Microsoft is preparing for a significant shift with Windows 12, which reports suggest could launch this year or by late 2026. This new operating system, potentially codenamed 'Hudson Valley Next,' is expected to be heavily AI-focused and may require specialized AI chips, known as NPUs, for its full capabilities. These NPUs would need to deliver at least 40 TOPS of performance, potentially limiting the functionality of older PCs. Chip manufacturers like Intel and AMD are already gearing up for this requirement, which could lead to increased PC costs but promises faster, smarter performance.
Windows 12 is also rumored to feature a modular design based on the CorePC architecture, allowing users to customize features. Some advanced AI services might even be offered through a subscription model, alongside a potential visual overhaul including transparent elements and a floating taskbar. Meanwhile, the broader implications of AI are drawing scrutiny, with Palantir CEO Alex Karp warning Silicon Valley that if AI displaces many white-collar jobs, governments might nationalize the technology to serve national security interests and prevent societal unrest.
The strategic importance of AI is further highlighted by concerns that a potential ban on Anthropic's AI tools by a future Trump administration could jeopardize the US military's AI advantage over China. Anthropic's AI has been tested in military operations, and its removal could slow progress. In enterprise, EY is deploying AI software agents called Droids to over 5,000 engineers globally, integrating them into tools like GitHub and Jira for tasks such as code refactoring. This represents one of the largest enterprise uses of AI agents in production, aiming to accelerate innovation.
Supporting this AI expansion, CoreWeave and Perplexity have partnered, with CoreWeave Cloud providing GPU infrastructure for Perplexity's AI inference workloads, helping the AI-powered answer engine scale. For developers, LangWatch has open-sourced a platform offering tools for evaluating, tracing, simulating, and monitoring AI agents, addressing their non-deterministic nature. In specific industries, Swiftex showcased its AI Revenue Orchestration Engine at the e4m Auto Summit, aiming to tackle the automotive industry's $2.6 million annual loss from unconverted leads by improving follow-up times and boosting lead-to-test-drive conversions by 3.2x.
However, the reliability of AI tools themselves is under question, as seen in Colombia where the Supreme Court rejected an appeal claiming AI authorship of a ruling, only for the appellant's lawyer to find a 93% probability of AI generation using the court's own detection software. Separately, health policy makers are set to learn about AI's role in healthcare at a March 26, 2026 webinar, discussing both its potential benefits like improved outcomes and efficiency, and risks such as bias and safety concerns, with a focus on AI adoption in Europe.
Key Takeaways
- Microsoft's upcoming Windows 12, expected this year or by late 2026, may require AI chips (NPUs) with at least 40 TOPS for full functionality, impacting older PCs and driving chip development by Intel and AMD.
- Windows 12 is anticipated to feature a modular design (CorePC architecture) and could offer advanced AI services via subscription, alongside a visual refresh.
- Palantir CEO Alex Karp warns that widespread AI-driven job displacement could lead to government nationalization of AI technology, citing concerns about private industry power and national security.
- A potential ban on Anthropic's AI tools by a future US administration could compromise the US military's AI advantage, as Anthropic's technology has been tested in military operations.
- EY has deployed AI software agents, called Droids, to over 5,000 engineers globally, integrating them into existing tools like GitHub and Jira to accelerate software development and address technical debt.
- CoreWeave and Perplexity have partnered, with CoreWeave providing specialized GPU infrastructure to support Perplexity's AI inference workloads and enable scaling of its AI-powered answer engine.
- Swiftex's AI Revenue Orchestration Engine addresses the automotive industry's $2.6 million annual loss from unconverted leads, demonstrating a 3.2x increase in lead-to-test-drive conversions for clients.
- The reliability of AI detection tools is under scrutiny after a Colombian court's ruling, deemed AI-free by its own software, was found to have a 93% probability of AI generation by the same tool when used by an appellant's lawyer.
- LangWatch has open-sourced a platform for evaluating, tracing, simulating, and monitoring AI agents, providing tools to address their non-deterministic nature and streamline development.
- Health policy makers will attend a webinar on March 26, 2026, to understand AI's benefits (improved outcomes, efficiency) and risks (bias, safety) in healthcare, focusing on European adoption.
Windows 12 May Need AI Chips for Top Performance
Microsoft is hinting that Windows 12, expected in late 2026, might require special AI chips called NPUs for its full capabilities. These chips, capable of at least 40 TOPS, will help run AI features smoothly. This means older PCs might have limited functions or not work with the new OS. The move could increase PC costs but promises faster, smarter performance for users and businesses. Chip makers like Intel, AMD, and Qualcomm are preparing for this shift.
Windows 12 Could Be Modular and AI-Focused
Reports suggest Microsoft plans to release Windows 12 this year, featuring a modular design and a strong focus on AI. The new operating system, codenamed 'Hudson Valley Next,' may require NPUs for full performance, potentially excluding millions of users with older computers. Its modular structure, based on the CorePC architecture, will allow users to add or remove features for customization. Some advanced AI services might also be offered through a subscription model, alongside a potential visual overhaul including transparent elements and a floating taskbar.
Palantir CEO Warns of AI Nationalization Over Job Loss
Palantir CEO Alex Karp warned Silicon Valley that if AI eliminates many white-collar jobs, the government might nationalize the technology. He argued that societies won't accept powerful private industries displacing educated voters while not serving national security interests. Karp's comments come amid government scrutiny of AI companies, like the recent issues with Anthropic. He believes ignoring the political impact of AI job losses could lead to government intervention and loss of control over AI technology.
Swiftex AI Engine Tackles Auto Sales Losses, Plans Global Growth
Swiftex showcased its AI Revenue Orchestration Engine at the e4m Auto Summit in New Delhi, addressing the automotive industry's $2.6 million annual loss from unconverted leads. The platform aims to improve lead follow-up times, which currently average 47 hours, by ensuring leads receive timely attention. Swiftex clients have seen a 3.2x increase in lead-to-test-drive conversions. The company is preparing for international expansion with upcoming presentations at conferences in Houston and Dubai.
Colombian Court's AI Ruling Questioned by Its Own Detector
Colombia's Supreme Court of Justice rejected an appeal based on claims that a ruling was written by AI. The court stated its AI detection tools found no evidence of AI authorship. However, the lawyer for the appellant ran the court's own ruling through the same software and found a 93% probability it was AI-generated. This incident raises doubts about the reliability of AI detection tools and the court's reliance on them in legal proceedings.
Health Policy Makers Learn About AI's Role in Healthcare
A webinar on March 26, 2026, will help health policy makers understand artificial intelligence in the health sector. The event, titled 'Demystifying artificial intelligence in health,' will discuss AI's potential benefits, such as improved outcomes and efficiency, as well as its risks, including bias and safety concerns. Speakers from the London School of Hygiene and Tropical Medicine, Royal Free London NHS Foundation Trust, and the Department of Health in Catalonia will share insights on AI adoption in Europe and key considerations for decision-makers.
US Military AI Edge Threatened by Anthropic Ban
A potential ban on Anthropic's AI tools by the Trump administration could jeopardize the US military's AI advantage over China. Anthropic's AI has been tested in military operations, providing valuable experience. Removing or replacing this technology could slow down the US military's progress in AI development. The dispute centers on how and when Anthropic's tools are used, with the Pentagon concerned about restrictions hindering operations. While alternative AI systems exist, the loss of Anthropic's battle-tested technology might impact the US's lead in military AI.
CoreWeave to Power Perplexity's AI Inference Workloads
CoreWeave and Perplexity have signed a multi-year partnership where CoreWeave Cloud will support Perplexity's AI inference workloads. Perplexity, an AI-powered answer engine, will use CoreWeave's specialized GPU infrastructure to efficiently run its large language models. This collaboration aims to help Perplexity scale its operations and innovate in the AI field. Both companies will also explore co-developing new AI services. CoreWeave focuses on providing high-performance cloud infrastructure for AI, while Perplexity offers accurate, source-cited answers.
EY Deploys AI Agents for Software Development at Scale
EY has rolled out agent-native development across its global engineering teams, deploying AI software agents called Droids to over 5,000 engineers. This marks one of the largest enterprise uses of such agents in production. The Droids are integrated into existing tools like GitHub and Jira, handling tasks such as code refactoring and documentation updates. Engineers maintain architectural control and final approval, allowing for delegation without compromising quality or security. This initiative aims to accelerate innovation and address technical debt within EY's large codebases.
LangWatch Open Sources AI Agent Evaluation Tools
LangWatch has released an open-source platform to address the non-deterministic nature of AI agents. The platform provides tools for evaluation, tracing, simulation, and monitoring, moving AI development towards a systematic approach. It allows developers to run end-to-end simulations, pinpointing failures in agent reasoning and tool usage. LangWatch also streamlines the process of converting traces into test cases for evaluation and prompt optimization. Built using OpenTelemetry, it integrates with various AI frameworks and model providers, supporting GitOps workflows for prompt version control.
Sources
- Windows 12 May Require AI Chips for Full Performance, Microsoft Signals
- Windows 12 Reportedly Set for Release This Year as a Fully Modular, Subscription-Based, AI-Focused OS
- Silicon Valley Is “Retarded” If It Thinks AI Won’t Be Nationalized: Palantir CEO Alex Karp
- Swiftex Showcases AI Revenue Orchestration Engine at e4m Auto Summit, Prepares for Global Expansion at Houston and Dubai Conferences
- Colombian Court Rejects Appeal for AI Writing, Then Gets Flagged By Its Own AI Detector
- Demystifying artificial intelligence in health: What health policy-makers need to know
- Anthropic ban may threaten the military's AI advantage over China
- CoreWeave Announces Agreement to Power Perplexity’s AI Inference Workloads
- Engineering for an Agent-First Era: How EY is Redefining Enterprise Development with Factory
- LangWatch Open Sources the Missing Evaluation Layer for AI Agents to Enable End-to-End Tracing, Simulation, and Systematic Testing
Comments
Please log in to post a comment.