Microsoft announced three new large language models are now available on the company’s AI platforms. Starting today on both Microsoft Foundry and MAI Playground, developers have access to MAI-Transcribe-1, a transcription model across 25 languages; MAI-Voice-1, which handles natural, expressive speech generation; and MAI-Image-2, the company’s most capable image model yet. Read more about how they work.

Microsoft is on track to spend $5.5 billion to power Singapore’s AI future before the end of 2029, the company announced Wednesday. The funding will largely go to cloud and AI infrastructure, as well as ongoing operations in the country. Microsoft Vice Chair and President Brad Smith also announced the expansion of Microsoft Elevate programs in Singapore to provide AI tools and skills to tertiary students, teachers and nonprofits with responsible AI to uplift all communities in the AI era.

The ever-increasing capabilities of AI have everyone wondering what it means for their career. LinkedIn’s Chief Economic Opportunity Officer Aneesh Raman joins the Worklab podcast to explore why curiosity, adaptability and action matter now more than ever. Drawing from his new book, “Open to Work: How to Get Ahead in the Age of AI” — cowritten with LinkedIn CEO Ryan Roslansky — Raman makes the case that careers aren’t being automated away. They’re being reshaped for the AI era, by the people willing to move first.

Imagine a secure modern AI system as a very new, very junior employee. It’s smart and eager to help but also can easily stumble into trouble. Like a junior person, AI works best when given clear, specific goals. The vaguer the directions, the more likely it is to misinterpret them. But how does that play out at a system level? Read on for some practical advice about security fundamentals in the age of AI.

Work is still evolving in the AI era, but it’s reached a turning point. In the new book released today, “Open to Work: How to Get Ahead in the Age of AI,” LinkedIn’s Ryan Roslansky and Aneesh Raman explore how people and organizations can move forward thoughtfully, keeping human ambition at the center.

Microsoft announced that Copilot Cowork is available starting today via the Frontier program. Workers using Microsoft 365 can describe the outcome they want and Copilot Cowork creates a plan, reasons across their tools and files and carries the work forward. All the while, the human workers will be able to see progress and have opportunities to steer the program’s next moves. With skills from Claude and Microsoft built in, such as calendar management and daily briefing, Copilot Cowork can handle everything from one-off tasks to repeatable workflows like a monthly budget review.

What will humans do in the AI era? In an upcoming book, Ryan Roslansky, the CEO of LinkedIn and executive vice president of Microsoft 365 and Copilot, explores how work is changing, what tools are accelerating that change and how human workers can find their footing in a rapidly changing landscape. Roslansky makes the case that as AI absorbs more of the drudgery of work, people will have the space to do the human work they’ve always been best at but rarely had time for.

Nonprofits tackle some of the world’s biggest challenges, from creating more affordable housing to alleviating poverty, often with limited resources. A new initiative, Microsoft Elevate for Changemakers, will help nonprofits do even more with AI. It will give them access to practical AI training, including guidance on responsible adoption, ways AI can streamline everyday work and how it can support fundraising, so they can drive even more change around the world.

In the traditional model of cyber security – built on siloed directories and disconnected access policies – cyberattacks don’t have to totally break digital defenses. They just move between them, in the gaps between different apps, resources and environments. As organizations manage growing numbers of human and AI agentic identities, the way they manage digital security has to evolve as well. Read more about the three critical layers that a modern identity security solution must unify.

Copilot doesn’t shorten your workday, KPMG Canada’s Christine Andrew says. It changes what you do with it. The AI executive says that previously, the mechanics of work took up most of her time. Now, with help from AI, she’s been able to strengthen her focus on “very high quality” execution and reallocate her attention toward deeper reflection on how the business is going.

The world is racing to meet a historic surge in power demand with an infrastructure pipeline built for the analog age. Driven by the exponential expansion of digital technologies and the reindustrialization of supply chains, the mandate for always-on, carbon-free power is urgent.

In a new blog post, Darryl Willis, Microsoft’s corporate vice president of Worldwide Energy and Resources Industry, writes that nuclear energy is the essential key to powering the future, but the industry remains trapped in a delivery bottleneck. Willis announces Microsoft is working with Nvidia on an AI-for-nuclear collaboration that will break the infrastructure bottleneck and provide end-to-end tools that will optimize build-out and operations across the industry.

AI infrastructure is still in its chaotic phase. Early on, teams made their own choices, using different tools, different reasoning and so on to build out. Initially, it felt flexible…but at scale, it was revealed to be fragmented. When companies switched to Kubernetes — an open-source platform designed to automate deploying, scaling and managing containerized applications — it was an answer to how to run AI systems safely.

Now, as the industry looks to the future, Microsoft’s teams have continued investing across open-source AI infrastructure. Read more about what was announced today at KubeCon + CloudNativeCon Europe 2026 in Amsterdam.

Takayuki Hirayama founded ARUM Inc. 20 years ago from his kitchen table in Kanazawa, Japan. In that time, the company’s morphed from making precision parts to designing AI-driven manufacturing systems using Microsoft Azure. Now he wants to take this automation technology global.

ARUM Inc is turning decades of Japanese machining expertise into software, and fully automated machining centers. Precision parts can now be crafted by junior workers, thanks to AI. And more guidance will come from KAYA, a prototype conversational AI built with Azure AI Speech and Azure OpenAI in Microsoft Foundry.

Microsoft is adding a new pillar to its Zero Trust security approach by extending the model to AI. Zero Trust, which means building security systems that assume security breaches and check every access point, is now being applied across the entire AI lifecycle, from data ingestion and model training to deployment and agent behavior. Read more about the company’s answer to the question, “We’re adopting AI fast, how do we make sure our security keeps pace?”

During tax season, it’s not that unusual to get time-sensitive emails like refund notices, payroll forms and filing reminders. That makes this time of year peak cyberattack season.

Every year, there is an uptick in tax-themed phishing campaigns as Tax Day (April 15) approaches in the United States. Read more about the most common scam email campaigns going around this year.

Every time technology takes a leap forward, workers wonder what they will do next. Steam power, electricity and then computing each reshuffled how people made a living, producing genuine anxiety about what would be left on the other side.

“And each time, the answer turned out to be: more than before, but different than expected,” Jared Spataro, Microsoft’s CMO of AI at Work, writes in a new post. He says the future of the workplace will be defined by humans who can identify “what’s worth doing and how the work should run.”

Today, Microsoft announced a new, in-house text-to-image AI model: MAI-Image-2. The model is designed primarily for high-quality creative work, like photorealistic images and infographics. Read more about how it works.

As cyberattacks on water and wastewater utilities become more frequent and dangerous, the U.S. Office of the Director of National Intelligence is warning that the critical infrastructure is an easy target for foreign adversaries because far too many have outdated software and poor password practices. Today, Microsoft is releasing a report that examines both the urgency of this challenge and what it will take to close the cyber readiness gap in the water sector. Read more for the three clear findings about what helps and what limits cybersecurity in the water sector.

Behind every computer or smartphone screen, there is a real network of cables flowing with data. Microsoft researchers have reimagined those connections, using energy‑saving MicroLEDs to move data between servers in a more efficient and reliable way. Another innovation called Hollow Core Fiber is complementary, passing data through air instead of fiber, increasing speed and cutting latency over longer distances. These innovations are a quiet revolution in the way our “digital plumbing” works.

Microsoft said it is expanding its partnership with Nvidia to help companies move artificial intelligence from experiments into real‑world use. At Nvidia’s GTC conference, the company announced new Microsoft Foundry tools for building and running production‑ready AI agents, alongside Azure infrastructure optimized for inference‑heavy, reasoning‑based workloads and deeper support for physical AI systems that connect software with real‑world operations.

When AI agents mess up, it can be hard to figure out where they first went wrong. AgentRx is a new open‑source tool from Microsoft Research that helps quickly pinpoint the exact step where an AI agent’s mistake became unavoidable — and explains why. Read more about how it works.

At some point, many of us have stared at a medical test result we didn’t understand or worn a device that tracked everything but revealed little. Sat down at a consultation and immediately forgot all of our questions for the doctor. Many people don’t really need more health information – they need help to make sense of what they already have. On Thursday, Microsoft launched Copilot Health, a separate, secure space within Copilot where medical intelligence makes sense of your information and delivers personalized health insights that you can act on. Read more about how it works and sign up for the waitlist.

Microsoft released new data showing how well its email defenses work after malicious messages slip through initial filters. In a new post, Jeff Pinkston, the vice president and general manager for Microsoft Defender for Office 365, writes that Microsoft Defender removes nearly 71% of harmful emails and that layering additional cloud email security tools can cut even more spam and promotional clutter. The goal, he says, is greater transparency into what actually stops cyberattacks.

Rajesh Jha, Executive Vice President, Experiences + Devices, is retiring from Microsoft after more than 35 years. He will stay on until the end of the fiscal year and continue as an advisor to Chairman and CEO Satya Nadella. Learn more on the Official Microsoft Blog.

The Xbox team is deep in development on its next generation Xbox console, now called Project Helix. In a speech at the 2026 Game Developer Conference, Jason Ronald, the vice president of Next Generation at Xbox, said his team is “pushing the boundaries of rendering and simulation” for the upcoming gaming system. Read more from Ronald, including the latest on Xbox mode and Xbox Play Anywhere.

As companies increasingly turn to open AI models to avoid getting stuck with one vendor, many struggle to run them efficiently and securely at scale. Microsoft is tackling the problem by bringing Fireworks AI into Microsoft Foundry, combining faster performance for open models with Azure tools for evaluation, deployment, governance and security.

AI agents don’t succeed in isolation; they depend on unified data, across the cloud and applications. Microsoft announced Tuesday a new season of “The Shift,” a podcast that will explore how agents can be catalysts for innovation across the entire work environment, performing best when layers of the proverbial stack work together.

The show will release eight episodes this spring — one each week — that bring engineering, product and strategy perspectives together as a team.

As AI continues to show up in health care – everywhere from exam rooms to security operations – one thing is becoming clear: some organizations are redesigning how work gets done while others are still doing trial runs.

Research that Microsoft conducted with senior health care executives in the United States, published in the New England Journal of Medicine, revealed a growing readiness divide. The result is that while some systems are building governance, security and workforce models to  scale AI safely, others are being left behind.

Read more about how some health care organizations are embedding AI into core workflows to move beyond pilots and deliver real impact.