“AI's Shifting Landscape: Power, Policy, and Productivity"
As AI continues to evolve, the tech world is witnessing both breakthroughs and barriers. Kompact AI, a lightweight solution from Ziroh Labs and IIT Madras, promises decentralized power by running AI without costly chips—potentially revolutionizing accessibility. Meanwhile, U.S. export restrictions on advanced AI chips send shockwaves through markets, denting giants like Nvidia and AMD. On the productivity front, Microsoft pushes boundaries with Copilot Studio’s new ‘Computer Use’ feature, empowering AI to seamlessly interact with web and desktop tasks. Together, these developments reflect a rapidly transforming AI ecosystem—balancing innovation, regulation, and real-world application.
Kompact AI, developed by Ziroh Labs and IIT Madras, can run AI models without heavy computing chips.
Ziroh Labs, the homegrown AI startup, has developed a new computing platform that claims to reduce reliance on costly and scarce GPUs for running large AI models. The system, called Kompact AI, has been built in collaboration with the Indian Institute of Technology (IIT) Madras, and is intended to allow advanced AI systems to operate on CPUs found in standard computers.
Kompact AI is part of a larger effort undertaken by the Centre of AI Research (COAIR), jointly established by IIT Madras and Ziroh Labs. The Centre focuses on developing AI systems that are accessible, affordable, and suited for real-world use in regions with limited computing resources. This also aligns with India’s broader national vision of “AI for All,” which seeks to foster inclusive AI development that can operate in diverse environments without the heavy infrastructure requirements seen in the West.This development comes at a time when access to advanced AI infrastructure has become a key concern for developers, particularly in countries like India, where resources for large-scale computing are limited. The global demand for GPUs has pushed up costs and made them difficult to acquire, even for large companies. IIT Madras Director V.
Kamakoti commented on this, saying, “The AI divide is because only those with high-end expensive GPU-powered resources can access, develop and deploy powerful AI,” he said. “We’re demonstrating that you don’t need a revolver to kill a mosquito.” Thus, it seems that the most immediate impact of Kompact AI is its potential to level the playing field in AI development, ensuring that smaller startups, educational institutions, and researchers in developing countries cease to struggle to meaningfully contribute to the AI landscape by enabling large AI models to run efficiently on standard CPUs (thus lowering the barrier to entry for organizations without deep pockets).tested by major US chip manufacturers, including Intel and AMD, to validate its performance.
The platform has reportedly been optimized to handle 17 widely used AI models on CPU-based systems. According to the startup, models with parameters under 50 billion typically used in various business and research settings perform well on Kompact AI without compromising accuracy. William Raduchel, a former executive at Sun Microsystems and an advisor to Ziroh Labs, suggested that it will have a very profound market impact in the years ahead.”
Ziroh Labs’ Kompact AI is designed specifically to focus on AI inference — the process of using trained AI models to perform tasks like answering queries or generating content. While GPUs are still critical in training large models due to their ability to process vast amounts of data simultaneously, Kompact AI aims to run these models after they have been trained, using the far more commonly available CPUs. This approach was demonstrated at an event held recently, where Ziroh Labs showcased AI models like Meta’s Llama 2 and Alibaba’s Qwen2.5 operating on a laptop powered by an off-the-shelf Intel Xeon processor. According to the company, the system could process queries effectively without specialized hardware.
US tech stocks fall as AI chip export curbs hit Nvidia, AMD and others.
US tech stocks experienced a significant plunge in the past couple days, primarily driven by fresh export restrictions on AI chips and escalating trade tensions under President Donald Trump’s administration. The Nasdaq Composite fell 1.8%, while the S&P 500 declined 0.9%, marking a sharp downturn as markets reacted to newly announced US tariff pressures.
In fact, semiconductor major Advanced Micro Devices (AMD) has revealed it anticipates an $800 million charge due to new US export controls requiring licenses for AI chip sales to China. These restrictions specifically impact AMD’s MI308 chips, which are integral to its AI product lineup. Even, AMD’s shares fell 6.1% today, closing at $89.51.“On April 15, 2025, AMD completed its initial assessment of a new license requirement implemented by the United States government for the export of certain semiconductor products to China (including Hong Kong and Macau) and D:5 countries, or to companies headquartered or with an ultimate parent in such countries.
The Export Control applies to the Company’s MI308 products. The Company expects to apply for licenses, but there is no assurance that licenses will be granted. The Company expects that the Export Control may result in charges of up to approximately $800 million in inventory, purchase commitments and related reserves,” the company’s filing with the SEC stated.Meanwhile, another leading AI chipmaker, Nvidia, disclosed a projected $5.5 billion revenue loss originating from the indefinite licensing requirements imposed on its H20 chips. In response to these developments, Nvidia’s stock dropped 6.3%. Notably, these chips previously generated a significant portion of Nvidia’s $17 billion revenue from China last year.
These measures have intensified investor concerns, leading to a widespread sell-off in the tech sector. Companies like Broadcom and Marvell Technology also saw their stocks decline by 6.3% and 19.8%, respectively. Also, Apple’s stock closed at $197.43, down 0.0233% from the previous close. Amazon’s stock also witnessed a decline of around 0.0174% and closed at $176.46.
Microsoft Copilot Studio gains ‘Computer Use’ allowing AI agents to do tasks on web and desktop apps
Microsoft has introduced a new capability – called Computer Use – in its Copilot Studio platform that allows AI agents to perform tasks across both web and desktop apps.
Unveiled as part of an early research preview, Computer Use allows AI agents to simulate the way humans interact with software—by clicking buttons, navigating menus, and entering text directly into fields. According to Microsoft’s Corporate Vice President of Business and Industry Copilot, Charles Lamanna, the feature is easy to use and effectively bridges gaps in automation where system APIs are either unavailable or insufficient.
“AI innovation is accelerating at an unprecedented pace, and Microsoft Copilot Studio is at the forefront—integrating the best AI advancements into a platform built to solve business challenges at scale. Last month, we introduced deep reasoning capabilities for agents, support for model context protocol (MCP), and the general availability of agent flows in Copilot Studio,” the blog post read. “Today, we are excited to announce that computer use is coming to Copilot Studio through an early access research preview. This new capability allows your Copilot Studio agents to treat websites and desktop applications as tools. With computer use, agents can now interact with any system that has a graphical user interface!”
Computer use opens the door for enterprises to deploy AI across a broader range of software environments. From internal data entry workflows to market research and external websites lacking developer support, Copilot agents can now carry out tasks while adapting to changes in interface design and layout in real time, according to the company. Microsoft says these agents will automatically detect and adjust to modifications in user interfaces, ensuring processes are not interrupted by routine updates or redesigns.
The rollout of Computer Use follows Microsoft’s earlier launch of a consumer-oriented Copilot feature named Actions, which permits users to perform background tasks such as making reservations or purchasing tickets. However, unlike Actions, the Studio version of Computer Use offers broader applicability and is not restricted to a select group of partner platforms. This feature is also similar to autonomous AI frameworks developed by OpenAI and Anthropic. OpenAI’s “Operator” and Claude’s identically titled “Computer Use” both offer AI agents the capacity to perform tasks across software environments. Microsoft’s version, however, extends functionality to include desktop apps in addition to browser-based tools, giving it a potentially wider range of use cases in enterprise settings.
Still, it is part of Microsoft’s growing suite of AI-powered services and capabilities. These include the Copilot Vision tool an AI assistant for web browsing and the Researcher and Analyst features in Microsoft 365 Copilot, which combine natural language querying with third-party data integration. The Analyst feature, for instance, runs on OpenAI’s o3-mini reasoning model and supports Python scripting for real-time data queries and visualizations. The Copilot Vision feature, recently made available to Windows Insiders, can analyze web content with user permission, offering contextual insights and simplifications. Meanwhile, the Researcher tool merges Microsoft’s AI systems with platforms such as Salesforce and Confluence, bringing powerful search and synthesis capabilities into workplace software.
