ImagiNxt Edge - Your 7 Minute Edge to the Future of Tech

Google to build mega AI hub in India with $15 billion investment

Google is investing ten billion dollars in Andhra Pradesh. A new one-gigawatt data center and artificial intelligence hub will be built in Vishakhapatnam. This is Google's largest investment in India. The facility will boost India's AI and cloud computing capabilities. It is expected to create over 180,000 jobs. This project aims to make Visakhapatnam a global tech hub.

Google has announced an investment of $15 billion to build a 1-gigawatt data centre and artificial intelligence hub in the southern state of Andhra Pradesh. The project, to be located on the port city of Vishakhapatnam. As reported by Reuters, this is said to be Google’s largest investment in India to date and is poised to become a cornerstone of the country’s AI cloud ecosystem.

The leading Indian telecom operator Airtel has entered into partnership with Google to set up the facility. As reported by Reuters, the upcoming facility will integrate advanced AI infrastructure, large-scale energy systems and an expanded fibre-optic network. This initiative is said to boost India’s capacity to support next-generation AI applications, cloud computing and data-intensive services.

“In an era where data is the new oil, such initiatives will serve as a strategic advantage,” said Nara Lokesh, Andhra Pradesh’s IT minister.

Formal agreement and national backing

As per the Reuters report, a formal memorandum of understanding is said to be signed in New Delhi with top dignitaries including Andhra Pradesh Chief Minister N. Chandrababu Naidu, Union Finance Minister Nirmala Sitharaman, and Union IT Minister Ashwini Vaishnaw in attendance. The project is being hailed as a transformative step toward making Visakhapatnam a global tech hub. The investment is projected to generate over 180,000 jobs, both directly and indirectly, making it one of the largest foreign direct investments (FDI) in India’s tech sector.

The facility will not only cater to India’s growing data needs but also serve as a critical node in Google’s global AI infrastructure.

Airtel partners with Google to establish India’s first mega AI hub

Airtel has partnered with Google to set up India’s first Artificial Intelligence (AI) hub in Visakhapatnam, Andhra Pradesh. Airtel and Google will jointly establish the purpose-built data center in Visakhapatnam, as well as a state-of-the-art Cable Landing Station (CLS) to host Google’s new international subsea cables that will join its extensive global terrestrial and subsea infrastructure. Airtel will also create a robust intra-city as well as inter-city fibre network as a part of this project.

This high-capacity, low-latency network will deliver faster experiences to Google users and customers; increase the resilience and capacity of India's digital backbone; as well as drive digital inclusivity and transformation across India, bringing the benefits of AI to more people and businesses nationwide. Gopal Vittal, Vice Chairman and Managing Director, Bharti Airtel Limited, said, “This partnership with Google is a defining moment in India’s digital future. By combining world-class AI infrastructure with our nation’s extraordinary talent and also expanding global connectivity, we are laying the foundation for India to become a leader in the AI-driven era. With Visakhapatnam becoming a new hub on the world’s AI map, we are ensuring that India has the opportunity to set the pace for innovation, digital inclusion and economic growth, not just for our people, but for the world.”

MakeMyTrip just made searching for hotels, homestays easier with AI help

MakeMyTrip has launched Semantic Search, an AI-powered feature enabling travelers to find hotels and homestays using natural language queries. This advanced tool interprets complex, open-ended requests, moving beyond traditional filters to offer personalized results with visuals and AI-curated content, simplifying the accommodation discovery process.

MakeMyTrip, one of India’s leading online travel company, announced on Thursday that it has introduced Semantic Search, an advanced AI feature to aid customers find hotels and homestays. The new tool allows travellers to search for accommodation using natural language with simple descriptions, eliminating the need to use cumbersome filters.

Built on the same AI framework as the recently announced “Myra” trip assistant, the system is designed to interpret and act on open-ended queries, focusing on the traveler’s intent, the company said in a press release.

Moreover, the feature moves beyond rigid search parameters, allowing for highly specific requests.

Users can now input specific demands such as “pet-friendly hotels in Manali with a swimming pool” or “heritage stays near Jaipur forts with parking.” The AI processes these layered descriptions and generates relevant results, presenting them with property visuals, AI-curated user-generated content, and personalized rankings.

According to the travel giant, the system also includes AI smart suggestions, offering recommendations like “beachfront hotels in Goa” or “Ooty stays with mountain views,” complete with tags explaining their appeal (for example, “great for families” or “private pools available”).

“This aims to speed up the discovery and decision-making process for users. ‘We’re simplifying one of the most important parts of the traveler journey, finding the right stay,’ Ankit Khanna, Chief Product Officer at MakeMyTrip said. ‘Instead of forcing users to think like a system, we’ve built a system that understands how people naturally express themselves.’”

The Semantic Search tool follows the Beta launch of the generative AI-enabled trip planning assistant, Myra, which handles everything from discovery and booking to in-trip support. Myra is currently available in English and Hindi, with plans to expand to more Indian languages.

Dubai deploys fully autonomous AI traffic system to detect road violations in real time without human input

Dubai has launched an AI-powered Intelligent Traffic System, instantly detecting and documenting major violations like seatbelt and phone use. This automated approach frees up police for strategic duties, enhancing road safety and efficiency. The system provides real-time data and analytics, supporting Dubai’s smart city vision and setting a benchmark for urban traffic management.

Dubai has introduced a fully autonomous Intelligent Traffic System (ITS), leveraging artificial intelligence and real-time video monitoring to identify and document major traffic violations instantly. Unveiled at GITEX Global 2025, the system aims to transform road safety enforcement and streamline traffic operations across the emirate through automation and smart data use.

A smarter approach to road safety

Dubai Police’s newly launched ITS operates without human intervention and is designed to detect traffic offenses as they happen. Built as a unified digital platform, the system incorporates live surveillance feeds and advanced data analytics to monitor driver behavior and flag violations. The ITS represents a significant shift in how traffic laws are enforced, moving routine tasks from field officers to an AI-driven system. According to Lieutenant Engineer Ahmad Al Hammadi, the automation of these responsibilities frees up police personnel for more strategic duties, ultimately improving operational efficiency and decision-making. He described the system as “smarter, fairer, and more efficient,” underlining its role in supporting Dubai’s broader vision to become a globally recognised smart and safe city.

Real time detection of five critical violations

• Not wearing a seatbelt
• Using a mobile phone while driving
• Obstructing traffic flow
• Stopping in the middle of the road without valid reason
• Tailgating

Instant updates and in-depth analytics

One of the system’s standout features is its ability to provide live updates on the number of vehicles being monitored and the violations recorded at any given moment. This constant stream of data gives authorities an up-to-date view of traffic compliance levels across the city.

Supporting Dubai’s smart city goals

By automating key aspects of traffic enforcement, the platform not only reduces the margin for human error but also introduces a fairer, more transparent, and highly accountable mechanism for holding drivers responsible.

Citigroup’s AI usage frees up 100,000 hours for developers a week

Citigroup CEO Jane Fraser said the company’s use of artificial intelligence has saved time and allowed it to free up 100,000 hours of weekly capacity for software developers. Almost 180,000 of the bank’s employees in 83 countries have access to Citi’s internal AI tools, she told analysts on an earnings call on Tuesday.

Google advances AI-powered cancer research with Gemma-based foundation model

Google has launched a new 27-billion-parameter foundation model for single-cell analysis, built on the Gemma family of open models. As part of its research collaboration with Yale University, Google is releasing Cell2Sentence-Scale 27B (C2S-Scale), a large-scale model designed to understand the language of individual cells. Built on Gemma, C2S-Scale represents a new frontier in single-cell analysis.

This announcement marks a significant milestone for AI in science. C2S-Scale generated a novel hypothesis about cancer cellular behavior, which was subsequently confirmed through experimental validation in living cells. The discovery reveals a promising new pathway for developing therapies to fight cancer.

The launch builds on Google’s earlier work demonstrating that biological models follow clear scaling laws, similar to natural language models, where larger models deliver stronger performance in biological tasks. That research raised a critical question, do larger models merely improve at existing tasks, or can they acquire entirely new capabilities? The results with C2S-Scale suggest that the true promise of scaling lies in generating new ideas and uncovering previously unknown biological mechanisms.

How C2S-Scale 27B works

One of the major challenges in cancer immunotherapy is that many tumors are “cold,” meaning they are effectively invisible to the body’s immune system. A key strategy to make these tumors “hot” is to induce antigen presentation, a process that forces cancer cells to display immune-triggering signals.

To address this, Google tasked C2S-Scale 27B with identifying a drug that could act as a conditional amplifier, one that would boost immune signaling only in an “immune-context-positive” environment, where low levels of interferon (a critical immune-signaling protein) are present but insufficient on their own to induce antigen presentation. Solving this problem required a level of conditional reasoning that emerged only at scale; smaller models were unable to capture this context-dependent effect.

To enable this, researchers designed a dual-context virtual screening framework with two stages:

Immune-Context-Positive:

The model was provided with real-world patient samples featuring intact tumor-immune interactions and low-level interferon signaling.

Immune-Context-Neutral:

The model was provided with isolated cell line data lacking any immune context.

Using this framework, the team simulated the effects of more than 4,000 drugs across both contexts and asked the model to predict which compounds would selectively boost antigen presentation only in the immune-context-positive setting, prioritizing patient-relevant outcomes. Among the drug candidates identified, approximately 10–30% were already documented in prior literature, while the remaining candidates represented unexpected findings with no previously known link to the screening objective.

From prediction to experimental validation

The model’s predictions were clear. It identified a striking “context split” for the kinase CK2 inhibitor silmitasertib (CX-4945). The model predicted a strong increase in antigen presentation when silmitasertib was applied in an “immune-context-positive” setting, but little to no effect in an “immune-context-neutral” one. What made this prediction particularly compelling was that it represented a novel insight. Although CK2 has been implicated in many cellular functions, including modulation of the immune system, inhibiting CK2 via silmitasertib has not been previously reported in the literature to explicitly enhance MHC-I expression or antigen presentation. This demonstrated that the model was generating a new, testable hypothesis rather than reproducing known facts.

A prediction, however, is only valuable if it can be validated in practice. The first step in this process is experimental validation in the laboratory, followed ultimately by evaluation in clinical settings.

For the next phase of the project, researchers tested this hypothesis at the lab bench using human neuroendocrine cell models, a cell type that was entirely unseen by the model during training. The experiments showed:

• Treating the cells with silmitasertib alone had no effect on antigen presentation (MHC-I).
• Treating the cells with a low dose of interferon alone produced a modest effect.
• Treating the cells with both silmitasertib and low-dose interferon resulted in a marked, synergistic amplification of antigen presentation.

Notably, laboratory tests showed that the combination of silmitasertib and low-dose interferon produced an approximately 50% increase in antigen presentation, making tumors significantly more visible to the immune system.

The model’s in silico predictions were confirmed multiple times in vitro. C2S-Scale successfully identified a novel interferon-conditional amplifier, revealing a new potential pathway for converting “cold” tumors into “hot” tumors and increasing their responsiveness to immunotherapy. While this represents an early step, it provides a strong, experimentally validated lead for developing new combination therapies that use multiple drugs in concert to achieve more robust therapeutic effects.

Beyond this specific finding, the results offer a blueprint for a new approach to biological discovery. They demonstrate that by following scaling laws and building larger models such as C2S-Scale 27B, it is possible to create predictive models of cellular behavior that are powerful enough to run high-throughput virtual screens, uncover context-conditioned biology, and generate biologically grounded hypotheses.

Teams at Yale University are now investigating the underlying mechanisms identified in this work and testing additional AI-generated predictions across other immune contexts. With continued preclinical and clinical validation, these approaches have the potential to significantly accelerate the discovery and development of new therapies.

AI & Marketplaces: How Intelligence Is Rewiring Commerce at Scale

Marketplaces were once digital noticeboards, match buyers with sellers, process transactions, and scale through brute force. AI is changing that DNA. The most advanced marketplaces today no longer host commerce; they orchestrate it.

When AI sits at the edges, a recommendation widget here, a chatbot there, gains are incremental. But when intelligence is embedded at the core, marketplaces begin to behave like living systems. Discovery shifts from keyword search to intent prediction, where platforms infer what users want before they articulate it. Demand and supply stop reacting to each other and start moving in sync, powered by real-time forecasting that accounts for geography, seasonality, price elasticity, and behavioral signals. High-frequency decisions, pricing, promotions, routing, inventory placement, fraud checks, move from human-led heuristics to machine-led execution, happening thousands of times a second.

This shift matters most at scale. In markets like India, where fragmentation, volatility, and hyper-local nuance are the norm, human decision-making simply can’t keep up. AI thrives in this chaos. Every interaction becomes training data. Every edge case improves the system. The marketplace learns faster than competitors can copy.

The strategic insight for businesses is not to “use AI,” but to hand it responsibility. Assign AI ownership of repeatable, high-volume decisions. Let humans focus on system design, governance, and exceptions, not manual optimization. Over time, this creates tighter margins, faster feedback loops, and a compounding advantage that looks less like a feature and more like infrastructure.

The winners won’t be the marketplaces with the best UI or the most sellers.
They’ll be the ones where intelligence runs the market, and humans just set the rules.

‍

‍