Today, we are excited to announce that GPT-5.2, Azure OpenAI’s newest frontier reasoning model, is available in Microsoft Azure for U.S. Government Secret and Top Secret cloud environments. This release marks another significant milestone in our mission to bring cutting edge AI capabilities to the nation’s most critical missions, securely, reliably, and at unprecedented speed.

Frontier AI for the Nation’s Most Sensitive Workloads

GPT-5.2 represents the next evolution of frontier level reasoning, planning, and problem solving. It brings improved contextual understanding, stronger multistep reasoning, and better performance on complex analytical tasks, capabilities that support the operational needs of defense and national security customers.

National security customers require solutions that can handle complex tasks with structured outputs and enterprise security and reliability. GPT-5.2 on Microsoft Foundry’s enterprise-grade platform is optimized for these capabilities.

• Multi-Step Logical Chains: Decomposes complex tasks, justifies decisions, and produces explainable plans.
• Context-Aware Planning: Ingests large inputs (project briefs, codebases, meeting notes) for holistic, actionable output.
• Analytics and Decision Support: Useful for wind tunneling scenarios, explaining trade-offs, and producing defensible plans for stakeholders.
• Application Modernization: Make rapid progress in refactoring services, generating tests, and producing migration plans with risk and rollback criteria.
• Data and Pipelines: Audit ETL, recommend monitors/SLAs, and generate validation SQL for data integrity.

By making GPT-5.2 available across classified clouds, mission owners can now apply state-of-the-art generative AI to accelerate analysis, reduce cognitive load, and improve decision support while aligning with U.S. Government security compliance and isolation requirements.

Driven by Customer Need

The arrival of GPT-5.2 in classified clouds is not just a technological achievement, it is a direct response to clear, mission critical demand. Customers across the U.S. Government asked for faster access to advanced reasoning capabilities to support time critical operations.

Their collaboration guided prioritization, influenced an accelerated release timeline, and reaffirmed our commitment to delivering frontier AI where the mission needs it most.

Looking Ahead

With GPT-5.2 now available in our Secret and Top Secret air-gapped clouds, we are continuing to expand the mission-ready AI ecosystem:

• Enhancing support for agentic workflows and orchestration across multiple models
• Improving observability, safety guardrails, and security posture for classified deployments
• Accelerating availability of future frontier models to support the customer mission

We are excited to see how mission teams apply GPT‑5.2 to solve their hardest problems, and we remain committed to enabling secure, rapid access to the most advanced AI capabilities in the world.

Learn more about GPT-5.2 and get started today!

The post Announcing GPT-5.2 Availability in Azure for U.S. Government Secret and Top Secret Clouds appeared first on Azure Government.

As artificial intelligence continues to reshape industries and redefine the boundaries of innovation, Microsoft is proud to announce a leap forward in secure, high-performance computing in our Secret and Top Secret clouds: the integration of NVIDIA H200 Tensor Core GPUs into Azure’s AI infrastructure. This advancement helps unlock unprecedented generative AI capabilities, large language models (LLMs), and high-performance computing (HPC) workloads, while meeting the stringent security and compliance demands of government and regulated industries.

Built for the Future of AI 

The NVIDIA H200 GPU is engineered for today’s generation of AI workloads. With 141 GB of HBM3e memory and 4.8 TB/s of bandwidth, it delivers a 76% increase in memory capacity and a 43% boost in bandwidth over its predecessor, the H100. This enables faster training, larger model support, and more efficient inference, fundamental for applications ranging from real-time language translation to scientific simulations.

The ND H200 v5 series starts with a single VM and eight NVIDIA H200 Tensor Core GPUs, interconnected with 900 GB/s NVLink. ND H200 v5-based deployments can scale up to thousands of GPUs with 3.2Tb/s of interconnect bandwidth per VM. Each GPU within the VM is provided with its own dedicated, topology-agnostic 400 Gb/s NVIDIA Quantum-2 CX7 InfiniBand connection. These connections are automatically configured between VMs occupying the same virtual machine scale set, and support GPUDirect RDMA.

Supported Features:

• Premium Storage: Offers the highest speed Azure Storage can provide to a VM.
• Premium Storage caching: Allows the VM to exceed the underlying disk IOPS limitations through multitier caching technology.
• Generation 2 VMs: Use the new UEFI-based boot architecture rather than the BIOS-based architecture used by Generation 1 VMs, leading to improved boot and installation times.
• Accelerated Networking: Helps enhance networking performance of the VM by enabling Single Root I/O Virtualization (SR-IOV).
• Ephemeral OS Disk: Disks that are created on the local virtual machine storage and not saved to the remote Azure Storage.
• Trusted Launch: Helps protect against advanced and persistent attack techniques through secure boot and other key integrity and trust guarantees.

Real-World Impact 

AI is beginning to deliver tangible benefits across government missions, streamlining operations and enabling new capabilities. Agencies adopting generative AI tools are reporting faster turnaround times for tasks such as grant review and technical content preparation, leading to improved accessibility and efficiency. Document intelligence solutions are helping to reduce manual effort in legal and financial analysis, while predictive analytics and intelligent assistants are enhancing decision-making and transforming how users access and interact with information. Early implementations have demonstrated measurable improvements in workflow speed, data accessibility, and service responsiveness, supporting government teams as they work to better serve citizens.

What’s Next 

Microsoft is actively expanding its GPU capacity across Azure Secret and Top Secret clouds, enabling customers to power their AI workloads. With support for Azure Machine Learning, Microsoft AI Foundry (including powering serverless models like Azure OpenAI and community models from the model catalog), and Azure Kubernetes Service, the ND H200 v5 series integrates seamlessly into existing capabilities.

As we continue to push the boundaries of AI, Microsoft remains committed to delivering secure, scalable, and high-performance infrastructure that empowers every organization to innovate with confidence.

To get started with the newly authorized products now available in Azure for U.S. Government Secret and Top Secret, please contact your Microsoft Account Team. Visit ND-H200-v5 size series – Azure Virtual Machines on Microsoft Learn to discover how to do more for your mission with Microsoft Azure.

The post Unlocking New Possibilities: Microsoft Azure Hyperscale AI Computing with H200 GPUs Accelerates Secure AI Innovation in Azure for U.S. Government Secret and Top Secret appeared first on Azure Government.

NASA’s vast archive of Earth observation data has long been a cornerstone for scientific discovery, environmental monitoring, and global sustainability efforts. Now, Microsoft has made it even easier for researchers, developers, and decision-makers to access and utilize this critical resource. Through a collaboration with NASA, the Harmonized Landsat and Sentinel-2 (HLS) dataset is now available on Microsoft Azure by using Microsoft’s Planetary Computer platform.

Whether you’re a researcher looking to push the boundaries of climate science, an entrepreneur exploring new commercial applications, or a scientist developing cutting-edge tools for agriculture, water resource management, or sustainable land use, NASA’s HLS data in Azure provides a foundation for next-generation Earth observation capabilities.

The island of Uvea (Wallis) of Wallis and Futuna can be seen in this true-color Harmonized Landsat Sentinel-2 reflectance image captured on April 29, 2025, by the Multispectral Imager (MSI) instrument aboard the Sentinel-2B satellite. Image Credit: NASA Worldview

Unlocking the Power of HLS Data in the Cloud

Microsoft Planetary Computer is a platform providing access to a multi-petabyte catalog of global environmental data presented in consistent, analysis-ready formats. Users can access the data through APIs and directly via Azure Storage. This flexible scientific environment allows users to answer questions about the data, and both build applications and use applications on top of the platform.

The HLS dataset is funded by NASA and designed to respond to the needs of the U.S. federal government identified by the Satellite Needs Working Group (SNWG) biennial survey to provide a seamless set of surface reflectance records from the USGS/NASA Operational Land Imager (OLI) aboard the NASA/USGS Landsat-8/9 and Multi-Spectral Instrument (MSI) aboard the European Space Agency’s Sentinel-2A/B remote sensing satellites. By harmonizing these datasets, NASA provides a continuous, high-resolution view of our planet, which allows for better analysis of land use changes, vegetation health, water resources, and more.

With HLS data hosted on Microsoft Azure and available via Microsoft Planetary Computer, users can scale complex environmental computations and apply Microsoft’s advanced AI and cloud-based analytics to drive innovation across industries.

Innovate with AI and Cloud-Based Analytics

For example, developers and researchers can use Azure OpenAI Service and Microsoft Planetary Computer to create intelligent applications that enhance Earth observation analysis. Already, NASA IMPACT and Microsoft have worked together to create a prototype NASA Earth Copilot, an AI-powered assistant that helps scientists and policymakers discover, interpret, and analyze geospatial data using natural language queries.

Earth Copilot enables users to interact with NASA data hosted in Azure repository through plain language queries. Instead, they can simply ask questions such as “What was the impact of Hurricane Ian in Sanibel Island?” or “How did the COVID-19 pandemic affect air quality in the US?” AI will then retrieve relevant datasets, making the process seamless and intuitive.

In short, Azure cloud and AI tools help researchers spend less time on data retrieval and more on analysis and discovery. It also opens the door wider for educators to more easily engage students in Earth Science, and policymakers to more easily make informed decisions on critical issues like climate change, urban development and disaster preparedness.

By combining HLS data with Azure AI, machine learning, and high-performance computing services, users can:

• Automate land cover classification and vegetation monitoring
• Detect changes in urban development and deforestation trends
• Develop AI models to predict environmental patterns
• Create interactive geospatial applications that simplify decision-making

Whether you’re a researcher looking to push the boundaries of climate science, an entrepreneur exploring new commercial applications, or a scientist developing cutting-edge tools for agriculture, water resource management, or sustainable land use, NASA’s HLS data in Azure provides a foundation for next-generation Earth observation capabilities.

Get Started Today

The power of NASA’s Earth observation data is now at your fingertips. To begin leveraging HLS data on the Microsoft Planetary Computer, visit the dataset’s landing page and explore how you can integrate it with Azure’s AI, machine learning, and data analytics tools.

Microsoft and NASA also offer comprehensive training resources to help you make the most of these capabilities. Learn how to build AI-powered applications using Azure OpenAI, process geospatial data efficiently, and deploy scalable solutions in the cloud:

• Build your own Earth Copilot in Azure to navigate Geospatial Data: https://github.com/microsoft/Earth-Copilot/
• Explore HLS Data on Microsoft Planetary Computer Using GitHub Copilot: https://github.com/microsoft/HLS-PlanetaryComputer-Copilot
• Microsoft Planetary Computer Pro (Public Preview): https://aka.ms/planetarycomputerpro
• Learn more about HLS data and its applications in a new NASA Earthdata blog post: https://www.earthdata.nasa.gov/news/blog/harmonized-landsat-sentinel-2-hls-data-now-available-microsofts-planetary-computer

Start exploring NASA’s HLS data today and unlock new possibilities with AI and cloud computing. Visit Microsoft’s Planetary Computer and begin your journey toward the future of Earth science innovation.

The post NASA’s Harmonized Landsat and Sentinel-2 (HLS) Data Now Available on Microsoft’s Planetary Computer appeared first on Azure Government.

In the coming years, artificial intelligence will continue to be foundational to technical innovations for national security missions. Already, U.S. defense and intelligence organizations are using AI to enhance productivity, support decision making, and improve mission outcomes. Microsoft Azure is a backbone for this type of AI-driven work, supporting U.S. government missions with high resiliency, sophisticated capabilities, flexibility, and advanced security, designed to meet the stringent compliance requirements of the nation’s most sensitive data.

Today, we are excited to share that Azure OpenAI Service has been authorized by Defense Information Security Agency for U.S. Department of Defense workloads at Impact Level 6 (IL6). With this announcement, Azure OpenAI Service is now authorized for workloads at all U.S. Government data classification levels.

Azure OpenAI Service is part of Microsoft’s comprehensive AI platform providing access to OpenAI’s industry-leading large language models as well as a host of other AI capabilities, including:  

• Document and text translation  

• Speech-to-text and text-to-speech 

• Text classification 

• Entity extraction 

• Object detection 

• Content safety  

• Search 

The platform also provides a full set of tools for data scientists to build custom AI models or run open-source models from our model catalog with full ML Ops capabilities. All these are available with the scalability, reliability, and enterprise-grade security of Microsoft Azure.  

Azure OpenAI authorized for workloads at all U.S. Government data classification levels  

In early February, the U.S. Department of Defense’s Defense Information Security Agency (DISA) authorized Azure OpenAI Services for IL-6 designated information with Microsoft’s Azure for U.S. Government Secret cloud.  

Earlier this year, Azure OpenAI Service was authorized for use in Microsoft Azure for U.S. Government Top Secret cloud, operating in accordance with Intelligence Community Directive (ICD) 503. In September 2024, Azure OpenAI Service was approved as a service within the FedRAMP High Authorization for Azure for U.S. Government cloud and approved as a service by DISA within the DoD Impact Level 4 (IL4) and Impact Level 5 (IL5) Provisional Authorization for Azure Government. 

With these authorizations, government customers and partners can leverage Azure OpenAI Service for their mission needs, accelerating innovation, improving efficiency, and increasing productivity across all data classification levels. 

A comprehensive portfolio of AI capabilities for your mission needs 

Azure OpenAI Service is one part of a comprehensive AI portfolio available on the Microsoft Azure for U.S. Government clouds.  

• Azure AI Services help developers build ready-to-use AI applications with out-of-the-box and customizable APIs and models. Even without direct AI or data science skills, developers can build applications using pre-built and customizable APIs and models for natural language processing, vision, speech, content safety, and decision-making.  

• Azure AI Search is an information retrieval platform, designed to optimize retrieval-augmented generation (RAG) and modern search. Organizations can store, index, and search their own data, delivering current information to AI models. With Azure AI Search, customers can surface the most relevant information with cutting-edge technology including semantic ranking, vector, and hybrid search.  

• Finally, while Microsoft is committed to providing industry-leading large language models from OpenAI, we also recognize that some mission requirements benefit from other models. Azure Machine Learning is a cloud service for accelerating and managing the machine learning project lifecycle. It supports the training, deployment, and management of machine learning models, including open-source models and open-source foundation models. It also provides tools for MLOps to monitor, retrain, and redeploy models.  

All services are available and authorized across Microsoft Azure’s government and classified clouds: Azure for U.S. Government, Azure for U.S. Government Secret and Azure for U.S. Government Top Secret. 

Expanding access to AI through Microsoft’s partner ecosystem 

Microsoft’s expanding partner ecosystem is leveraging the full capability of Microsoft’s AI-driven cloud platforms.  

• In early March, start-up Scale AI announced that they are partnering with Microsoft to develop and deploy AI-powered solutions for mission partners at U.S. Indo-Pacific Command and European Command, as part of a newly awarded contract by the U.S. Department of Defense’s Defense Innovation Unit. The solutions will enable operational commanders to quickly ingest, process, and summarize large volumes of information to produce draft operations plans, concepts, and operations orders. 

• Microsoft also recently announced AI-focused partnerships with Anduril Industries and Palantir Technologies. These partnerships establish Microsoft Azure as Anduril’s preferred hyperscale cloud for all workloads related to Anduril AI technologies, and makes Palantir one of the early adopters of Azure OpenAI Service within Microsoft Azure for U.S. Government Secret and Microsoft Azure for U.S. Government Top Secret clouds.  

• Additionally, Andworx, a Microsoft Cloud Solutions Partner with a Top Secret Facility Security Clearance, integrates Microsoft AI Services, including Azure OpenAI, into low-code Microsoft Power Platform solutions across the Federal Government and Department of Defense.  

Microsoft’s partner offerings continue to expand and evolve to meet an increasing number of mission requirements.  In addition to the partners mentioned above, there is a growing list of Solution Providers onboarding to our classified environments. You can see a snapshot of current partner offerings below: 

Getting started with Microsoft AI capabilities 

Microsoft is committed to providing U.S. government customers and their partners with access to highly resilient and secure AI capabilities through Azure’s commercial, U.S. government, and classified clouds. These capabilities are critical to enabling government customers and industry partners to transform America’s world-leading AI advancements into next-generation military and national security capabilities. 

To get started with Azure OpenAI Service to serve your mission objectives, contact your Microsoft representative or visit Transformative Cloud Solutions l Microsoft Federal to learn more about AI and other transformative cloud solutions. 

The post Azure OpenAI Service now authorized for all U.S. Government data classification levels appeared first on Azure Government.

The post Embracing Zero Trust: Never Trust, Always Verify appeared first on Azure Government.

Agencies in the U.S. federal government publish an average of 3,700 proposed rules yearly, according to the U.S. Government Accountability Office.  With each proposed rule, agencies generally provide an opportunity for stakeholders and members of the public to submit comments before the rules are finalized.  In some instances, thousands of comments are submitted, with no consistent government-wide process for intaking, analyzing and reporting the findings. A similar dynamic exists in the private sector, where organizations regularly solicit and analyze feedback and comments from customers to improve products and services. 

To harness the power of feedback effectively, Microsoft Federal’s customer success team developed a robust solution for Comment Analytics. The solution identifies, extracts, and analyzes aggregated comments to identify and report key insights. The following blog details the approach the team followed, along with a link to the code, shared in our repository. 

Discover themes, sentiment and suggestions  

The solution we’ve developed provides users with insights by understanding stakeholder perspectives across a series of elements. It can be leveraged for any scenario where you need to extract insights from multiple documents related to a particular topic (e.g. Loan documentation, contract documents, project proposals, and public rulemaking) and do further analytics. Key areas of insight that can be analyzed include: 

• Common themes: Identifying recurring issues that stakeholders frequently mention. Highlighting the most frequently discussed areas across all analyzed documents and comments.  

• Overall sentiment: Gauging the overall tone—positive, neutral, or negative—to assess stakeholder satisfaction. 

• Specific likes and dislikes: Understanding what stakeholders appreciate and what they find frustrating. Pinpointing specific pain points that need to be addressed promptly. 

• Stakeholder Suggestions: Collecting actionable ideas from stakeholders for potential improvements. 

Our solution supports various file types to ensure broad applicability: 

• PDF 

• Text 

• CSV: Each line is treated as a separate comment 

The solution can also be easily extended to support other file types like Word, PowerPoint, and more. 

How it works

1: Extract Insights from Individual Comments 

We start by extracting insights from each individual comment. For CSV files, each line is treated as a separate comment. If a comment is larger than a specified size, we chunk it to manage the data efficiently. This step generates a JSON file with all the extracted insights, including: 

• Summary: A summary of the overall comment. 
• Main Themes: Identification of themes with brief summaries for each. Predefined theme categories can be specified if the focus is on specific themes. 
• Aspect-Based Sentiment: Sentiment score for each theme. 
• Suggestions: Suggestions or remediations mentioned in the feedback. 

This step is critical to the solution, and the foundation for the three steps that follow. Extracting all relevant insights from each comment can be time-consuming if there are many comments. Leveraging the Batch API can expedite this process. Once the individual comments are processed, these insights can be utilized multiple times to generate actionable analytics and meaningful reports. 

2: Merge the Individual Comment’s JSON Files 

Next, we merge the summaries, themes, and suggestions from each comment’s JSON file into three separate files: 

• Merged Summary 
• Merged Themes 
• Merged Suggestions

This allows us to extract insights from each segment separately, such as identifying popular themes or suggestions. 

3: Generate Aggregated Insights from Merged Files 

We then generate final aggregated outputs: 

• Aggregated Summary: A comprehensive summary of all comments. 
• Aggregated Themes: Consolidation of themes to generate the Top 25 most frequently occurring themes. We also explore other options such as categorizing themes for easier consumption and occurrence count. Note that the occurrence count may not be effective if there are too many themes. 
• Aggregated Suggestions: A consolidated list of suggestions for improvement. 

4: Generate Final Consolidated Report (Executive Summary) 

Finally, we combine all aggregated outputs into a Consolidated Report, which includes: 

• Summaries 
• Themes 
• Suggestions 

Upcoming Updates 

We plan to add the following updates to enhance the solution further: 

• Use JSON mode to generate JSON files and structured outputs to create text based on a predefined schema. 
• Utilize the Batch API to generate individual comment JSON files efficiently. 
• Leverage Managed Identity to connect to various Azure Services and use Key Vault to store secrets securely. 
• Incorporate Azure Document Intelligence to extract text and sections from PDF files. Alternatively, we can use PyMuPDF to extract text from PDF files, as it is also adding support for chunking for LLM use cases. 

Get Started  

You can leverage the Comment Analytics solution and extend it to meet your requirements to make informed, data-driven decisions that align with stakeholder expectations and drive operational excellence. 

You can get started with solution today by accessing the code repository on GitHub: AllAboutUnstructuredData/CommentAnalytics/README.md at main · smallangi/AllAboutUnstructuredData · GitHub 

Additional Contributors: Ashish Talati, Chris Kahrs, J Lee, Jay Sen, Narasimhan Kidambi, Pamela Fox and Patrick Davis 

The post Successfully leveraging stakeholder feedback with Comment Analytics appeared first on Azure Government.

Microsoft’s leading AI capabilities are being used across industries to enhance data analysis, bolster security, and support key business priorities. Our U.S. Government customers and authorized partners consistently share that they are ready and eager to apply these same tools to serve their critical missions, and Microsoft is committed to delivering the breadth of our capabilities to them securely, responsibly and compliantly. These capabilities open new possibilities for government agencies to improve public services, streamline internal workflows, and support critical missions like disaster response and public health innovations.  

Today, I am excited to share that 26 additional products in Microsoft’s Azure for U.S. Government Top Secret cloud have been authorized to operate in accordance with Intelligence Community Directive (ICD) 503, including Azure OpenAI Service and Azure Machine Learning. These authorizations are another step forward in our commitment to deliver AI-enabled capabilities to U.S. Government customers. 

Access best-in-class AI capabilities with Azure AI   

With today’s authorizations, the same best-in-class AI capabilities available to commercial customers can be leveraged in accordance with ICD 503 by U.S. Government agencies and authorized partners across the defense and intelligence communities. 

Among the authorized services is Azure OpenAI, which allows agencies and authorized partners operating in Microsoft’s Azure Government Top Secret cloud to benefit from multimodal generative AI models, such as GPT4-o, while meeting the rigorous security and compliance requirements necessary for the nation’s most sensitive data. Authorized users can easily access and integrate Azure OpenAI Service and further ground it on their data for more specialized and accurate intelligence. 

With these authorizations, national security customers can realize the benefits of generative AI, including: 

Enhanced productivity: Reduce time consuming tasks and free up team members to focus on higher value and more satisfying work. 

Augmented cognition: Expedite the task of combing through vast amounts of data to pull insights and interpret recommendations through co-reasoning with AI. 

Accelerated discovery and analysis: Recognize patterns and anomalies in code to detect possible vulnerabilities for analyst review, triage open-source events, and better understand and simulate complex situations. 

Just as customers operating in any Azure environment, customers using Azure OpenAI in the Azure Government Top Secret environment do so with the knowledge that their data remains secure. Proprietary information, from prompts to core data sets, remains within the boundary of Azure Government Top Secret. Prompts (inputs) and completions (outputs), embeddings, and training data are: 

• Not available to other customers 

• Not available to OpenAI nor used to train other models 

• Not used to improve any Microsoft or third-party products or services 

• Not used to fine-tune Azure OpenAI models for your use in your resource unless you chose to fine-tune models with your own training data 

In other words, your data is your data.   

Finally, while Microsoft is committed to providing industry-leading large language models from OpenAI, we also recognize that some mission requirements benefit from other models. Now, Azure Machine Learning is also authorized to operate in accordance with ICD 503, allowing users to train, deploy, and manage open-source models – including open-sourced foundation models. Customers can also use Azure Machine Learning to bring/build their own large language models.    

Bringing Microsoft capabilities together to do more on Azure  

Along with Azure OpenAI and Azure Machine Learning, 24 additional Microsoft products have received authorization to operate in accordance with ICD 503, including: 

• Databases: Azure Database for PostegreSQL and Azure Database Migration Service  

• Identity: Microsoft Entra Domain Services  

• Analytics: Microsoft Graph 

• Insights and Reporting: Power BI 

These authorizations allow customers to access and easily integrate best-in-class data management, security, analytics, and reporting capabilities. Customers can bring these capabilities together uniquely on Microsoft’s Azure Government Top Secret cloud, supporting more informed decision making to deliver mission outcomes.  

To get started with the newly authorized products now available in Azure for U.S. Government Top Secret, please contact your Microsoft Account Team. Visit Azure for U.S. Government to learn how to do more for your mission with Microsoft Azure. 

The post Unlocking new possibilities: Microsoft AI, database, and identity products receive authorization for Top Secret workloads appeared first on Azure Government.

Editor’s note (September 3, 2024): this post has been updated to note that Azure OpenAI has been approved as a service within the DOD IL4 and IL5 Provisional Authorization for Azure Government by Defense Information Systems Agency (DISA). 

As part of our ongoing effort to deliver the latest AI innovations to our government customers, our team is proud to announce several significant updates: 

• Azure OpenAI Service is now approved as a service within the FedRAMP High Authorization for Azure Government. 
• Azure OpenAI has been approved as a service within the DoD Impact Level 4 (IL4) and Impact Level 5 (IL5) Provisional Authorization for Azure Government.
• GPT-4o is now available as part of Azure OpenAI Service for Azure Government and included as part of this latest FedRAMP High and DoD IL4/IL5 Authorization. 

Agencies requiring FedRAMP High Authorization are now able to access these leading AI capabilities within their Azure Government tenant, enabling secure and responsible access to the latest AI technologies while maintaining strict security and compliance requirements. 

GPT-4o, engineered for speed and efficiency, is a ground-breaking, multimodal OpenAI model that integrates text, vision, and audio capabilities to transform how users interact with large language models through more natural and engaging experiences.  

GPT-4o can be used for various AI scenarios, such as natural language understanding, natural language generation, text summarization, text classification, sentiment analysis, question answering, conversational agents, and more. Azure OpenAI Service provides a simple and intuitive interface for customers to interact with GPT-4o, as well as tools and APIs for integrating GPT-4o with their own data and applications.  

Here are some possible use cases for GPT-4o that address needs our customers have identified: 

• Accelerated Discovery: GPT-4o is capable of recognizing patterns and anomalies in code to detect possible vulnerabilities, while also suggesting code modifications that can reduce or mitigate those vulnerabilities. 

• Augmented Cognition: Coupling GPT-4o with large-scale data sets across disparate modalities enables customers to rapidly spot patterns, emerging trends, and reasoned connections to support decision making and actions that improve mission outcomes. 

• Enhanced Productivity: By combining GPT-4o with grounding data from past examples of RFPs and other relevant docs, GPT-4o can create a ready-to-edit first draft RFP to help reduce time spent writing and creating higher quality requirements, while simplifying the process for non-contracting professionals. 

The addition of Azure OpenAI Service to the Azure Government FedRAMP High Authorization is a significant milestone. FedRAMP is a government-wide program that provides a standardized approach to security assessment, authorization, and continuous monitoring for cloud products and services. FedRAMP enables government agencies to adopt cloud services with confidence, knowing that they meet high security standards and comply with federal regulations. FedRAMP authorization is a rigorous and comprehensive process that involves extensive documentation, testing, and auditing by independent third-party assessors. FedRAMP authorization demonstrates Microsoft’s commitment to delivering cloud services that meet the most stringent security and compliance requirements of the US Government. 

Get started with Azure OpenAI Service today in Azure Government 

By making Azure OpenAI Service available in the Azure Government cloud, Microsoft remains committed to enabling government transformation with AI. Along with delivering innovations that help drive missions forward, we make AI easy to procure, easy to access, and easy to implement. Microsoft is committed to delivering more advanced AI capabilities across classification levels in the coming months.

The post Azure OpenAI, including GPT-4o, Approved as a Service within the FedRAMP High Authorization for Azure Government appeared first on Azure Government.

Rising complexity, IP intensity and need for security of physical systems products requires a transformation in how to develop, manufacture and operate these systems. To meet time-to-market demands, software developers and hardware producers alike need optimized cloud + AI solutions that can provide higher levels of collaboration, higher protection and access to scalable computing resources, and the latest AI technologies. Additionally, government and industry require increasing validation of security and access controls to many physical systems that are key to commercial and national security.

In March, Microsoft announced the public preview of Azure Modeling and Simulation Workbench (Azure MSWB), a fully managed engineering environment that enables secure user collaboration while protecting data and IP via multi-layered security and access control solution.

While early use of Azure MSWB by United States Department of Defense, Defense Industrial Base (DIB) and commercial industry has centered around the modernization of secure, collaborative systems design, we envision a future secure, intelligent cloud gateway for the physical systems product lifecycle. A coherent, integrated, secure, and dynamic platform such as this will be essential to enable a seamless DevOps cycle for physical systems products. This is needed for further technology-driven advancements such as generative AI, which has already begun transforming digital software products.

Transforming the DevOps cycle for physical products

We have already seen the complete transformation of how digital software products and services are designed and built. As networking and compute capabilities became more reliable, accessible, and ubiquitous, software developers moved from a linear, siloed process with limited access to run-time data to a continuous DevOps cycle. The modern DevOps loop for software has been critical in better understanding how products are being used, rapidly iterating, and pushing updates to customers in near real-time.

A similar transformation is underway for the design and production of physical products. While nearly every physical product today is already born digitally (i.e., in a CAD system), physical product development still largely follows the linear model – even in instances where physical products have digital technology embedded within.

In design-time, hardware is designed to fulfill requirements. It undergoes simulation and testing before moving into the build and launch. Because of limited feedback in run-time, few companies have visibility into post-sale usage. This impacts a company’s proactive ability to provide any maintenance support or to improve products based on real-world data. Each of these steps uses proprietary tools that often don’t integrate or easily allow data sharing, making it difficult for engineers and product managers to coordinate upstream or downstream.

In the future, we envision that DevOps for physical products will connect design-time and run-time through cloud, edge, and connectivity capabilities across a common compute, data, and collaboration fabric.

This transformation:

1. Transforms design into a faster iterative loop by digitizing more of the process, which enables faster design, simulation, and tapping into cloud-native collaboration tools
2. Creates an operational / run-time loop on the edge by capturing more data with IoT and leveraging it to optimize the operation of products across diverse use cases. These continuous streams of real-time data will allow for the monitoring and adjustment of products’ run-time parameters.
3. Enables integration of the operational / run-time loop into a unified DevOps process by using the edge to calibrate simulations, making them dramatically more accurate – enabling much faster product design and testing.

This future state requires an integrated data, compute, and collaboration platform. Azure MSWB will be essential for facilitating the collaborative framework that supports DevOps for physical systems, while linking engineers and software developers as they work with different engineering apps.

Azure MSWB – design, simulate, collaborate

In development of both software and hardware, security and integrity of IP is critical. Already, Azure MSWB has been used by the US Department of Defense as the foundation for silicon design under the Rapid Assured Microelectronics Program (RAMP).  During phase 1 of the program, Azure MSWB enabled design teams to tapeout (handoff to manufacturing) multiple silicon designs in a matter of weeks.

For example, BAE Systems used Azure MSWB under a stringent Multi-Project Wafer deadline. Azure MSWB enabled physical implementation with commercial and customs EDAs and data management tools, allowing the creation of two chips in ten months – with three times faster runtimes –  that typically would have taken 12 to 18 months.

Raytheon discovered several benefits using Azure MSWB while collaborating across multiple customer and partner teams, including:

• Faster compute runtimes with the ability to scale up quickly as needed with virtually no cap.​
• Enablement of higher-level security classifications due to the secure cloud access.​
• Successful design a state of the art SoC Chiplet, and see a path towards a “one stop shop” storefront model: Computer resources, CAD tools, PDKs, templates, support, using the latest CAD tools.

It was recently announced that Microsoft received a RAMP extension. In this phase 2 extension, the US Government will help grow the RAMP Platform – powered by Azure MSWB – by supporting proofs of concept and design activities for 15 DIB and government organizations, agencies, or laboratories.

As part of the execution of this extension, Microsoft signed and started the onboarding of 15 additional Defense Industrial Base (DIB) companies onto Azure MSWB.

This level of collaboration is an early indicator of how advanced systems development – such as space, air, ground, and other electronic systems – will be transformed through cloud + AI platforms. In the coming years, companies and countries will have the ability to collaborate across companies, countries, and classification domains securely with full visibility into the integrity of their data.

For example, any company designing and building military aircraft will be able to securely coordinate with subcontractors and end customers, while ensuring security and sovereignty of sensitive components, subsystems, and system designs.

This kind of sophisticated, multi-party design is both possible – using Azure MSWB –  and essential for mission customers to achieve their mission objectives.

Creating a secure cloud gateway for physical systems

Together with our partners, a vision for the future is emerging and centers around a secure cloud gateway through which national investments – including end-to-end microelectronics design – can occur.

A secure cloud gateway – like Azure MSWB – is a key platform for the safe and effective delivery of cloud services. This platform enables smooth and secure exchange of data between different computing environments, such as on-premises infrastructure and multiple cloud services. Additionally, as a secure cloud gateway, Azure MSWB uses sophisticated security measures to handle, filter, and monitor the traffic between these environments, ensuring that data transfers are both protected from threats and compliant with policies.

More specifically, Azure MSWB can support the semiconductor R&D and manufacturing sectors, in particular, by facilitating safe and efficient cloud interactions that are vital for the sensitive data and intellectual property involved in this industry. Furthermore, in alignment with the CHIPS Act initiatives, Azure MSWB will enhance national security and economic competitiveness by protecting the technological infrastructure of the US semiconductor and other critical industries.

Azure MSWB also supports the CHIPS for America’s vision by strengthening the US role in technological innovation and ensuring that the R&D ecosystem is resilient and secure against emerging threats. By using cloud-native security models, Azure MSWB reduces risks and improves the performance of the cloud services utilized within the semiconductor sector and beyond, helping to maintain a reliable and secure supply chain.

By expanding beyond design into manufacturing and operations Azure MSWB helps to ensure provenance and efficiency of the physical supply chain. The natural next step in the process will build on lessons learned through the DoD RAMP program to enable scale for a wide variety of physical products – e.g. enable chip manufacturing atop a hyperscale cloud.

Azure MSWB enables both government and industry to collaborate and innovate securely while protecting everyone’s IP. Just as important, Azure MSWB facilitates a more open environment of collaboration by providing the security necessary for new resources – such as a GitHub equivalent for physical design – to be launched and used at scale. This environment, in turn, increases the value that can be derived from applying AI across a large and diverse data set – accelerating insights and informing design, build and improvements.

Azure MSWB also plays a pivotal role in addressing workplace shortages and enhancing educational programs by enabling remote access to crucial resources, centralizing security management, and automating key security processes. Further, Azure MSWB allows organizations to tap into a global talent pool and operate efficiently with fewer on-site personnel, thereby mitigating the impact of localized workforce shortages. It also simplifies IT operations through robust, centralized security management across cloud services and on-premises environments, reducing the need for large, specialized IT teams and automating routine security tasks to compensate for the shortage of cybersecurity professionals.

A platform such as MSWB can also be leveraged for workforce development by providing access to advanced educational tools and fostering collaborative opportunities. By facilitating secure, remote access to a variety of cloud-based resources and industry-standard technologies, these gateways enable educational institutions to extend learning beyond traditional campus boundaries. This not only enriches the learning environment but also prepares students for the technological demands of the modern workforce, effectively bridging the gap between current industry needs and academic offerings. This can be particularly motivating for students in fields like data science, engineering, and digital arts, where large-scale computing resources are a necessity.

Looking ahead to generative AI

For digital software products, generative AI is being adopted for scenarios ranging from code entry – GitHub Copilot, productivity applications – Microsoft 365 Copilot and security – Microsoft Copilot for Security, with new scenarios rapidly emerging. As physical systems design evolves towards a software-like DevOps paradigm, similar opportunities abound for using generative AI to accelerate, reduce cost of, and secure physical systems lifecycles. Last November, our industry partner Synopsys announced Copilot for EDA. Transformative application of generative AI for physical systems will require a secure, collaborative, scalable data and compute platform, such as the one Azure MSWB alongside Azure and Azure OpenAI provide.

Learn more

Azure MSWB is now in public preview (request here) and we are bringing the service to more regions. We’ll have global availability by the end of the summer 2024 with Azure MSWB being deployed in the Sweden region in May ’24 and deployed in the APAC region later this summer.

If you’d like to stay up to date on news and announcements about our secure collaboration cloud solutions for engineering systems and modeling, we invite you to sign up here for the latest, or you can speak to our technical specialists by requesting an outreach here.

The post Anticipating the future of physical systems design appeared first on Azure Government.