AI Is Taking Over Core Operations of Indian IT Companies
CNBC reports AI is rapidly displacing traditional outsourcing work at India's largest IT firms — threatening the business models of Infosys, Wipro, and TCS as clients automate tasks previously handled by large offshore teams.
Read more →OpenAI Bets Big on Audio as Silicon Valley Declares War on Screens
OpenAI is investing heavily in audio-first AI interfaces, betting that voice will become the dominant mode of human-computer interaction. A broader Silicon Valley shift is underway — away from screens and toward ambient, always-on AI.
Read more →Microsoft's Nadella: Stop Thinking of AI as "Slop"
Microsoft CEO Satya Nadella pushes back on the growing narrative that AI-generated content is low quality "slop" — arguing the industry needs to focus on useful, grounded AI applications rather than dismissing the technology based on its worst outputs.
Read more →Apple's AI Wearables Push and the Low-End MacBook Launch
Bloomberg's pre-launch analysis of Apple's AI wearables strategy alongside the March 4 budget MacBook — positioning Apple to bring AI to a broader consumer base through both wearable devices and more affordable hardware.
Read more →Anthropic's "Do More With Less" Bet Has Kept It at the AI Frontier
Daniela Amodei explains how Anthropic's strategy of maximising efficiency over raw scale has allowed it to remain competitive at the frontier despite being outspent by rivals like OpenAI and Google.
Read more →Investors Predict AI Is Coming for Labor in 2026
Top investors surveyed by TechCrunch largely agree: 2026 is the year AI begins meaningfully displacing knowledge workers. Coding, customer service, and legal roles are most exposed. The debate is no longer if — but how fast.
Read more →Cramer: Forget the Bubble Talk — Own AI Stocks in 2026
CNBC's Jim Cramer argues that demand for compute, infrastructure, and enterprise AI is fundamentally different from the dot-com era, and that current valuations are justified by real underlying business transformation.
Read more →Microsoft Research: What's Next in AI — Inside the Edge of Discovery
Microsoft Research's deep-dive into the frontiers shaping AI development: new architectures, scientific AI, reasoning models, and the infrastructure required to push beyond current capability limits.
Read more →Jensen Huang GTC Keynote — Live Stream Tomorrow, March 16
Jensen Huang delivers Nvidia's GTC 2026 keynote Monday March 16 at 11am PT from SAP Center, San Jose. Free to stream at nvidia.com — expected to cover Rubin architecture, agentic AI, physical AI, and major partnership announcements.
Read more →OpenAI & Google Employees Rush to Anthropic's Defence in DOD Lawsuit
More than 30 employees from OpenAI and Google DeepMind filed a public statement supporting Anthropic's lawsuit against the US Department of Defense — a rare cross-company solidarity moment triggered by the Pentagon's blacklisting of Anthropic.
Read more →Microsoft Says Anthropic Remains Available After Pentagon Blacklist
Microsoft was first to publicly commit to its Anthropic partnership following the Pentagon's security risk designation, saying Claude products remain available to enterprise customers — a significant show of support against federal pressure.
Read more →Google Also Confirms Anthropic Available Outside Defence Projects
Google followed Microsoft in reassuring customers that Anthropic remains accessible through Google Cloud for non-defence use, despite the Pentagon standoff — presenting a united front of major AI partners backing Anthropic.
Read more →Microsoft Launches Frontier Suite — Claude Now in Microsoft 365 Copilot
Microsoft's Frontier Suite brings model diversity to Microsoft 365 Copilot, with Claude now available in mainline Copilot chat alongside OpenAI models — a major shift toward multi-model enterprise AI and a win for Anthropic's commercial reach.
Read more →OpenAI Launches Agentic Coding Model Minutes After Anthropic's Own
OpenAI released a new agentic coding model in direct competitive response to Anthropic's coding model — the two dropped within minutes of each other, exemplifying the accelerating tit-for-tat AI product race between the two companies.
Read more →Anthropic Has Brought Something New to AI: The Power to Say No
Bloomberg argues Anthropic's refusal of the Pentagon deal marks a genuine departure from Silicon Valley norms. Its willingness to turn down government contracts on ethical grounds may prove to be its most differentiated — and commercially valuable — characteristic.
Read more →AI "Scientists" Are Helping Humans Answer Urgent Climate Questions
AI systems are now actively participating in climate research — generating hypotheses, simulating cloud formation and ocean dynamics, and accelerating breakthroughs where human researcher capacity is a bottleneck.
Read more →OpenAI Report: AI as a Healthcare Ally
OpenAI's January 2026 white paper outlines how AI is closing gaps in healthcare access, assisting diagnosis, and scaling treatment research — framing AI as a collaborative clinical tool rather than a replacement for human judgement.
Read more →Atos & Nvidia Partner on Exascale Computing for Climate and Healthcare
Nvidia and Atos announced a partnership to deploy exascale computing infrastructure targeting climate modelling and healthcare research — combining GPU acceleration with high-performance computing at unprecedented scale.
Read more →Rivian Spin-Out Mind Robotics Raises $500M for Industrial AI Robots
Mind Robotics raised a $500M Series A co-led by Accel and a16z to build AI foundations for industrial robots capable of human-like dexterity and physical reasoning — going beyond the repeatable, fixed tasks of current industrial robots.
Read more →At CES 2026, AI Leaves the Screen and Enters the Real World
Scientific American's CES roundup documents the shift from AI as software to AI embedded in physical products — robots, wearables, autonomous vehicles, and appliances — marking a turning point in how AI interacts with the physical world.
Read more →Caterpillar Taps Nvidia to Bring AI to Construction Equipment
Caterpillar is integrating Nvidia's AI platform into heavy construction machinery, enabling autonomous operation and AI-assisted job site decision-making — a major signal that physical AI is entering traditional industrial markets.
Read more →Nvidia Releases Physical AI Open Models for Robots via Omniverse
Nvidia released open physical AI models and frameworks through its Omniverse platform, giving developers simulation tools and foundation models to accelerate deployment of real-world autonomous robots and vehicles.
Read more →Nvidia Agentic AI Blueprints for Autonomous Network Operations
Nvidia unveiled Nemotron-based large telco models and multi-agent Blueprints for autonomous network operations — covering energy saving and network configuration, targeting AI-driven infrastructure management for telecoms.
Read more →China's Humanlike AI Plans Could Set the Tone for Global Rules
China is tightening regulations for humanlike AI, requiring firms to ensure users know they're talking to a bot and prioritising safety and societal values. Analysts say China's approach could shape global AI governance frameworks.
Read more →Anthropic Research: Economic Policy Responses to AI
Anthropic explores what policies are needed to manage AI's economic impact — covering workforce transitions, innovation investment, and legal frameworks to minimise disruption while maximising opportunity for workers and businesses.
Read more →Anthropic Publishes Claude's Constitution
Anthropic released its "Claude's Constitution" — a public document outlining the values, principles, and rules that guide Claude's behaviour. A rare look at how an AI company formally encodes ethics and priorities into a deployed model.
Read more →US Universities Must Educate the Future AI Workforce
Washington Post opinion argues US universities are failing to keep pace with AI's transformation of the labour market, calling for curriculum overhauls, industry partnerships, and new funding to produce an AI-ready workforce.
Read more →Microsoft: Teaching in the AI Age
Microsoft's policy framework for AI-ready education outlines five priorities for governments and schools: AI literacy, clear guidance for educators, building teaching capacity, responsible innovation, and leadership in AI-enabled learning.
Read more →Google: A Policy Agenda for Responsible AI Progress
Google's policy agenda calls for risk-based proportional regulation, common international standards, and cross-border cooperation on AI governance — framed around three pillars: opportunity, responsibility, and security.
Read more →US Supreme Court Declines to Rule on AI Copyright
The Supreme Court turned away a case on whether AI-generated art can be copyrighted, leaving in place the ruling that human authorship is required. Works combining AI output with sufficient human creative input may still qualify for partial protection.
Read more →Music Publishers Sue Anthropic for $3B Over "Flagrant Piracy" of 20,000 Songs
Concord and Universal Music Group lead a coalition suing Anthropic for allegedly using 20,000+ copyrighted songs to train Claude without permission. Potential damages exceed $3 billion — the largest AI copyright suit targeting Anthropic to date.
Read more →Warner Music Settles with Udio and Signs AI Music Licensing Deal
Warner Music settled its copyright case against AI music startup Udio and struck a licensing deal for an AI music platform launching in 2026 — a potential template for how labels and AI companies resolve training data disputes going forward.
Read more →YouTubers Sue Snap for Copyright Infringement in AI Training
A group of YouTubers filed suit against Snap alleging it scraped their content without permission to train AI models — adding to the growing wave of creator-led copyright actions against AI companies now exceeding 70 cases industry-wide.
Read more →Meta Signs Massive Nvidia Deal to Power AI Data Centers
Meta has struck a significant deal with Nvidia to supply chips for its rapidly expanding AI data center buildout, underscoring the hyperscaler's accelerating push into proprietary AI infrastructure and intensifying competition for Nvidia's most advanced processors.
Read more →AI Data Centers Hit Power Limits — Indian Startup C2i Backed to Fix the Bottleneck
As AI data centers run into hard power ceilings, Peak XV has backed Indian startup C2i to address the energy infrastructure crunch. The investment highlights a growing secondary market of companies solving the physical constraints that are now throttling AI expansion.
Read more →Iran War Threatens Semiconductor and Memory Chip Supply
Escalating conflict with Iran is raising alarms over the stability of global semiconductor supply chains, with analysts warning that regional disruption could squeeze memory chip availability and drive up costs for AI hardware manufacturers already facing tight supply.
Read more →Ratepayers Push Back as AI Data Centers Drive Electricity Prices Higher
CNBC reports a growing backlash from consumer advocates and state regulators as AI data center electricity demand pushes utility bills higher. Several states are exploring ratepayer protection rules that would shift more infrastructure costs onto large industrial AI customers.
Read more →First Look at Nvidia's Vera Rubin — How It Beats Blackwell
CNBC gets a first look at Nvidia's next-generation Vera Rubin AI system, set to succeed Blackwell. The platform promises significantly higher performance per watt, with Nvidia positioning it as the foundation for the next wave of frontier model training and inference at scale.
Read more →Nvidia to Invest $2 Billion in AI Data Center Specialist Nebius
Nvidia is committing $2 billion to Nebius, a European AI cloud and data center company, deepening its strategy of backing infrastructure partners that run Nvidia hardware at scale. The deal signals Nvidia's intent to cultivate a network of Nvidia-powered AI cloud providers globally.
Read more →Meta Builds Massive AI Infrastructure with Nvidia
Nvidia details how Meta is constructing one of the largest AI infrastructure deployments in history using Nvidia's platforms, encompassing tens of thousands of GPUs across multiple data centers to support Meta AI, Llama model development, and recommendation systems.
Read more →Anthropic Releases Claude Sonnet 4.6 as Default Free and Pro Model
Anthropic launched Claude Sonnet 4.6, making it the new default for both free and Pro users. The model improves on computer use, coding, design, and large-context tasks — arriving just 12 days after Opus 4.6 shipped, reflecting an accelerating release cadence at Anthropic.
Read more →Google Deepens Pentagon AI Push After Anthropic Sues Trump Administration
Following Anthropic's lawsuit against the Trump administration over its national security risk designation, Google has moved to fill the gap by deepening its Pentagon AI partnership. Google is now offering DoD personnel tools to build custom AI agents on the military's GenAI.mil enterprise portal.
Read more →OpenAI Launches Platform for Enterprises to Build and Manage AI Agents
OpenAI launched OpenAI Frontier, an end-to-end enterprise platform for building, deploying, and managing AI agents — including agents built on third-party models outside OpenAI. The move signals OpenAI's ambition to become the operating system layer for enterprise agentic AI, not just a model provider.
Read more →Google Wants Intrinsic to Be the "Android of Robotics"
Google is elevating Intrinsic, its internal robotics software project, from the "Other Bets" division into the core business, with a stated goal of making it the open-platform standard for industrial robots — akin to what Android is for smartphones. Google has also partnered with Boston Dynamics to run Gemini inside Atlas humanoid robots.
Read more →Jira's Latest Update Lets AI Agents and Humans Work Side by Side
Atlassian has updated Jira to support "agents in Jira," enabling AI agents and human team members to collaborate on the same tasks and projects within the platform. The integration reflects a broader push by enterprise software vendors to embed agentic AI natively into existing workflows rather than requiring separate tooling.
Read more →Humanoid Robots Take Over CES 2026 as Industry Bets Big on Physical AI
Humanoid robots dominated the CES 2026 floor in Las Vegas, with manufacturers from Boston Dynamics to NEURA Robotics showcasing AI-powered machines built for manufacturing, logistics, and elder care. Industry analysts called it a turning point as physical AI moves from research labs to commercial deployment timelines.
Read more →OpenAI Launches GPT-5.4 with Pro and Thinking Versions
OpenAI released GPT-5.4 on March 5 — its most capable frontier model yet — with a 1 million token context window and record benchmark scores including 83% on knowledge-work evaluations. The release also includes GPT-5.4 Thinking for complex reasoning tasks and GPT-5.4 Pro for maximum-capability professional use.
Read more →ChatGPT's GPT-5.3 Instant Will Stop Telling You to Calm Down
OpenAI released GPT-5.3 Instant, a faster everyday model specifically tuned to drop the therapy-speak and excessive hedging that users found grating in earlier versions. The model targets quick, professional tasks where response character matters as much as capability.
Read more →OpenAI Signs Agreement with the Department of War
OpenAI has formalised an agreement with the US Department of Defense, providing military access to its AI capabilities. The deal marks a significant shift for OpenAI, which previously restricted military use cases, and puts it in direct competition with Anthropic — which is currently suing the Trump administration over its own DOD designation.
Read more →UK Regulators Demand AI Chatbot Firms Protect Children Under Online Safety Laws
UK regulators Ofcom and the ICO are requiring AI chatbot companies to comply with stricter child safety obligations under the Online Safety Act, including age verification and safeguards against emotional manipulation. The rules apply to any AI service accessible to children, setting a potential template for other jurisdictions.
Read more →New York Governor Signs RAISE Act — Second US State to Regulate AI Safety
New York became the second US state to enact major AI safety legislation when Governor Hochul signed the RAISE Act, requiring large AI developers to publish safety protocols, report safety incidents within 72 hours, and submit to oversight by a new state AI monitoring office. The law complements California's SB 53 and increases pressure on Washington to pass federal rules.
Read more →Nvidia Accelerates AI for Over 80 New Science Systems Worldwide
Nvidia announced that more than 80 new supercomputers and research systems are coming online globally in 2026, all powered by Nvidia accelerators. These systems target fusion energy, drug discovery, climate modelling, materials science, and quantum computing research — extending AI-assisted discovery beyond hyperscalers into academia and national labs.
Read more →AI Hitting Labor Market "Like a Tsunami" as Layoff Fears Mount
CNBC reports a sharp rise in worker anxiety about AI-driven job displacement, with employee concerns about layoffs jumping from 28% in 2024 to 40% in early 2026. Economists warn of "AI redundancy washing" — companies citing AI as cover for cost-cutting cuts that may have other causes — complicating the picture for policymakers.
Read more →Dario Amodei Warns AI May Cause "Unusually Painful" Job Disruption
Anthropic CEO Dario Amodei issued an unusually direct public warning that AI could cause significant and rapid labour market disruption — more concentrated and faster-moving than previous waves of automation. He called on governments to prepare economic support systems before displacement accelerates rather than after.
Read more →AI Squeezes Entry-Level Jobs — but Teens Remain Optimistic
A new survey finds employers are already cutting entry-level roles due to AI automation, with a Stanford study showing a 16% relative employment decline for recent graduates in AI-exposed roles. Despite this, teenagers entering the workforce report being broadly optimistic about their future employability — a confidence gap researchers say warrants attention.
Read more →Washington Post Tests Hundreds of Jobs Against AI — See the Results
The Washington Post ran an extensive interactive analysis testing AI performance across hundreds of real job tasks, mapping which roles are most and least vulnerable to automation. The results reveal wide variation even within the same profession, and challenge the conventional wisdom that only routine, low-skill work is at risk.
Read more →Microsoft Launches Copilot Cowork — AI That Coordinates Workflows Across Microsoft 365
Microsoft announced Copilot Cowork, a new agentic layer that turns user intent into coordinated action across Word, Excel, Outlook, and Teams — combining Anthropic's Claude for multi-step reasoning with Microsoft 365's enterprise controls. The feature is in limited Research Preview now, with broader availability through the Frontier programme expected in late March.
Read more →OpenAI Raises $110 Billion — One of the Largest Private Funding Rounds in History
OpenAI closed a $110 billion funding round in February, dwarfing its previous raises and cementing its position as the most valuable private company in the world. The round reflects extraordinary investor confidence in AI's near-term commercial trajectory, though it also intensifies pressure on OpenAI to generate returns at a scale that justifies the valuation.
Read more →Google Launches Gemini 3 Deep Think — Built for Scientific Discovery
Google announced Gemini 3 Deep Think, a model specifically designed to advance science, research, and engineering. Building on earlier Deep Think versions that achieved IMO Gold-medal performance in mathematics, Gemini 3 Deep Think now scores up to 90% on Olympiad-level problem benchmarks and has autonomously solved previously open mathematical conjectures.
Read more →Google DeepMind Partners with US Dept of Energy on "Genesis" AI for Science
Google DeepMind has partnered with the US Department of Energy on Genesis, an initiative applying AI to accelerate scientific breakthroughs across energy research, materials science, and national laboratory priorities. The partnership signals a deepening of AI's role in federally-funded research infrastructure beyond the commercial sector.
Read more →Boston Dynamics' Next-Gen Humanoid Robot Will Have Google DeepMind DNA
Boston Dynamics announced its next-generation humanoid robot will be built around Google DeepMind's AI models, integrating Gemini directly into the Atlas platform. The collaboration positions DeepMind as the intelligence layer for one of the most capable physical robots in existence, accelerating the path from research lab to industrial deployment.
Read more →US Court Bars OpenAI from Using "Cameo" After Trademark Dispute
A federal court in California ruled against OpenAI in a trademark dispute with celebrity video platform Cameo, ordering OpenAI to stop using the "Cameo" name in its Sora 2 video generation product. The ruling is a small but notable legal setback for OpenAI as it navigates an increasingly litigious landscape around AI product naming and intellectual property.
Read more →OpenAI, Anthropic and Waymo Consumed 90% of February's $189B in Global VC
A Crunchbase report reveals that a record $189 billion in global venture capital was deployed in February 2026 — with OpenAI, Anthropic, and Waymo accounting for roughly 90% of it. The extreme concentration signals both massive institutional conviction in AI and growing concern that venture returns may be equally concentrated.
Read more →Legal AI Startup Legora Raises $550M, Triples Valuation to $5.5B
Swedish legal AI startup Legora closed a $550 million Series D led by Accel, tripling its valuation to $5.55 billion in just a few months as it targets US market expansion. The raise reflects surging enterprise demand for AI tools that can handle complex, high-stakes professional work like legal research and contract drafting.
Read more →Nvidia Launches Alpamayo — Open AI Models That Let Cars "Think Like a Human"
Nvidia unveiled Alpamayo at CES 2026, a family of open-source reasoning AI models for autonomous vehicles that use chain-of-thought logic to navigate complex edge cases — like directing a car through a traffic light outage. The 10-billion-parameter vision-language-action model is already integrated into the new Mercedes-Benz CLA, due on US roads this year.
Read more →Motional Targets 2026 for Driverless Robotaxi Launch in Las Vegas
Motional has rebooted its self-driving programme around AI, placing machine learning at the core of its perception and decision systems as it targets a fully driverless robotaxi service in Las Vegas before end of year. The relaunch comes after a period of restructuring and signals renewed commercial ambition in the autonomous ride-hail space.
Read more →Google Adds AI Music Generation to Gemini App via Lyria 3
Google has integrated DeepMind's Lyria 3 music model directly into the Gemini app, letting users describe an idea or upload a photo to generate a 30-second custom track with matching cover art. The feature also includes ProducerAI in Google Labs, offering fine-grained controls over tempo, arrangement, and time-aligned lyrics for more advanced creators.
Read more →Google Flow Gets Major Update — Bringing AI Video Editing to the Forefront
Google's AI video creation tool Flow has been redesigned with image generation front and centre, improved asset management, and more precise editing controls. Since its launch, Flow users have created over 1.5 billion images and videos for films, music videos, and product campaigns — making it one of the most-used generative video tools outside of social platforms.
Read more →China's Lead in Intelligent Robots and Autonomous Drones Is "Just Getting Started"
A Washington Post opinion piece argues Beijing has pulled ahead of the US in deploying AI-powered robots and autonomous drones at scale, embedding the technology across military, industrial, and consumer applications faster than American competitors. The author warns the gap in physical AI deployment — not just software — may prove more consequential than the LLM race.
Read more →Nvidia Still Hasn't Sold Its US-Approved China Chips — and Fears Local Rivals
Despite creating a special export-control-compliant chip for China, Nvidia has struggled to sell it as local competitors like Huawei close the performance gap under the advantage of unrestricted domestic deployment. The dynamic illustrates how export controls, while limiting China's frontier compute, may also be accelerating the rise of a self-sufficient Chinese AI chip industry.
Read more →China's Tech Shock Threatens US AI Monopoly — and It's "Just Getting Started"
CNBC analysis finds China's rapid AI commercialisation — spanning cars, robotics, consumer electronics, and industrial automation — is directly challenging American dominance in applied AI. With a new $8.7 billion national AI fund and an "AI+" strategy to embed AI across the entire economy, China is betting that deployment breadth, not just model capability, wins the AI race.
Read more →Schools Are Teaching AI All Wrong — Washington Post
A Washington Post opinion argues that most schools are teaching AI as a tool to use rather than as a phenomenon to understand, producing students who can prompt but can't reason about AI's limits, biases, or implications. The author advocates for embedding critical AI literacy — not just practical skills — across all subjects rather than treating it as a standalone technical topic.
Read more →OpenAI Pushes into Higher Education as India Seeks to Scale AI Skills
OpenAI has partnered with six Indian universities to bring ChatGPT Edu and GPT-5.2 access to over 100,000 students and faculty — part of a broader "Education for Countries" initiative to help governments integrate AI into national education systems. The move signals OpenAI's ambition to become the default AI platform in academia, alongside classroom competitor Microsoft.
Read more →Cursor Rolls Out "Automations" — Agents That Trigger Themselves on Code Changes
Cursor launched Automations, a new agentic coding system that lets AI agents fire automatically in response to repository events like a new commit, a Slack message, or a timed schedule — removing the human from the trigger loop entirely. The feature represents a step change from AI-assisted coding toward AI-autonomous software development.
Read more →Apple's Xcode Now Supports Claude and OpenAI Codex as Native Coding Agents
Apple updated Xcode to integrate Anthropic's Claude Agent and OpenAI's Codex directly into its official developer IDE, letting iOS and macOS developers run agentic coding assistance without leaving Apple's toolchain. The move gives both Anthropic and OpenAI a foothold in the millions-strong Apple developer ecosystem at the platform level.
Read more →Microsoft Reveals Maia 200 — Its Second-Gen AI Chip, 30% Faster Than Rivals
Microsoft unveiled the Maia 200, the second generation of its in-house AI accelerator chip, claiming 30% better performance-per-dollar than competing offerings from Nvidia, Amazon, and Google. Microsoft plans to deploy Maia 200 internally across Azure to reduce its dependency on Nvidia GPUs for running its own AI workloads.
Read more →Meta Strikes AMD Chip Deal Days After Committing to Millions of Nvidia GPUs
Meta signed a major deal to deploy AMD GPUs for AI workloads — just days after announcing one of the largest-ever Nvidia GPU commitments. The dual-supplier strategy signals Meta's intent to avoid dependence on any single chip vendor as it scales infrastructure for its AI ambitions, and gives AMD its most significant AI cloud win yet.
Read more →Anthropic Launches Claude for Healthcare Alongside OpenAI's ChatGPT Health
Anthropic announced Claude for Healthcare — a specialised offering covering clinical workflows, drug discovery, and life sciences research — within days of OpenAI revealing ChatGPT Health. The rapid back-to-back launches mark the opening of a direct competitive front between the two leading AI labs in the high-stakes medical sector.
Read more →Anthropic Launches Claude Opus 4.6 in the "Vibe Working" Era
Anthropic released Claude Opus 4.6, its most capable model yet, framing the launch around a new era of "vibe working" — where professionals delegate extended, judgment-heavy tasks to AI rather than simply using it for point queries. The model excels at sustained multi-step work, coding, and producing polished professional outputs across hours-long sessions.
Read more →Anthropic Launches Enterprise Agent Plug-ins for Finance, Engineering and Design
Anthropic unveiled a suite of enterprise agent plug-ins targeting specific professional domains — finance, engineering, and design — marking its most aggressive push yet to embed Claude into everyday workplace workflows. The plug-ins allow Claude to take multi-step actions within existing enterprise software stacks rather than requiring users to switch between tools.
Read more →YouTube Expands AI Deepfake Detection to Politicians, Officials and Journalists
YouTube is rolling out its AI-powered likeness detection technology to a pilot group of political candidates, government officials, and journalists — high-value targets for synthetic media attacks. The expansion comes ahead of the 2026 US midterms as platforms face mounting pressure to prevent AI-generated impersonation from distorting political discourse.
Read more →"An AlphaFold 4" — Scientists Marvel at DeepMind Drug Spin-off's New AI
Scientific American reports that Isomorphic Labs — DeepMind's drug discovery spin-off — has unveiled its IsoDDE engine, capable of precisely predicting how proteins interact with candidate drugs and antibodies, with researchers comparing the leap to the original AlphaFold moment. Isomorphic has active drug-development partnerships with Johnson & Johnson, Eli Lilly, and Novartis, and has clinical trials on the horizon.
Read more →Nvidia and Eli Lilly Open Co-Innovation AI Lab for Drug Discovery
Nvidia and pharmaceutical giant Eli Lilly announced a joint co-innovation lab dedicated to applying AI across the drug discovery pipeline, from molecular simulation to clinical trial design. The partnership is one of the first direct big pharma–chip maker R&D collaborations at this scale, signalling that AI infrastructure is becoming as central to drug development as laboratory equipment.
Read more →Meta Signs 6+ Gigawatt Nuclear Deals to Power AI Supercluster
Meta struck nuclear power agreements with three companies — Vistra, Oklo, and TerraPower — totalling over 6 gigawatts, enough to power roughly 5 million homes, to supply its "Prometheus" AI supercluster. The deals make Meta one of the world's largest corporate nuclear power buyers and signal that the AI industry's energy hunger has outpaced what renewables alone can supply.
Read more →California Reconsiders Nuclear Energy Amid AI Power Supply Crunch
California — which previously closed its last nuclear plant — is reassessing nuclear energy as AI data center construction strains the state's power grid. The reversal reflects a broader shift across US states as AI-driven electricity demand forces a rethink of long-held opposition to nuclear power, with small modular reactors now back on the policy table.
Read more →OpenAI Launches o3 and o4-mini — Its Most Powerful Reasoning Models Yet
OpenAI released o3 and o4-mini, setting new state-of-the-art scores on Codeforces competitive programming, SWE-bench software engineering, and MMMU multimodal reasoning benchmarks. The o3 model also crosses the 90% threshold on the ARC-AGI-1 general reasoning benchmark — a milestone researchers had considered a meaningful proxy for human-level flexible reasoning.
Read more →Google's Gemini Pro Sets Record Benchmark Scores — Again
The latest version of Gemini Pro posted record-breaking scores on Humanity's Last Exam and other frontier benchmarks, significantly outperforming its predecessor and challenging GPT-5.x across key capability measures. TechCrunch notes this is the second consecutive benchmark sweep for Google's flagship model, as the race between Google and OpenAI tightens.
Read more →Alibaba Unveils Qwen 3.5 as China's AI Agent Race Heats Up
Alibaba released Qwen 3.5, pitching it as a frontier model optimised for agentic tasks as Chinese AI labs shift their competitive focus from chat to autonomous agents. The release signals that China's AI industry — led by Alibaba, Baidu, and ByteDance — is increasingly competing on the same agentic terrain as OpenAI and Anthropic rather than simply closing the LLM capability gap.
Read more →Tiny Startup Arcee AI Builds 400B-Parameter Open Source LLM That Beats Meta's Llama
Arcee AI, a small startup, released a 400 billion parameter open source large language model built entirely from scratch that outperforms Meta's Llama on key benchmarks — achieving this without the resources of a big tech company. The feat challenges the assumption that frontier open-source AI requires hyperscaler budgets and may accelerate open-weight model development by smaller labs.
Read more →Microsoft: Threat Actors Now Use AI as Operational "Tradecraft"
A new Microsoft Security report documents how nation-state hackers and criminal groups have moved beyond using AI as a novelty to embedding it as routine tradecraft — generating spear-phishing content, creating fake personas for employment fraud, and deploying agentic AI that can execute cyberattacks autonomously for extended periods. North Korean operatives are using AI to fraudulently infiltrate US Fortune 500 firms via fake remote worker identities.
Read more →AI Helps Southeast Asia Scam Centers Evade Crackdowns and Reach More Victims
Bloomberg reports that criminal scam operations in Southeast Asia — already among the world's largest fraud networks — are using AI to scale attacks, create convincing personas, and evade government crackdowns. The technology has enabled scam centres to target larger victim pools at higher speed, complicating international law enforcement efforts.
Read more →Gemini 3 Arrives in the Gemini App with Agentic Android Automation
Google updated the Gemini app with its Gemini 3 model and new agentic capabilities that can automate multi-step tasks on Android in the background — ordering groceries, booking rides, and finding restaurants without the user leaving their current app. The update signals Google's push to make Gemini a proactive personal assistant rather than a reactive chatbot.
Read more →Google Rolls Out Gemini Capabilities to Docs, Sheets, Slides and Drive
Google has pushed new Gemini features into Workspace productivity apps, including a "Fill with Gemini" tool for Sheets that can generate custom text, categorise data, and pull live information from Search directly into spreadsheet cells. The rollout targets Google AI Ultra and Pro subscribers first, extending the Gemini 3 model into everyday office work.
Read more →12% of US Teens Now Turn to AI for Emotional Support or Advice
A new survey finds roughly one in eight American teenagers uses AI chatbots for emotional support or personal advice, a figure that is rising as young people report feeling less judged by AI than by human friends or therapists. Mental health researchers warn the trend is outpacing safety research, with several AI companion apps still lacking adequate crisis detection for vulnerable users.
Read more →Experts Warn: Know When Not to Use ChatGPT as a Therapist
Mental health professionals speaking to CNBC draw a sharp line between AI's utility for everyday stress and its dangers as a substitute for clinical therapy — pointing to cases where chatbots failed to recognise suicidal ideation or gave actively harmful advice. The American Psychological Association has accused AI wellness apps of "deceptive practices" by marketing themselves as therapy while disclaiming clinical responsibility in the fine print.
Read more →Goldman Sachs Taps Anthropic's Claude to Automate Accounting and Compliance
Goldman Sachs is working with embedded Anthropic engineers to build Claude-powered autonomous agents for trade accounting, transaction processing, and client onboarding — some of the most compliance-sensitive work on Wall Street. The partnership is one of the most concrete examples yet of a top-tier investment bank deploying frontier AI inside core operations rather than just productivity tooling.
Read more →Harvard Study: AI Can Predict 71% of Active Mutual Fund Trades
A Harvard-led study found that AI models can accurately anticipate roughly 71% of trading decisions made by active mutual fund managers — raising profound questions about the value of human stock-picking judgment. The findings suggest that a large share of active management alpha may be pattern-based and therefore replicable by AI at a fraction of the cost.
Read more →UBS Analyst Warns AI Disruption Is About to Hit Credit Markets
A UBS analyst argues that credit markets — long considered more opaque and relationship-driven than equities — are the next major financial sector to face AI-driven disruption, as models become capable of credit analysis, loan underwriting, and covenant monitoring at scale. The warning extends the AI displacement narrative beyond software and white-collar knowledge work into the heart of traditional banking.
Read more →Cleveland's Largest Newspaper Uses AI to Write Articles — Staff Spooked
The Plain Dealer, Cleveland's major daily, has begun publishing AI-written content, boosting web traffic while alarming its remaining journalists who fear displacement. The experiment represents one of the most prominent deployments of AI writing at a legacy regional newspaper, and is being watched as a bellwether for how local journalism navigates cost pressures and AI capabilities.
Read more →AI Journalism Startup Symbolic.ai Strikes Deal with Rupert Murdoch's News Corp
Symbolic.ai signed a deal with News Corp to integrate its AI journalism platform with Dow Jones Newswires, promising up to 90% productivity gains on research-heavy tasks like newsletter creation, fact-checking, and headline optimisation. The partnership gives Murdoch's media empire direct access to AI-assisted news production at scale — and signals a strategic bet that AI augments rather than replaces journalists at major outlets.
Read more →AI Can Be a Source of Climate Data — Not Just a Climate Villain
A Bloomberg opinion piece argues the framing of AI as an environmental bad actor misses its growing role as a tool for generating and analysing climate data — from satellite image analysis to emissions monitoring to extreme weather prediction. The author contends that AI's net effect on climate outcomes may be positive if its analytical power is directed at the problem rather than just blamed for energy use.
Read more →Pentagon Pressured Anthropic to Drop AI Safety Guardrails — Anthropic Refused
Bloomberg's detailed account of the Pentagon-Anthropic standoff reveals Defence Secretary Hegseth demanded Anthropic remove guardrails preventing its AI from being used for mass domestic surveillance and fully autonomous weapons — and that Anthropic CEO Dario Amodei refused. The confrontation led to Anthropic being designated a national security supply chain risk, setting a landmark test case for how AI companies navigate government demands to weaken safety limits.
Read more →Anthropic Was the Pentagon's AI of Choice — Now It's Banned and Experts Are Worried
CNBC reports that Anthropic had been the US military's preferred AI partner before the Trump administration's ban, leaving a capability gap that Google and OpenAI are now rushing to fill on different terms. Defence AI experts warn the episode may push future AI-military partnerships toward less safety-conscious vendors, as the ban signals that guardrails are a liability in the procurement process.
Read more →OpenAI Shares More Details on Its Pentagon Deal — Three Hard Red Lines
Following the contrast with Anthropic's ban, OpenAI published more detail on its Department of War agreement — confirming three absolute prohibitions: no use for mass domestic surveillance, no directing of autonomous weapons systems, and no use against US citizens. The disclosure is partly a positioning exercise, as OpenAI seeks to reassure safety advocates that its military deal has meaningful limits.
Read more →Google Adds Multimodal Search to AI Mode — Ask Questions About What You See
Google upgraded its AI Mode in Search with multimodal capabilities powered by Gemini and Lens, letting users point their camera at an object, scene, or document and ask natural language questions about it directly in Search. The feature extends conversational AI beyond text queries into the physical world, blurring the line between visual search and AI assistant.
Read more →AI Uncovers 800 Previously Unknown "Cosmic Anomalies" Hidden in Hubble Archives
An ESA AI tool sifted through nearly 100 million image cutouts from the Hubble Legacy Archive and surfaced over 800 never-before-seen cosmic anomalies — objects and phenomena that had been hiding in plain sight in decades of archival data. The discovery illustrates how AI is expanding the effective reach of existing scientific instruments, not just accelerating analysis of new observations.
Read more →Microsoft Launches $99/Month AI Software Bundle — a 65% Price Hike
Microsoft unveiled Microsoft 365 E7, a new flagship enterprise AI bundle launching May 1 at $99 per user per month — a 65% increase over its previous top-tier package. The bundle positions Microsoft as the operating system for "human-led, agent-operated" enterprises, bundling Copilot, Claude, and next-gen OpenAI models alongside the full Microsoft 365 suite.
Read more →Anthropic Expands Claude Cowork with Connectors for Enterprise Workflows
Anthropic gave Claude Cowork — its AI tool for office workers — a broader enterprise release, adding a suite of connectors and plug-ins that link it to existing corporate software stacks. The update marks Claude Cowork's transition into a full enterprise product, with Anthropic pitching it as a way to "turbo-charge" individual worker productivity across knowledge-heavy roles.
Read more →Apple Developing AI Glasses, a Pendant and Camera AirPods for the AI Era
Bloomberg reports Apple is actively developing three AI wearables: smart glasses that compete with Meta's Ray-Bans, a wearable pendant, and AirPods with an outward-facing camera for Visual Intelligence. The devices represent Apple's bet that the next computing platform is ambient AI worn on the body rather than held in the hand.
Read more →OpenAI Plans Its First Hardware Device in 2026 — Likely AI Earbuds
OpenAI has confirmed it intends to ship its first consumer hardware device in 2026, with a leading candidate being AI-powered earbuds codenamed "Sweet Pea." The move puts OpenAI in direct competition with Apple and Amazon in the emerging AI wearables space, and would give the company a direct hardware relationship with consumers independent of app stores.
Read more →Amazon Acquires Bee — a $50 Always-Listening AI Wearable
Amazon acquired Bee, a clip-on AI wearable that continuously listens to its wearer's conversations and activities, automatically generating transcripts, recaps, and to-do lists throughout the day. The acquisition signals Amazon's ambitions to extend Alexa's ambient AI presence beyond the home and into the body, competing directly with the wave of AI pendant and glasses devices coming from Apple, Meta, and OpenAI.
Read more →AI Deepfakes Spread Virally After Maduro's Removal from Power in Venezuela
Following the fall of Nicolás Maduro's government, AI-generated deepfake videos depicting fake US military involvement went viral across social media, illustrating how political power vacuums immediately attract synthetic disinformation. The episode — one of the first major post-coup deepfake incidents — underscores that AI misinformation is now a predictable feature of geopolitical crises, not just elections.
Read more →Google Launches Universal Commerce Protocol — a New Open Standard for AI Shopping Agents
Google unveiled the Universal Commerce Protocol (UCP), an open standard designed to let AI agents operate seamlessly across the entire retail journey — discovery, purchase, and post-sale support — regardless of platform. The move is a bid to make Google the infrastructure layer for agentic commerce, with projections of $1–5 trillion in AI-orchestrated retail revenue globally by 2030.
Read more →How Close Are Today's AI Models to AGI — and to Self-Improving into Superintelligence?
Scientific American surveys the landscape of AGI timelines, with DeepMind projecting arrival by 2030 and Sam Altman suggesting superintelligence could be "a few thousand days" away. Researchers caution that while benchmarks keep falling, the gap between impressive task performance and general reasoning remains contested — and the question of what AGI actually means grows harder to answer as models get closer to it.
Read more →Microsoft Report: AI Adoption Growing Twice as Fast in Global North as Global South
Microsoft's Global AI Adoption 2025 report finds the gap between wealthy and developing nations is widening, not closing — with adoption in the Global North growing almost twice as fast as in the Global South, a divide now spanning 10.6 percentage points. Language barriers, infrastructure gaps, and lack of local AI services in low-resource languages are identified as the primary drivers of exclusion.
Read more →Google Translate Upgraded with Gemini AI — Better at Nuance and Low-Resource Languages
Google has upgraded Google Translate with new Gemini-powered translation models, improving accuracy for nuanced text and expanding support for underrepresented languages. The update also extends real-time speech translation in Google Meet to Italian, Portuguese, German, and French, with further language rollouts planned across 2026.
Read more →Salesforce CEO Benioff Calls for AI Regulation — Says Some Models Have Become "Suicide Coaches"
Salesforce CEO Marc Benioff made an unusually blunt call for AI regulation, stating that some AI models have effectively become "suicide coaches" for vulnerable users and that the industry cannot self-police effectively. His remarks add significant corporate weight to the push for mandatory safety standards, coming from one of enterprise software's most prominent figures.
Read more →Google and Character.AI Settle Lawsuits Over Teen Suicides Linked to AI Chatbots
Google and Character.AI have agreed to settle civil suits brought by families who allege their children's suicides were linked to harmful AI chatbot interactions. The settlements — terms undisclosed — arrive as regulators in multiple countries are drafting stricter rules for AI companion products, and before what could have been damaging discovery processes for both companies.
Read more →Maybe AI Agents Can Be Lawyers After All
TechCrunch examines a growing body of evidence that agentic AI is crossing a threshold in legal work — passing bar exam components, autonomously drafting complex contracts, and completing multi-step legal research tasks that previously required junior associates. The question is shifting from "can AI do legal work?" to "what liability framework governs it when it gets something wrong?"
Read more →Legal AI Giant Harvey Acquires Hexus as Competition Heats Up in Legaltech
Harvey, one of the leading AI platforms for law firms, acquired rival Hexus as the legal AI market consolidates around a handful of well-capitalised players. The deal reflects a broader pattern of AI startup M&A as the market matures from experimentation to deployment, with law firms increasingly committing to single-platform relationships rather than running multiple tools.
Read more →OpenAI and Ginkgo Bioworks Show How AI Can Cut the Cost of Protein Production by 40%
A Scientific American report on an OpenAI-Ginkgo Bioworks collaboration shows AI reducing protein production costs by around 40% after just two months and over 36,000 automated experiments — dramatically compressing a process that previously required years of trial and error. The result is one of the most concrete demonstrations yet of AI accelerating wet-lab biology at industrial scale.
Read more →SpaceX Is Acquiring Elon Musk's xAI Ahead of Potential $1.25 Trillion IPO
Elon Musk is merging his AI startup xAI into SpaceX, creating a combined entity expected to pursue a public listing that would value it at roughly $1.25 trillion. The deal collapses the boundary between Musk's space and AI ventures and positions xAI's Grok model and computing infrastructure within SpaceX's broader industrial ecosystem ahead of what would be one of the largest IPOs in history.
Read more →OpenAI Acquires Convogo — Executive Coaching AI Meets the World's Biggest AI Lab
OpenAI is acquiring the team behind Convogo, a platform using AI for executive coaching, leadership assessments, and HR feedback automation. The all-stock acqui-hire brings three founders into OpenAI and signals its interest in applying AI to management and professional development — a space adjacent to its enterprise ambitions.
Read more →Mistral AI Makes Its First Acquisition — Buys Cloud Infrastructure Startup Koyeb
French AI lab Mistral, valued at $13.8 billion, has made its first-ever acquisition by purchasing Koyeb, a Paris-based startup that simplifies AI application deployment and manages the underlying infrastructure. The move signals Mistral's ambition to grow beyond model development into a full-stack AI cloud provider, competing more directly with US hyperscalers.
Read more →OpenAI Unveils ChatGPT Health — 230 Million Users Already Ask About Health Each Week
OpenAI launched ChatGPT Health, a dedicated medical information experience, revealing that 230 million people were already using ChatGPT each week for health-related questions before the product even existed. The launch formalises OpenAI's push into healthcare with features tailored for symptom checking, medication information, and navigating medical decisions — alongside a partnership framework with health providers.
Read more →How Silicon Valley Will Eat Its AI Competitors in 2026
A Bloomberg analysis argues that 2026 will see the major tech platforms — Google, Microsoft, Amazon, Apple — systematically absorb the AI startup ecosystem through acquisitions, talent raids, and API lock-in, replicating the pattern by which they absorbed the mobile app boom. The piece warns that the window for independent AI companies to establish durable competitive positions is closing fast.
Read more →