गूगल के CEO, Thomas Kurian ने बताया कि यह अमेरिका के बाहर कंपनी का सबसे बड़ा इनवेस्टमेंट होगा। कंपनी की योजना देश में 1 GW आउटपुट वाला डेटा सेंटर कैम्पस बनाने की है। इससे Google Cloud का क्लाइंट बेस बढ़ाने में सहायता मिल सकती है। यह डेटा सेंटर आंध्र प्रदेश के विशाखापट्टनम में बनाया जाएगा। अगले पांच वर्षों में यह इनवेस्टमेंट किया जाएगा।
Nvidia has announced the DGX Spark, touted as the world’s smallest AI supercomputer, which will start shipping on October 15. Priced at $3,999 (roughly Rs. 3.5 lakh), it offers one petaflop of AI performance and 128GB of unified memory. Despite its compact size, Nvidia says it can train and fine-tune large AI models efficiently.
Slack has announced new AI-powered features aimed at creating an “agentic OS” experience. Integrating Salesforce tools such as Sales, IT, HR, and Tableau, Slack now enables users to perform AI-driven tasks within chats. The platform will also roll out an enterprise search API, allowing secure cross-platform data searches via a single interface.
Microsoft has unveiled MAI-Image-1, its first in-house AI image generation model. The model debuted in the top 10 on LMArena’s text-to-image leaderboard and will soon be integrated into Microsoft products like Copilot and Bing Image Creator. It follows MAI-Voice-1 and focuses on producing photorealistic, fast-rendering visuals.
Google, on Tuesday, announced that it will invest $15 billion (1.3 lakh crore) in India to set up an artificial intelligence (AI)-focused data centre. Google Cloud CEO Thomas Kurian confirmed that this data centre will be built in Andhra Pradesh’s port city, Visakhapatnam. The investment will be made over a period of the next five years.
Users can now quickly generate webpage summaries through the Gemini overlay in Chrome for Android, following its initial launch last month in Chrome for Mac and Windows for US users. The Summarise Pages option appears above the Gemini prompt bar, between Share Screen with Live and Ask About Page. Selecting it produces a floating summary window, expandable with follow-up questions or read-aloud options. Powered by Gemini 2.5 Flash, the shortcut works across Chrome, Custom Tabs, Search, Discover, and Google News, streamlining summarisation without manual link pasting.
OpenAI has entered a multi-year partnership with Broadcom to co-develop custom AI accelerators capable of delivering up to 10 gigawatts of computing power. Under the deal, OpenAI will design the chips while Broadcom handles fabrication and networking. The move marks OpenAI’s second hardware venture, following its collaboration with Jony Ive on an AI device.
Google on Monday announced new features for NotebookLM to help users better understand complex topics. The company is bringing Nano Banana to Video Overviews, taking advantage of its generation speed, element-based editing, and consistent character. As per the company, there are six new visual styles in Video Overviews, including Anime, Heritage, and Retro Print. Alongside, Google is also rolling out two new formats to generate a detailed summary with the AI feature.
Anthropic has warned that even a few poisoned samples in a dataset can compromise an AI model. A joint study with the UK AI Security Institute found that as few as 250 malicious documents can implant backdoors in LLMs up to 13B parameters, proving model size offers no protection.
Slack has opened its platform for developers to build AI apps and agentic workflows using its conversational data. Powered by a real-time search API and an MCP-based server, it enables secure, context-aware integrations. Partners like Anthropic, Google, and Notion are already adopting it, with a wider rollout planned for 2026.
Microsoft Azure has launched the NDv6 GB300 VM series, the first supercomputing-scale cluster of Nvidia GB300 NVL72 systems. Designed for OpenAI’s demanding AI workloads, it features 4,600 GPUs, high-speed NVLink/NVSwitch connectivity, and 37TB of memory per rack. The system supports agentic and multimodal AI workflows, enabling model training in weeks instead of months while optimising throughput and efficiency.
OpenAI is reportedly developing a new “clinician mode” for ChatGPT, which could enable AI-assisted medical or health-related discussions. Code references also hint at a “model speaks first” voice feature that allows ChatGPT to greet users proactively. While these capabilities sound promising, OpenAI hasn’t confirmed their release yet.
Google has introduced Gemini Enterprise, a full-stack AI platform for businesses that integrates data from Workspace, Microsoft 365, Salesforce, and SAP. It enables companies to build custom AI agents using a no-code workbench for automated workflows. The system includes central governance and security tools to manage and audit AI agents across organisations.
Intel has unveiled more details about its upcoming Panther Lake chip, officially called Intel Core Ultra series 3. Built using the 18A process at its Arizona facility, it will feature a multi-chiplet SoC design, RibbonFET transistors, and Foveros packaging. The processor promises 50% faster performance, improved efficiency, and 180 TOPS AI capability.
Meta has expanded its Reels translation feature on Instagram and Facebook to include Hindi and Portuguese, in addition to the existing English and Spanish support. The AI-powered tool replicates the creator’s voice and offers a lip-sync option for natural-looking translations. Dubbed Reels carry a “Translated with Meta AI” label, and users can choose to enable or disable translations or view the original video by adjusting the audio and language settings.
At India Mobile Congress 2025, Samsung showcased its “AI for All” vision across five experience zones, highlighting Galaxy AI, SmartThings, and AI-powered security and education platforms. The company emphasised its commitment to building an inclusive, connected future through technology. Samsung also demonstrated its 5G and AI-driven network solutions at the event.
Razorpay and OpenAI, along with National Payments Corporation, are coming together to bring "India's first agentic payments" solution to the country. Using the company's payments stack, people will soon be able to search, explore, and buy groceries and other products directly through the artificial intelligence (AI)-powered chatbot. Currently, the feature is in the pilot stage. However, Razorpay is allowing people to sign up for early access.
At India Mobile Congress 2025, Ericsson showcased 5G and AI-led innovations, including an AI-powered robot dog that inspects driverless railcars for defects. It also unveiled Annapurti, a 5G-enabled grain ATM for efficient food distribution, and StethoECG, a wireless digital stethoscope with ECG support. Ericsson’s FWA tech now serves 8.4 million users in India.
OpenAI has updated its Sora app policies to give copyright holders and users more control. Rightsholders must now opt in, rather than opt out, for their characters to appear in AI-generated videos, meaning many copyrighted characters could disappear. Users can also restrict how their likeness is used through new Cameo preference settings.
Google has expanded its AI-powered “Try It On” feature to include shoes, letting users preview how footwear would look on them before buying. The tool generates realistic images by asking users to upload a full-length photo. The process mirrors Google’s apparel try-on tool, where users select a shoe listing, tap “Try It On,” and view or share the results. The update builds on Google’s earlier AI clothing try-ons, making online shopping more interactive and personalised.
Jio has unveiled its latest series of JioBharat Phones, which now come with multiple safety features. The company said that the handsets will be most useful for the caretakers of the elderly, parents of children, and women, as they ship with location monitoring capabilities, usage manager, and up to seven-day battery backup. Additionally, the company unveiled its new Jio AI Classroom course.
Vi has introduced Vi Protect, an AI-driven initiative focused on ensuring safety and security measures against the spams and cyber threats. As part of this initiative, Vi introduced two measures at India Mobile Congress (IMC) 2025. An AI-based Voice Spam Detection System and Cyber Defense and Incident Response System. The former detects and flags fraudulent and spam calls in real time.
OpenAI has banned several accounts, potentially linked to China, that attempted to use ChatGPT for mass surveillance, social media monitoring, and profiling. Some accounts also sought assistance in developing tools targeting Uyghurs and other sensitive topics. OpenAI confirmed no such tools were actually built, and accounts from Russia attempting phishing-related uses were also blocked.
Anthropic's new office will be located in Bengaluru, widely recognized as the technology hub of India, and operations will start early 2026, the company said. The location would serve as its second office in the Asia Pacific region after Tokyo.
Google is bringing Search Live in AI Mode to India, the company announced on Wednesday. It makes India the first country outside the US to receive the feature. Search Live in AI Mode, notably, is a two-way conversational feature which allows users to hold voice conversations with the AI chatbot. Alongside, Google also announced the expansion of AI Mode to seven new Indian languages.
IndiaAI Mission, under the Ministry of Electronics and Information Technology (MeitY), has selected five artificial intelligence (AI) projects aimed at making AI systems safer and fairer. Announced on Monday, these projects emerged from over 400 proposals received under the “Safe & Trusted AI” pillar launched in December 2024. The selected initiatives will receive funding and guidance to build tools for deepfake detection, bias mitigation, and AI security testing.
OpenAI hosted DevDay 2025 on Monday, unveiling several updates for users and developers. The biggest highlight is a new ChatGPT feature that lets users interact with third-party apps like Canva and Spotify without leaving the interface. The company also introduced AgentKit for developers to build and optimise agentic workflows, while GPT-5 Pro, Sora 2, and Sora 2 Pro are now available via API.
Elon Musk announced that xAI’s gaming division, xAI Game Studio, will release an AI-generated game before the end of next year. The company has also posted a job listing for a video game tutor to train its AI model, Grok, in gaming concepts. The announcement coincides with updates to Grok Imagine’s video generation capabilities.
Google DeepMind has introduced CodeMender, an AI agent that can automatically detect and fix security vulnerabilities in software code. It analyses source code, identifies flaws, and generates and verifies patches without developer input. In trials, CodeMender produced 72 verified patches across major open-source projects. Currently, its patches are reviewed by human researchers.
AMD and OpenAI have signed a multi-year partnership to build AI infrastructure, with OpenAI set to deploy 6GW of AMD GPUs over several years and chip generations, starting with Instinct MI450 GPUs in 2026. The deal could generate tens of billions in revenue for AMD. OpenAI can purchase up to 160 million AMD shares upon meeting deployment and share price milestones.
India-born global impact leader Kunal Sood has launched AudacityAI during the 80th United Nations General Assembly (UNGA). Announced at the Audacity UNGA 100 Disruptors Summit, the initiative aims to align artificial intelligence (AI) systems with human values. It seeks to guide AI through principles like compassion, service, and purpose to ensure technology benefits society.
OnePlus has announced the launch date for its new Android 16-based operating system. Dubbed OxygenOS 16, it will introduce a suite of artificial intelligence (AI)-enabled features. On top of this, it will also bring freshly designed app icons and other user interface elements. OnePlus recently announced that it will integrate Google's Gemini generative AI models with OxygenOS 16 to allow users to execute tasks directly on their handsets.
OpenAI’s long-rumoured AI device, developed in collaboration with former Apple design chief Jony Ive, is reportedly facing delays due to mounting technical and infrastructure challenges. While the hardware side of the project is progressing smoothly, a report suggests that software development and compute limitations are proving to be major hurdles, potentially pushing the launch timeline beyond 2026.
Meta, the Mark Zuckerberg-led tech giant, has announced that interactions between a user and the company's AI chatbot will soon be used for showing recommended content and ads. Users will have the option to choose the accounts from which the said data will be collected. The change will come into effect by the end of 2026, and Meta will start notifying users from next week. Both voice and text-based interactions will be used for displaying ads and content.
Over a month after Elon Musk’s xAI sued OpenAI for anticompetitive behaviour and stealing trade secrets, the San Francisco-based firm has issued its first public response. On Thursday, it published court documents addressing the claims while seeking dismissal of the case. OpenAI also called Musk’s lawsuit a “harassment tactic” and accused him of bullying its employees.
Perplexity has rolled out its AI-powered Comet browser for all users, CEO Aravind Srinivas announced on Thursday. Previously limited to Perplexity Max subscribers and invite-only access, the browser is now open on Windows and Mac, with mobile apps in development. A key feature of the browser is the sidebar AI assistant, which can summarise content, analyse web pages, and perform agentic tasks.
Google has introduced new smart home devices powered by its Gemini for Home artificial intelligence (AI) assistant. As per the company, the updated Nest Cam Indoor (3rd gen), Nest Cam Outdoor (2nd gen) and Nest Doorbell (3rd gen) come with 2K HDR video capabilities, built-in intelligent alerts, and up to a 166-degree diagonal view. Meanwhile, the new Google Home Speaker is claimed to be built for more natural conversations, powered by Gemini. It has support for 360-degree audio, Google TV Streamer pairing, and more features.
Google has announced a new feature for AI Mode which adds visual exploration capabilities. It enables users to search for and find something soecific that they cannot express in words. As per the company, they can further fine-tune the search results with prompts or ask follow-up questions about what’s on their screen. The new feature is rolling out in the English language for users in the users this week.
OpenAI has announced the Sora 2 video generation model, nearly a year after the first version’s debut. The new AI-powered tool will be accessible for free through an invite-only iOS app, designed in a TikTok-style format. Users can create AI videos of themselves, share them with friends, or use the “Cameo” feature to generate videos of others.
Nothing has announced Essential, it's vibe coding suite aimed at AI-centric devices. It allows users to generate their own Essential apps using natural language commands. These can be shared on Playground, a new community-driven hub. The existing Essential Space app has also been further enhanced with the introduction of Essential Memory.
WhatsApp is reportedly developing a new Incognito Mode feature for Meta AI chats, which could offer enhanced privacy and confidentiality. The feature was spotted in the latest WhatsApp beta for Android 2.25.28.1 update. When a user turns on Incognito Mode, the instant messaging service could show users a prompt warning that their conversations will not be recorded for future reference and personalisation.
The Adobe Premiere app for iOS is now live on the App Store. First listed for pre-order on September 4, it officially launched on Monday. While free to download, certain features need a paid subscription. Adobe says the app closely mirrors its desktop software, with tools like multi-track timeline and dynamic audio waveforms.
Anthropic has launched Claude Sonnet 4.5, calling it its most advanced coding model yet. The AI firm says it offers upgrades in coding, reasoning, agentic operations, and domain knowledge, with the ability to run tasks autonomously for up to 30 hours. The model is available via Claude’s website, apps, API, Claude Code, and a Chrome extension.
OpenAI has rolled out new parental controls for ChatGPT, weeks after first announcing them in response to concerns around minors’ safety. The controls let parents link accounts with teenagers, restrict features like Voice Mode and image generation, and set Quiet Hours. All teen accounts also get stricter safeguards by default, with OpenAI also building distress notifications for parents.
Google has announced the expansion of Gemini's features to the Drive apps on Android and iOS. This integration allow users to interact with their files and folders by asking questions and receiving conversational responses. The update lets you summarise documents, discuss topics, and draw details across files on mobile. Until now, the feature was limited to the web.
OpenAI has introduced agentic shopping capability to ChatGPT with the new Instant Checkout feature. Presently, it is available to all ChatGPT users in the US, whether they are ChatGPT Plus, Pro, or free users. Initially, people will be able to use the Instant Checkout feature for products listed by Etsy sellers. Soon, the company would also allow users to buy products within the chatbot from Shopify merchants, too.
WhatsApp has announced several new features for its app. Users can now share Live Photos on iOS and Motion Photos on Android via the app, complete with sound. Further, Meta AI can be used to generate chat themes. Similarly, background for video calls can be created using text prompts. WhatsApp has introduced new sticker packs, support for document scanning on Android, and an easier group search function.
Canva AI, the company’s suite of artificial intelligence (AI) tools, has expanded support to 16 new languages, including Hindi, Arabic, Chinese, and Spanish. Users can now write prompts and receive outputs in their native languages. The update also adds support for 31 locales, allowing more culturally relevant and context-aware designs across regions.
Apple is reportedly working on a new application, codenamed Veritas, that resembles ChatGPT. The new application is designed to test conversational capabilities ahead of a sweeping overhaul of Siri expected next year. Veritas is being used by Apple engineers to stress-test generative AI features, context understanding, and more human-like responses.
Adobe has integrated Google’s Gemini 2.5 Flash Image model, also known as Nano Banana, into its Firefly app’s text-to-image module. The premium AI tool, requiring a subscription, also extends to Firefly Boards, Photoshop’s Generative Fill, and Adobe Express. This integration bolsters Adobe’s suite with advanced image generation and editing powered by Google’s AI.
Google has upgraded its Gemini 2.5 Flash AI model, enhancing homework help, answer formatting, and image analysis. The chatbot now breaks down complex questions, presents responses with headers and lists for easier reading, and better interprets detailed images. These improvements are rolling out globally across the Gemini app and website, benefiting free, Pro, and Ultra tier users.
