As AI systems increasingly rely on multiple models working together to complete complex tasks, the computing infrastructure needed to run them has become more resource-intensive. Researchers at the Massachusetts Institute of Technology (MIT) and Microsoft have introduced Murakkab AI system, an automated system designed to improve how these AI workflows are built and deployed. By selecting the most efficient combination of AI models, software tools and computing resources, the system reduces computational demand, lowers cloud operating costs and cuts energy consumption without affecting performance.

Named after the Urdu word meaning “a composition of things,” Murakkab is designed to optimize the entire lifecycle of an agentic workflow, from application design to cloud deployment.

Murakkab AI System Simplifying Complex AI Workflows

Many modern AI applications are powered by agentic workflows, multi-step systems that coordinate several AI models and external tools, such as databases and programming environments, to complete tasks ranging from video analysis to code generation. Configuring these workflows typically requires developers to manually choose models, define execution sequences and allocate computing resources, making deployment both time-consuming and resource-intensive.

Murakkab, which will be presented at the upcoming USENIX Symposium on Operating Systems Design and Implementation, replaces much of that manual process. Developers only need to describe the application’s objective in plain language. The system then determines which AI models and tools are best suited for the task while selecting the most efficient hardware configuration and resource allocation for cloud deployment.

Optimizing Resources in Real Time

Beyond configuring workflows, Murakkab continuously adjusts computing resources during execution based on user-defined priorities, such as lowering deployment costs or improving response times.

In evaluations across multiple agentic workloads, the system required fewer computational units than conventional deployment approaches, reducing both energy consumption and infrastructure costs while maintaining comparable application performance.

Speaking during a press briefing, Gohar Chaudhry, lead author of the study, said the rapid growth of agentic AI systems has made resource optimization increasingly important.

“Agentic workflows are getting very complicated and quickly becoming the backbone of what cloud providers are doing. Energy usage is a huge concern, so we need to be very careful about how efficient these workflows are. It is very easy to over-allocate resources, wasting energy and money. Enabling a cloud provider to intelligently make these workflows more resource-optimal is a win for everyone involved,” Chaudhry said.

The research was conducted by Chaudhry along with Adam Belay, associate professor in MIT’s Department of Electrical Engineering and Computer Science and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL), Ricardo Bianchini, technical fellow and corporate vice president at Microsoft Azure, and other Microsoft Azure researchers.

Why Existing AI Workflows Waste Resources

Traditional agentic workflows require developers to make numerous technical decisions in advance, including selecting AI models, defining the sequence in which AI agents interact and choosing the hardware resources needed for deployment. Because these workflows often combine proprietary models and tools from multiple providers, introducing a newly released AI model or updating an existing workflow can require extensive reconfiguration.

The researchers said the enormous number of possible combinations makes manual optimization increasingly difficult.

“Even if you wanted to do all this manually, it is unlikely that you’ll be able to configure the workflow optimally because the space of possible configurations is so large,” said Chaudhry.

Murakkab also addresses a key challenge for cloud providers, which often over-allocate computing resources because they have limited visibility into how agentic workflows operate internally. The system automatically selects suitable hardware and adjusts resource allocation during execution, reducing unnecessary computation and energy use while maintaining performance.

Apple has announced a price hike in India for several of its products, including MacBooks, iPads, Apple TV and HomePod devices, as rising global memory chip costs driven by artificial intelligence (AI) infrastructure increase manufacturing expenses. iPhone prices remain unchanged.

The revised prices are now reflected on Apple’s India online store and come amid a global surge in demand for DRAM and NAND flash memory, essential components used in laptops, tablets and other consumer electronics.

MacBook Prices See Sharp Increase

Among the biggest revisions, the 13-inch MacBook Air (M5) now starts at ₹1,49,900, up from ₹1,19,900. The 15-inch MacBook Air (M5) has increased from ₹1,44,900 to ₹1,74,900.

Meanwhile, the 14-inch MacBook Pro now starts at ₹2,39,900, compared to its earlier price of ₹1,69,900. Premium MacBook Pro models equipped with the M5 Max chip have also witnessed price increases of up to ₹1 lakh.

iPad Prices Also Revised

Apple has also increased prices across several iPad models. The entry-level 11-inch iPad now starts at ₹49,900, up from ₹34,900, while the 11-inch iPad Air has risen from ₹59,900 to ₹74,900. The 11-inch iPad Pro now starts at ₹1,19,900, compared with ₹99,900 earlier.

Apple TV and Home Pod devices have also become more expensive, although the company has not revised prices for iPhones, Apple Watches or AirPods.

Why Has Apple Increased Prices?

According to Reuters, Apple attributed the revision to rising costs of memory components such as DRAM and NAND flash storage.

The rapid expansion of AI data centres has significantly increased demand for advanced memory chips, tightening global supply and driving up component prices. Industry analysts say manufacturers across the consumer electronics sector are facing higher production costs as AI infrastructure investment continues to accelerate.

Why iPhone Prices Remain Unchanged

Despite the latest revision, Apple has kept iPhone prices in India unchanged. Analysts believe the company may be waiting until the launch of its next-generation iPhone lineup before making any pricing changes to its smartphones. However, continued increases in semiconductor costs could influence future pricing decisions.

AI Boom Reshaping Consumer Electronics

The price hike in India highlights the wider impact of the AI boom on the technology industry. As companies invest billions of dollars in AI infrastructure and data centres, demand for high-performance memory chips has surged, increasing manufacturing costs for laptops, tablets and other electronic devices.

The development reflects a broader trend where AI is beginning to influence not only software innovation but also the pricing of consumer hardware worldwide.

“We have to upskill ourselves every six months now. Earlier, learning a new software was enough. Today, the competition is not just with other people. It is with AI.”

For Vishnu, a customer service professional at Infopark in Kochi, keeping pace with technological change has become part of the job. New AI-powered tools are increasingly handling routine customer queries, summarizing conversations and assisting with problem-solving—tasks that once relied entirely on human workers.

His experience reflects a broader shift taking place across industries. As artificial intelligence becomes more capable, workers are being pushed to continuously adapt, raising concerns about whether technological progress will create enough employment opportunities to match its economic gains.

The global economy is undergoing one of its most significant technological transformations since the internet age. Yet alongside optimism about innovation and productivity, policymakers and business leaders are grappling with a growing concern: jobless growth.

What happens in jobless growth?

The issue took centre stage this week at the World Economic Forum’s Annual Meeting of the New Champions, popularly known as “Summer Davos,” in Dalian, China. The gathering brought together more than 1,800 leaders from governments, businesses and academia from over 90 countries to discuss how emerging technologies can drive economic growth while ensuring that workers are not left behind.

A recurring theme throughout the summit was the need to prevent economic growth from becoming detached from job creation. While artificial intelligence is expected to improve productivity across sectors, leaders stressed that technology alone cannot guarantee employment opportunities. Investments in skills, education, entrepreneurship and workforce transition were repeatedly highlighted as essential to ensuring that innovation benefits a wider section of society.

The concern is not without basis.

According to the World Economic Forum’s Future of Jobs Report 2025, technological change is expected to transform 22 percent of jobs globally by 2030. The report estimates that while around 170 million new jobs could be created during this period, approximately 92 million existing jobs may be displaced, resulting in a large-scale restructuring of the labour market.

The report also found that nearly 59 percent of the global workforce will require reskilling or upskilling by 2030. Meanwhile, 41 percent of employers surveyed said they expect to reduce workforces where artificial intelligence can automate specific tasks, even as a majority indicated plans to invest in retraining employees.

Why is India significant?

Home to one of the world’s largest young populations, the country adds millions of job seekers to the workforce every year. At the same time, sectors such as information technology, customer support, finance and administrative services—areas where India has built a strong global presence—are among those experiencing rapid AI adoption.

Research by the International Labour Organization has suggested that generative AI is more likely to transform jobs than eliminate them entirely. Many occupations, particularly in clerical and support services, are expected to see specific tasks automated rather than whole roles disappearing. This means workers may increasingly find themselves collaborating with AI systems instead of competing directly against them.

That possibility has shifted attention toward preparedness rather than panic.

Rather than debating whether AI will change the nature of work, attention is increasingly shifting to how workers can be prepared for that change. Policymakers, educational institutions and employers are under growing pressure to ensure that people have access to the skills needed in an AI-driven economy. From digital literacy and vocational training to continuous learning opportunities, reskilling is emerging as a key part of the response.

The World Economic Forum echoed this sentiment in Dalian, emphasizing that the next phase of economic growth will depend not only on technological breakthroughs but also on investments in human capital.