsoft Consulting Services (MCS), worth a total $1.7 million. The reason for not pursuing a public tender is recorded as Microsoft being the only company that can supply these services.
The military units and systems that have used Microsoft’s services during 2023–2024 include:
- Unit 8200: Israel’s main intelligence gathering agency that specializes in telecommunications surveillance and cyberwarfare.
- Unit 81: a secretive unit that develops spying and surveillance technologies.
- Rolling Stone: the system that monitors Israel’s permit system that restricts and controls the movement of Palestinians in the occupied West Bank and Gaza Strip.
- Matspen (Compass): a technology unit that develops software and applications to support emerging needs of combat troops. During the 2023–2024 war on Gaza, it reportedly “transformed” the military’s drone warfare and developed an app that analyzes combat zones and alerts troops of irregularities or increased activities, as well as a system that manages airstrikes.
- Ofek: the Israeli Air Force’s main software development unit, which develops and manages all of its databases, including the target bank.
- Sapir: a unit that “maintains the ICT infrastructure in the Military Intelligence Directorate.”
- The Military Advocate General’s Corps: the unit tasked with, among other things, "prosecuting Palestinians and lawbreaking soldiers in the occupied territories.”
The Israeli military also makes extensive use of Microsoft Azure’s Artificial Intelligence (AI) capabilities. These include “translation (about half of the average monthly consumption during the first year of the war), OpenAI’s GPT-4 model (about a quarter of the consumption), a speech-to-text conversion tool, and an automatic document analysis tool,” as reported by +972 Magazine. In October 2023, the military’s monthly consumption of Microsoft Azure AI services jumped sevenfold compared to September. By March 2024, it was 64 times higher than before the genocide.
These revelations potentially implicate Microsoft in the mass killing of tens of thousands of Palestinians, many of them unarmed civilians, using multiple AI systems (The Gospel, Lavender, and Where's Daddy?) that the Israeli military developed to generate targets and hit them at unprecedented scale with minimal human intervention.
The Israeli military's use of GPT-4, developed by OpenAI, began in August 2023 and increased 20-fold after October that year. This powerful tool, which can quickly analyze billions of pieces of information, is available through Microsoft Azure's offerings. Microsoft owns 49% of OpenAI, and the two companies collaborate closely.
Notably, OpenAI “quietly deleted language expressly prohibiting the use of its technology for military purposes from its usage policy” in January 2024. And in May 2024, Microsoft started marketing its Azure OpenAI Service as a “powerful tool” that offers a “paradigm shift” for military and intelligence applications. In October 2024, OpenAI announced new guidelines for assessing its “national security partnerships.” However, it is unclear how these policies relate to Israel’s conduct, as OpenAI stated that it has no partnerships with the Israeli military since its GPT-4 service is delivered through Microsoft Azure.
Microsoft employees have raised serious concerns over the company’s involvement in Israel’s apartheid regime and war on Gaza. The company fired two of its employees after they organized a vigil for Palestinians killed in Gaza. Employees have also criticized the company’s policy of matching employee donations to organizations that support the Israeli military and Israel’s illegal settlement enterprise.
Prior to Israel’s 2023–2025 war on Gaza, Microsoft collaborated with the Israeli military in various other ways, including by working with military personnel to develop applications specifically for military operations. For example, according to Who Profits, Microsoft has participated in several “hackathon” events with Israeli military personnel, including a 2021 event at an Israeli military base, where, under the guidance of Microsoft employees, Israeli soldiers developed a weapons calibration app for training combat soldiers. At another such event, Microsoft collaborated with military personnel to develop a “Settlement Defense” app, designed to assist soldiers guarding illegal Israeli settlements.
Microsoft has also provided the Israeli military with technologies used for training. In 2021, for instance, the company provided virtual AI training courses—developed in partnership with the Israeli military—to Israeli soldiers. The Israeli military has also used Microsoft Xbox controllers to train its soldiers on the steering, weapons systems, and “all manner of other operations” of armored fighting vehicles. Additionally, Microsoft Azure software has been used to power Elbit Systems’ OneSim—military simulation software used at Israeli military training sites to mimic real-life “battle scenarios” for tank crews.
In 2002, the Israeli Ministry of Defense signed a $35 million contract with Microsoft for software licensing and cybersecurity products for the Israeli military and other security forces. The purchase, which was described as “strategic,” was the largest software licensing contract in Israel at the time and was mostly funded by U.S. taxpayers through the Foreign Military Sales program. Under the agreement, Microsoft provided the military with “unlimited software products,” computer systems, security tools, and related services.
Israeli Police, Prisons, and Illegal Settlements
The Israel Police use Microsoft Azure cloud computing for all of its databases and systems. In 2020, the Israel Police attested in contracting documents that working with Microsoft “is necessary for the continued function of [its] operational systems, some of which are classified.” Microsoft reportedly claimed that its software “helps Israeli police intelligence officers complete data searches in seconds.”
The Israel Prison Service (IPS), responsible for imprisoning thousands of Palestinians under administrative detention, also contracts with Microsoft for software and related consulting and support services. Between at least 2017 and 2022, Microsoft provided the IPS with software, consulting services, server support, and other computing products/services.
Additionally, Microsoft provides technologies and services to agencies that oversee aspects of Israel’s illegal settlements. The company’s Azure platform powers the Israeli Civil Administration’s “Al Munaseq” app, as reported by Who Profits. this app is used to issue permits to Palestinians in the occupied West Bank and Gaza Strip. Palestinians are required to obtain these permits for work, commerce, family visits, legal appointments, and more. The app is also used to schedule appointments for obtaining biometric ID cards, which Israel requires Palestinians to carry in order to cross Israeli checkpoints.
The company has also provided technologies and services to educational institutions in illegal Israeli settlements. For example, it provides free software and email services to students and staff at Ariel University, located in the illegal Israeli settlement of Ariel in the occupied West Bank. Microsoft has also partnered with the Education Department of Ma’ale Adumim, another illegal Israeli settlement, to provide its software to all students and teachers in the settlement.
Acquiring Israeli High-Tech Companies
Since at least 2014, Microsoft has been acquiring Israeli high-tech startup companies, including Aorato, Adallom, CyberX, and Hexadite, all of which reportedly use technologies originally developed for the Israeli military. Adallom, for example, uses Israeli military–developed “technologies that were used to combat terrorism using machine intelligence and anomaly detection.”
In 2020, Microsoft divested its shares in AnyVision, an Israeli company whose technology powers mass surveillance of the Palestinian civilian population in the occupied West Bank. Microsoft’s venture capital fund M12 made the multimillion-dollar investment in AnyVision a year earlier. The divestment was prompted by a multi-stakeholder campaign demanding that Microsoft live up to its own principles on facial recognition technologies and an official Microsoft investigation into AnyVision’s involvement in mass surveillance of the occupied Palestinian population.
AnyVision’s main surveillance product, Better Tomorrow, uses facial biometrics and artificial intelligence to identify specific people within large crowds. This technology has been integrated during 2018–2019 into Israel’s illegal military checkpoints in the occupied West Bank. AnyVision reportedly has another, more secretive project, using a network of thousands of cameras deployed “deep inside the West Bank” that places the Palestinian civilian population under persistent surveillance. The project reportedly includes vehicle tracking using license plate readers and has led to the arrest of hundreds of Palestinians in 2018 alone.
AnyVision’s technology is used by governments and private actors in at least 44 other countries. Among others, the company sells its technology to state actors in Russia and was expanding its operations in Hong Kong in the summer of 2019, as facial recognition technologies were documented being used to repress protests there. The U.S. Navy also bought “AnyVision equipment” in 2019, and the company’s CEO said that it employs lobbyists in the U.S. Congress to “explain why artificial intelligence is a good thing.”
In 2020, following Microsoft’s divestment, AnyVision split its activities into two companies, a military and a commercial one. SightX, the company developing military applications of AnyVision’s technology, is a joint venture between some of AnyVision’s original shareholders and Israeli state-owned military contractor Rafael. AnyVision’s commercial arm was rebranded as Oosto and in 2025 was acquired by Metropolis, an AI-powered parking platform. Microsoft still offers AnyVision’s facial recognition product on its Azure Marketplace platform.
US Immigrant Surveillance
Microsoft provides the U.S. Department of Homeland Security (DHS) and its agencies Customs and Border Protection (CBP), Immigration and Customs Enforcement (ICE), and U.S. Citizenship and Immigration Services (USCIS) with products and services ranging from basic Microsoft 365 support services to cloud services and data visualization software.
Microsoft Azure Government provides DHS with cloud computing infrastructure. In 2018, ICE started hosting some of its “most sensitive unclassified data” on Azure. This includes “data that supports the core agency functions,” and Microsoft stated that it was “proud to support” ICE’s work. ICE relies on local and state databases such as this one to conduct deportation efforts.
ICE’s Cloud General Support System, which hosts many ICE “mission-oriented applications,” uses both Azure and Amazon Web Services (AWS). ICE stated in 2021 that it requires IT services to “continue to host existing ICE systems that are currently deployed in AWS and Microsoft Azure.” In addition to ICE, in 2022, CBP awarded Microsoft a contract worth a potential $19.1 million for Azure and other support services.
In addition to cloud services, DHS is increasingly relying on Microsoft and other corporations for artificial intelligence (AI) tools used to automate decision-making around the arrest, incarceration, and deportation of immigrants. As part of a $5 million 2024 pilot program involving the use of AI models in investigations, asylum interviews, and other immigration processes, DHS has reportedly been using Microsoft, Amazon, and Google cloud services in conjunction with OpenAi, Antropic, and Meta AI tools. DHS has provided little information about these AI tools, including what data is used and how, what the results are, and how it “identifies or manages errors or conducts oversight,” according to Mijente and Just Futures Law.
ICE’s Repository for Analytics in a Virtualized Environment (RAVEn) platform—which helps ICE “analyze large datasets” to “more easily identify enforcement targets”—uses Microsoft-owned GitHub. GitHub is a code repository and coder collaboration site that, according to a leaked internal email from 2019, ICE has been using Github since 2016. Hosted on AWS, RAVEn draws on biometric data from various sources, including fingerprints and DNA, government data, social media, surveillance photos and videos, GPS, and financial data from private companies.
Microsoft also partners with companies whose technologies are used to power CBP and ICE surveillance systems . For example, Axon’s cloud-based digital evidence platform, Axon Evidence, which is built on Microsoft Azure, is used by CBP’s Incident-Driven Video Recording Systems (IDVRS). Through IDVRS, Border Patrol agents equipped with body-worn cameras record their “encounters with members of the public” in public areas, as well as in or near CBP facilities and ports of entry. These cameras are connected to Axon’s Microsoft-powered evidence platform, which stores recordings for analysis. In 2021, CBP stated that it would begin deploying IDVRS in “a targeted, multi-phased approach to areas of operation where the U.S. Border Patrol lacks adequate fixed camera surveillance technology,” including “immigration checkpoints.”
Microsoft employees have pushed back against the company’s relationships with U.S. immigration authorities. In 2020, in light of the Trump administration’s family separation and other “zero-tolerance” immigration policies, more than 100 Microsoft employees signed an open letter calling on the company to cancel its contracts with ICE and other clients directly enabling ICE. Company CEO Satya Nadella responded to employee concerns by stating that Microsoft was “not working with the US government on any projects related to separating children from their families at the border.”
US Prison Surveillance and Prison Labor
Microsoft has several “offender management” and “offender data visualization” technologies used to surveil people in prison. Microsoft does not generally provide these products directly to prison authorities, but rather through third-party providers, making it difficult to identify the full extent of its involvement in the prison industry. Below are just some known examples.
Microsoft developed its Digital Prison Management Solution based on its policing platform Aware (see below). The system combines Aware’s data from outside the prison (see below) with “corrections operational knowledge” to support prisons “streamline” their operations. It is unclear which, or how many, U.S. prisons and jails use this technology.
In 2009, Microsoft partnered with technology company Tribridge to develop its first product for prisons: Offender 360, a searchable web-based platform used to track, identify, and run mass searches on incarcerated individuals. The product was subsequently adapted into Youth 360, a version for monitoring youth on probation. This version could link to other data systems, such as school and public health systems.
Microsoft has marketed and offered its prison surveillance technologies outside of the U.S. as well. For example, it has previously provided U.K. authorities with Azure-based electronic monitoring tools for “next generation offender tracking.” In 2016, the company published a blog post about the “prison of the future” in the U.K., in which it claimed that prison overcrowding, cost issues, and high recidivism levels could be overcome by the company’s “intelligent technologies” and a “big data” prison management system, and suggested that the increased use of digital technologies within prisons “makes it possible to consider prison as a business.”
In addition to surveillance and prison management systems, Microsoft provides prison authorities, including the U.S. government’s federal prison system, with technology support services, computers, computer software, and other products and services, either directly or through third-party vendors. In 2020, for example, the Department of Justice awarded Microsoft a contract worth a potential $75 million for software and support for various agencies, including the BOP and the U.S. government’s prison labor program, Federal Prison Industries (FPI)/UNICOR. In the 1990s, Microsoft used prison labor to package some of its products through another company; however, in 2018, it committed to prohibiting the use of prison labor.
US Police Surveillance
Microsoft’s police surveillance platform Domain Awareness System (DAS)—also called Microsoft Aware—integrates disparate sources of information to assist police investigations in real time. Initially developed with the New York Police Department in 2012, DAS allows police officers to track and watch people’s movements throughout a city by ingesting data from closed-circuit television (CCTV) cameras and automatic license plate readers (ALPRs), and cross-referencing these data against arrest records, 911 calls, complaints, warrants, and other police sources. DAS also performs video analytics, automatic pattern recognition, and predictive policing. Racial justice and civil liberties organizations have expressed significant concerns that this technology “could amplify the deep-rooted tendency by law enforcement to perceive BIPOC—in particular Black men—as criminal threats.” The technology has also been adopted by police in Brazil and Singapore.
In addition, Microsoft’s HoloLens 2, an augmented reality headset, has been used by police officers to capture evidence at crime scenes and store it for later review. The technology allows officers to map and record entire scenes, including the exact location of individuals and pieces of evidence, and enables officers to later walk through virtual renderings of crime scenes. It is unclear which or how many U.S. police departments have adopted this technology, but it has been used in the Netherlands, U.K., and Australia. It has also been touted for its military applications, including its ability to control unmanned ground vehicles.
Through Azure Government, Microsoft’s Coptivity software, an AI-enabled “conversation mobile app” developed by the San Diego County Sheriff’s Department in 2018, uses “intelligent voice assist” technology to deliver immediate assistance to officers on patrol without a dispatch operator on the other side of the call. The app includes instant access to vehicles’ registration status and drivers’ “criminal and mental health background.”
In 2018, Microsoft boasted that its Azure platform has helped the State of Georgia fight gangs. The company worked with Georgia-based company Formulytics to provide the state with the technical backbone for its Anti-Gang Network. As with DAS, this technology has been used to store and analyze vast amounts of data, helping state law enforcement agencies “create over 25,000 investigative profiles of gang members and identify tens of thousands of connections across the State.”
Following calls from company employees to drop its contracts with U.S. police departments in June 2020, Microsoft stated that it will not sell its facial recognition software to law enforcement agencies until there is a national law, “grounded in human rights,” that governs its use. This announcement followed in the wake of public pressure influencing mass surveillance technology providers Amazon and IBM to issue public statements on police use of their facial recognition services. In 2024, Microsoft reaffirmed this ban, stating that its Azure AI system is prohibited for use “by or for” police departments in the U.S. and that no “law enforcement globally” can use the system on mobile cameras, like body-worn cameras or dashcams, for “real-time facial recognition.”