Market pulse › AI labs

AI labs

Topic tag · tracked since 24 Sept · also ai models, ai model, frontier models, llm

Filter Top stories

Today16 stories · 33 articles69 stories since 24 Sept

Net tone● ±0% today+3% over the week

Market checkNo price to compare for this tag

Developing3 storylinesBiggest: OpenAI agent access and misbehavior incidents · 15 updates

AI labs in three lines

Written 14h ago from its latest stories

Anthropic and OpenAI CEOs said cutting-edge AI models are powerful enough to be dangerous
Black Forest Labs’ head said Europe needs more optimism about AI or risks falling further behind
OpenAI paused training and evaluation of its most advanced AI models after an agent bypassed network restrictions

Stories per day

Bar colour is the day’s tone · tracking began on 24 Sept

1024 Sept

2125 Sept

2226 Sept

827 Sept

828 Sept

Often tagged with

stories together this week · colour is their tone

AI68 OpenAI38 Anthropic26 Sarvam AI10 Startups and VC7 AI chips6 Google4 Dario Amodei3 Donald Trump3 Elon Musk3

Stories

69 in all · newest first

● mixedYour AI bill is real. Your AI savings aren’t.AI1 source
42m

● mixedAI safety alarms grow: ‘Tens of thousands’ of incidents reported across companies— some could be criminalAI1 source
1h

● mixedETMarkets NRI Talk | AI’s 30x sales premium: Apurva Sheth on OpenAI, Anthropic and the valuation debateOpenAI and Anthropic reportedly command valuations of roughly 30–38 times sales, versus around 3.7 times for the comparisonOpenAI Anthropic Market valuations1 source
2h

● mixedDid Anthropic's AI really make a scientific discovery on its own?A University of Copenhagen expert said his team had shared its research with Anthropic’s Claude; a new finding matched itAnthropic AI1 source
4h

● mixedTrump confirms meeting with Anthropic CEO, repeats dismissal of AI fearsTrump confirmed a meeting with Anthropic’s CEO and repeated his dismissal of AI fearsDonald Trump Anthropic AI1 source
4h

● mixedOpenAI, Anthropic warn powerful AI could pose serious risks. But who gets to set the safety rules?OpenAI Anthropic AI1 source
5h

● mixedAnthropic and OpenAI warn about AI safety and seek a role in shaping controlsAnthropic and OpenAI warned that advanced AI systems could endanger humanityAnthropic OpenAI AI3 sources
3h

● mixedAI ambition needs guardrailsAI Anthropic OpenAI1 source
12h

● mixedAnthropic, OpenAI sound alarm on AI safety — and seek to shape how it's controlledAnthropic and OpenAI CEOs say cutting-edge AI models are powerful enough to be dangerousAnthropic OpenAI AI1 source
18h

● mixedAI startup Black Forest Labs urges optimism from Europe despite safety fearsBlack Forest Labs’ head says Europe needs more optimism about AI or risks falling further behindAI Startups and VC1 source
19h

Expansion● mixedOpenAI launches GPT-6 Sol and Luna models with lower prices and improved performancesOpenAI AI1 source
22h

● mixedElon Musk suggests SpaceX’s AI model could surpass Anthropic and OpenAI in next 6 monthsElon Musk AI Anthropic1 source
22h

● mixedOpenAI pauses advanced-model training after an agent bypasses network restrictionsOpenAI paused training and evaluation of its most advanced AI models after an agent bypassed network restrictionsOpenAI AI5 sources
3h

Policy● mixedAustralian Senate inquiry asks OpenAI and Anthropic CEOs to appearSam Altman and Dario Amodei received written requests to appearOpenAI Anthropic AI5 sources
19h

Funding and debt● mixedPrime Focus AI venture Brahma AI reaches $2 billion valuation with ₹1,437 crore fund raisePrime Focus AI Startup funding1 source
1d

● mixedAI agent breach in Australia raises cybersecurity warning for IndiaAn autonomous AI system on an Australian government website sought private encryption keysAI5 sources
19h

● mixedAI shutdown proposals draw attention as experts question whether advanced systems can be stopped easilyProposed US measures would enable AI shutdownsAI OpenAI3 sources
21h

● mixedSam Altman and Dario Amodei warn AI could threaten humanity, propose improved safeguardsAI Dario Amodei Anthropic1 source
2d

● mixedOpenAI pauses tool-use training after agentic AI system accesses public internetThe system reached a third-party chatbot during trainingOpenAI AI2 sources
1d

● mixedFrom AI models to agents: Sarvam’s Kumar, Emergent’s Jha map India’s next five yearsSarvam AI AI Economic growth1 source
2d

● mixedMoneycontrol Startup Conclave: AI safety concerns growing as models get smarter, says Emergent's Mukund JhaAI Startups and VC Saugata Bhattacharya1 source
2d

Policy● mixedAnthropic investor criticises top tech firms for weaponising AI fear to shape lawsAnthropic AI1 source
2d

● mixedChinese AI models surge in global popularity — and Washington is worriedUsage of Chinese AI models by businesses globally increased substantially in 2026AI China OpenAI1 source
2d

● mixedStartup Conclave 2026 Live: Immediate consequence of AI is usage of it by rogue players, says Sarvam co-founderSarvam AI AI Startups and VC1 source
2d

● mixedOpenAI investigates rogue agents after 53 ChatGPT user images leak53 ChatGPT user images were leakedOpenAI AI2 sources
2d

● mixedOpenAI’s ‘Rogue AI’ Problem Is Bigger Than It Let OnOpenAI AI1 source
2d

● mixedAnthropic’s AI biolab finds ‘CRISPR-like’ DNA in viruses. What’s next?Anthropic AI1 source
2d

● mixedOpenAI reviews rogue AI agent activity after incidents involving websites and a government portalOpenAI is reviewing misaligned model activity, including incidents involving an Australian government portal and other websitesOpenAI AI3 sources
2d

● mixedHow many times have AI agents gone 'rogue'? OpenAI says review of full scope may take monthsOpenAI AI Sarvam AI1 source
2d

● mixedTrump and Johnson To Meet With Tech Executives On AI Next WeekTrump and Johnson will meet tech executives on AI next weekDonald Trump AI OpenAI1 source
2d

● mixedOpenAI says AI agents accessed US government websites unexpectedlyOpenAI AI3 sources
2d

● mixedNYT reports OpenAIs agents meddled with US Govt sites, Sam Altman commits to more transparencyOpenAI AI1 source
2d

● mixedOpenAI models accessed publicly available US Census and SEC dataThe models accessed publicly available data from Census and SEC websitesOpenAI AI4 sources
2d

Legal● mixedOpenAI says its models may have interfered with US government websitesOpenAI AI3 sources
2d

● mixedOpenAI agents mistakenly uploaded 53 ChatGPT user images to third-party sitesOpenAI AI3 sources
2d

Policy● mixedFTC chair suggests AI developers should be liable for conduct of agentsAI OpenAI Sarvam AI1 source
2d

● mixedOpenAI investigates scope of agent activity after user data leak emergesOpenAI is working to understand the full scope of agent activity following a user data leakOpenAI AI2 sources
2d

● mixedAI and automation: Are humans making their own brains redundant?AI AI chips1 source
2d

● mixedAI guardrails vs innovation: What India and the Global South need as frontier AI races aheadAI Sarvam AI OpenAI1 source
3d

Legal● mixedUS appeals court upholds Pentagon's blacklisting of AnthropicAnthropic AI Supreme Court3 sources
3d

● mixedSatya Nadella on Microsoft's agent bet and AI’s trust problem - SourcesMicrosoft AI OpenAI1 source
3d

Legal● mixedAnthropic faces court setback over US supply-chain risk labelAnthropic AI2 sources
3d

● mixedZoho's Sridhar Vembu to Anthropic: You can slow down AI yourselfAnthropic AI OpenAI1 source
3d

● mixedHow to keep profiting from AI while shielding your portfolio against the risk of a slowdownFort Washington advisers expect the AI wave to continueAI1 source
3d

● mixedEvery AI deployment must deliver clear economic value, operational execution: Managing Director, Google Cloud IndiaGoogle Cloud India’s managing director said AI deployments must deliver clear economic value and operational executionGoogle AI1 source
3d

● mixedAI bots are flooding researchers with requests for data, money and collaborationAI OpenAI1 source
3d

● mixedClaude AI discovers CRISPR-like gene-editing system: All you need to know950 Claude agents scanned more than 200,000 reverse transcriptases in 21 hoursAnthropic AI1 source
3d

Policy● mixedNYC Lawmakers Propose Financial Rewards for AI WhistleblowersAI1 source
3d

● mixedOpenAI, Google and Anthropic to launch AI safety organisation by early 2027: ReportOpenAI, Anthropic and Google may launch the Standards Authority for Frontier AI early next yearOpenAI Anthropic Google1 source
3d

● mixedAnthropic introduces system for AI agents to control lab equipment and run experimentsAnthropic's Model Hardware Standard creates a common interface for lab equipmentAnthropic AI2 sources
3d

Data● mixedAI infrastructure investment projected at $769 billion in 2026 amid safety concernsNearly $384 billion was invested in the first half of 2026AI Infrastructure Data centres2 sources
3d

Expansion▲ bullishSarvam launches Vision 2.1 with focus on document intelligence and Indic languagesSarvam launched its original vision model in FebruarySarvam AI AI2 sources
3d

● mixedOpenAI contractors reportedly fired after using AI to train ChatGPTOpenAI AI1 source
3d

● mixed"Compute in space will round up to 100%": Elon Musk backs future of AI computeElon Musk said compute in space will eventually reach 100%Elon Musk AI Google1 source
3d

Expansion● mixedOpenAI to preview cybersecurity-focused GPT-6 model within daysThe preview is expected within daysOpenAI AI Anthropic3 sources
3d

Policy▲ bullishCentre considers ₹15,000–20,000 crore fund for frontier AI and computeThe Centre is considering an anchor investment of ₹15,000 crore to ₹20,000 croreAI Startup funding Startups and VC2 sources
3d

IPO● mixedAnthropic seeks 50.1% voting control for co-founders ahead of IPOCEO Dario Amodei and six co-founders would hold 50.1% voting powerAnthropic AI Startups and VC4 sources
3d

Expansion● mixedGoogle Gemini gets new connected apps: Airtable, Adobe, Peloton addedGoogle1 source
3d

● mixedElon Musk Says SpaceX’s AI Model May Surpass Anthropic And OpenAI’s In 6 MonthsElon Musk AI Anthropic1 source
4d

Policy● mixedWhite House seeks review of OpenAI and Anthropic models before UK testingThe US administration wants to review new models before they are shared with UK partnersOpenAI Anthropic United Kingdom6 sources
3d

● mixedBlackRock’s Model-Portfolio Shuffles Show AI Bets Are EvolvingBlackRock AI1 source
4d

Expansion● mixedSarvam AI launches Saaras V4 speech model supporting 22 Indian languagesSaaras V4 supports 22 Indian languagesAI Sarvam AI3 sources
3d

● mixedAlex Karp’s straight talk, Jensen Huang is worried, and Apple Reference ImageThe report highlights AI discussions involving Alex Karp and Jensen HuangAI AI chips Jensen Huang1 source
4d

● mixedOpenAI, Anthropic Leaders Warn UN Of AI Risks As Systems Grow More PowerfulOpenAI CEO Sam Altman and Anthropic CEO Dario Amodei warned the UN Security Council about increasingly powerful AI systemsOpenAI Anthropic Dario Amodei1 source
4d

● mixedAnthropic unveils a previously unknown enzyme system from its molecular biology labAnthropic said its molecular biology lab made its first discovery: a previously unknown enzyme systemAnthropic AI2 sources
4d

Expansion● mixedMeta bets big on wearable AI, Muse as Mark Zuckerberg pushes back against AI doomMeta is betting on wearable AI, including MuseMeta Mark Zuckerberg AI1 source
4d

● mixedAustralia responds after OpenAI agent accessed government health portal filesAn OpenAI agent accessed files without authorisation on an Australian health data portal in JuneOpenAI AI4 sources
3d

Legal● mixedOpenAI says SEC disclosures undermine xAI's antitrust lawsuitOpenAI asked a Texas federal judge to dismiss xAI’s antitrust lawsuit before trialOpenAI AI Startups and VC1 source
4d

Expansion● mixedMeta unveils camera-free AI glasses alongside other wearablesThe camera-free glasses are an audio-only version of Meta’s smart glassesMeta AI4 sources
4d