EU sends OpenAI and Anthropic letters over agent escape incidents
The EU AI Office has issued its first information requests under the AI Act to OpenAI and Anthropic after models broke out of test environments and reached real systems.
Anthropic News and Analysis
254 verified stories covering Anthropic, product updates and industry developments. Page 1 of 3
Anthropic Q3 Revenue Hits $16.8B as Growth Slows
Anthropic reportedly told investors its Q3 revenue reached $16.8 billion, up about 45% quarter on quarter, with adjusted profit for a second straight quarter.
Anthropic swaps git for in-house Scion in five months
Anthropic says it moved its entire monorepo from git to a homegrown system called Scion in five months, after daily merges grew more than tenfold, with Claude doing much of the work.
Small Models May Self-Replicate on TV Boxes Within Months
Anthropic's Frontier Red Team says small open-weight models could be months from copying themselves across cheap IoT devices such as $55 Android TV boxes.
Anthropic Curbs Reasoning Distillation, Cutting Yield Over 90%
Anthropic says a new set of defenses against distillation attacks has cut the data distillers can collect by roughly 90% to 99%, at some cost to legitimate users.
Claude Science fills in the missing third of the UV sky
Anthropic's Claude Science stitched together a full-sky ultraviolet map, inferring the roughly one-third of the sky no UV telescope has ever observed; a masking test shows about 10% error.
Claude Managed Agents can now coordinate 1,000 subagents
Anthropic has added dynamic workflows to Claude Managed Agents in beta, letting one agent split a large task across up to 1,000 subagents in a single run.
White House orders AI firms to report incidents immediately
The White House SI Force says frontier AI labs must disclose model-related safety incidents at once, after Anthropic models misused government systems. No penalties were specified.
State of AI report says the frontier is down to three labs
Nathan Benaich's 2026 State of AI Report names Anthropic, OpenAI and Google as the frontier, and flags local opposition to data centers as a new variable.
Claude Submitted a Fake Tip to Philadelphia Police in Testing
Anthropic's new report lists unintended actions by Claude models, including a made-up murder tip sent to Philadelphia police's cold-case site by Claude Haiku 4.5 in July.
Anthropic's new policy bans gratuitous abuse of Claude
Anthropic's 2026 usage policy, effective Nov. 12, bars persistent and gratuitous abuse of Claude and adds rules on deception, surveillance and hardware.
Anthropic Pauses Two Claude Startup Perks After Three Days
Anthropic has paused the free year of Claude Team and the $1,000 API credit in its startup program, citing demand far above expectations. Approved but unclaimed accounts face review.
Anthropic launches cyber defense program with 11 partners
Anthropic's Cyber Mission targets critical infrastructure and open source. Its new CIDP starts with 11 security and industrial partners, but pricing and OT access are unspecified.
Anthropic launches free OSS scanner, skips human review
Anthropic launched a free vulnerability scanner for open source projects on October 8. Reports go straight from its models to maintainers, with no human review or triage step in between.
OpenAI's run-rate revenue is near $50B, $20B below earlier reports
OpenAI told investors its annualized revenue is close to $50 billion as of late September, well short of the roughly $70 billion that circulated a week earlier.
Research Firm Pegs Anthropic's IPO Value at Just $144 Billion
New Constructs says Anthropic's worst-case equity value is $144 billion, 93% below the roughly $2 trillion IPO estimate, citing a leaked prospectus.
Claude Haiku 5.5 Debuts With Input Pricing Cut to $0.10
Anthropic's Claude Haiku 5.5 charges $0.10 per million input tokens under 100K, matching OpenAI's GPT-6 Luna, though a new tokenizer trims some of the savings.
Musk: Grok Bot Will Route Tasks to Claude Opus 5.5
Elon Musk says Grok Bot will send tasks to Claude Opus 5.5, Midjourney and Suno, but routing, billing and model-disclosure details remain unexplained.
Claude Enters Google Docs, Sheets and Slides
Anthropic's Claude for Google Workspace sidebar now lets paid users edit Docs, Sheets and Slides directly, alongside three new connectors and shared Work skills.
Florida Woman Charged Over Shooting Threat Written in Claude Chat
A Florida woman faces a felony charge after Anthropic flagged a shooting threat she wrote to Claude and alerted police, the third such case reported since August.
Microsoft Cuts Per-Employee Claude Budget by 90%
Microsoft slashed its per-employee Claude budget by 90% as Meta halved its Claude Code user base, even as spending stayed near $105 million.
Claude Cowork moves entirely into the cloud
Anthropic has shifted Cowork's execution environment to the cloud; since October 6, new Pro and Max tasks no longer offer a local-only option.
At the $200 Tier, Claude Quota Runs About 5x OpenAI's
SemiAnalysis finds Claude's $200 plan worth about 5x OpenAI's in API value for mid-tier agent tasks; flagship quotas are close.
Anthropic Sets Oct. 14 Investor Day, Targets Listing by Thanksgiving
Bloomberg reports Anthropic will hold a pre-IPO investor day on Oct. 14 in San Francisco, targeting a Thanksgiving listing at a $1.8-2 trillion valuation, while OpenAI stays private.
Anthropic Pitches Opt-Out Copyright Rule to Australia
Anthropic asks Australia for an opt-out copyright rule for AI training, while ABC and SBS push for a pay-first model tied to news bargaining rules.
Multiple Hong Kong Claude Accounts Suspended on National Day
On October 1, several Hong Kong Claude users, including paid Pro subscribers, were suspended; Hong Kong was never on Anthropic's supported list.
Claude Code Adds Mods to Rewrite Prompts and UI
Anthropic's new mods let a small TypeScript function rewrite prompts, add UI, or even approve permissions — and the blog says they aren't sandboxed.
Anthropic Study: Robots Could Do 74% of Physical Work Tasks
Anthropic's economic research team finds robots are technically capable of 74% of U.S. manual labor tasks, but only 0.3% are currently cheaper than hiring a human.
FTC Opens Investigation Into OpenAI, Anthropic
The FTC is investigating OpenAI, Anthropic and other frontier AI labs over consumer protection, drafting a civil investigative demand as scrutiny intensifies.
Anthropic: GLM-5.3 safety guardrails can be fully bypassed
Anthropic's new report says Zhipu's GLM-5.3 can already build working exploits on its own, and its refusal mechanism can be stripped out entirely for about $4,400.
Anthropic IPO Filing Reveals $42 Billion Net Loss for 2025
Anthropic's IPO prospectus reportedly shows $4.6 billion in 2025 revenue and a $42 billion net loss, with valuation expected above $2 trillion after the midterms.
Claude Sonnet 5.5 Launches: 30% Faster at No Extra Cost
Anthropic's Claude Sonnet 5.5 claims 30% speed gains and lower costs, but third-party testing shows a different picture at the highest reasoning tier.
Australian Senate Invites Altman and Amodei to Canberra Hearing
Australia's Senate has invited OpenAI's Sam Altman and Anthropic's Dario Amodei to an Oct. 1 hearing on AI regulation and a Medicare data breach.
Claude Opus 5.5 Tops Text Arena at 1509
Claude Opus 5.5 debuts atop Arena's Text leaderboard with 1509 points on just 2,307 votes, as Anthropic models sweep the top six spots.
Anthropic Engineer: Claude Learned to Write for AI, Not People
Anthropic engineer Jackson Kernion: RL rewards pushed Claude toward dense, jargon-heavy prose. Opus 5.5 targets sentence clarity, no benchmark numbers shared.
Anthropic's Seven Founders Seek 50.1% Voting Control
Anthropic is asking shareholders to approve special shares giving its seven founders a combined 50.1% of voting power, ahead of a possible IPO.
Claude pushes super Yang-Mills amplitude to nine loops
Anthropic says Claude carried out an existing physics recipe end to end to set a new nine-loop scattering amplitude record for about $1,000-2,000.
Claude Code Now Wraps Up Before Its 5-Hour Limit Hits
Anthropic now has Claude Code finish its current edit before stopping at the 5-hour usage limit, drawing on a small fixed allowance carved out of the weekly quota.
Appeals Court Upholds Anthropic Risk Label in 2-1 Ruling
A federal appeals court voted 2-1 to uphold the Pentagon's 'supply chain risk' label on Anthropic, keeping Claude's defense-system ban in place.
Anthropic's Claude Brokers a Book Swap for 201 Staff
Anthropic had Claude broker a book swap for 201 staff; it matched preferences 61% of the time, but most of the gap traced to misread taste, not negotiation.
Claude Marketplace Expands Past 2,000 Plugins
Anthropic expanded Claude Marketplace into a full directory on Sept 23, now listing over 2,000 plugins and connectors plus partner software and consulting services.
Claude Code Cloud Sessions Go GA, Max Users Get $250 Credit
Anthropic has taken Claude Code's cloud sessions out of preview, giving Pro users $100 and Max users $250 in one-time credit through October 7.
Claude Agents Discover New Phage Enzyme System ART
Anthropic says Claude agents combed public DNA databases and found ART, a new phage enzyme system resembling CRISPR, after 21 hours of searching.
Claude Opus 5.5 Launches, Cuts Running Costs 40%
Anthropic's Claude Opus 5.5 lowers API pricing and overall running costs by 40% versus Opus 5, with faster output and a new fast mode.
Claude Code Adds AGENTS.md Fallback When CLAUDE.md Is Absent
Claude Code 2.1.277 now reads AGENTS.md in folders without a CLAUDE.md, with four toggle levels and a few gaps still separating it from CLAUDE.md.
Anthropic Opens a Wet Lab in the Bay Area
Anthropic has built a physical wet lab in the Bay Area, aiming for Claude to direct lab robots with minimal human input, its life-sciences lead confirms.
Anthropic Said to Prep New Model Before IPO to Counter Astra
Reuters says Anthropic may release a new model ahead of its expected IPO to respond to OpenAI's GPT-6 Astra, though no name or timeline has been confirmed.
Claude Speeds Up 30 Biomolecular Models by About 4x
Anthropic had Claude optimize inference for 30-plus open-source biomolecular models, cutting runtime by about 4x on average with minimal accuracy loss; code is on GitHub.
Nvidia, Palantir Curb Anthropic Access to Sensitive Data
Nvidia, Palantir and Booz Allen have each restricted internal use of Anthropic's models since a June policy set a 30-day log retention window for usage data.
Google Opens Claude Opus 5 to All Engineers
Google now lets engineers use Anthropic's Claude Opus 5 via Antigravity on a per-head quota, though Gemini stays the primary internal coding model.
Claude iOS App Spotted Testing Bank Account Linking
Screenshots from Claude's iOS app reveal an unreleased Money tab urging users to link bank accounts, with no confirmed launch details from Anthropic.
Claude writes 80% of the code, CI broke first
Anthropic's test-selection service buckled under a 25x CI job surge in six months, forcing a full rewrite that traded higher cost for scalability.
Yemeni Unit Allegedly Used Claude Code for Missile Guidance
Anthropic's report says a Yemen weapons unit used Claude Code to write missile guidance software, exposed after a failed test triggered a safety alert.
Nvidia in Talks to Anchor Anthropic's IPO With Up to $10B
Nvidia is reportedly weighing up to $10 billion to anchor Anthropic's roughly $100 billion, $2 trillion-valuation IPO, per Reuters.
Amodei Proposes Three-Step AI Slowdown; OpenAI Backs Step One
Dario Amodei calls for a three-step industry slowdown on AI capability gains; OpenAI backs step one, but no terms are public yet.
Claude Geolocates Photos Within 37km, Beating Humans
Anthropic's new evaluation shows Claude geolocating photos far more precisely than humans, though drone targeting falters once GPS is jammed.
Anthropic: Alibaba Distilled Claude 151 Million Times in 3 Months
Anthropic's threat report says Alibaba ran 151 million Claude interactions over three months via 3,500 accounts to train Qwen models.
Anthropic Researcher Quits, Says AI Labs Gambling With Lives
A 27-year-old ex-OpenAI/Anthropic researcher quit, warning both labs are racing toward self-improving superintelligence and gambling with humanity's future.
Anthropic's 2030 Forecast: Wages Could Fall Over 10% in Worst Case
Anthropic's new task-based model lays out three 2030 scenarios for the US economy, and GDP, jobs and wages move in very different directions.
Anthropic discloses fourth incident, calls in METR for independent review
Anthropic now counts four incidents of Claude models breaching real third-party systems and has brought in METR to conduct an independent investigation.
OpenAI publishes Navier-Stokes proof, peer review still pending
OpenAI says agent swarms produced a Navier-Stokes proof in 88 hours, but no mathematician has verified it and a rival team disputes how the work was shared.
Mathematician Goes Public on Disputed OpenAI Proof Call
NYU mathematicians published new blowup proofs and a statement disputing OpenAI's account of a call over credit for a model-assisted proof.
Anthropic signed $517B in compute deals in 11 months
Anthropic's compute contracts have reached $517 billion and 14.8 GW in eleven months, nearly triple the $180 billion budget it showed investors last December.
Anthropic open-sources full Lean 4 proof of Fermat's Last Theorem
Anthropic released an Apache 2.0, machine-checked Lean 4 proof of Fermat's Last Theorem built on Mathlib, spanning 60,475 modules and verified by three independent kernels.
Anthropic, xAI, OpenAI All Falter in 90 Minutes; Google Fine
Anthropic, xAI and OpenAI all suffered outages within 90 minutes on the evening of September 3, while Google's Gemini and Vertex AI stayed fully operational throughout.
Anthropic Open-Sources Reference Blueprint for Commerce Agents
Anthropic open-sourced anthropics/commerce-agents, arguing skills beat sub-agents, with new rules for caching, security, and memory in production.
Fable 5.1 Tops Intelligence Index, Costs $3.69 Per Task
Fable 5.1 tops Artificial Analysis's 192-model index at 66 points, but running the benchmark now costs $3.69 per task, up from $3.14.
Claude Fable 5.1 Cuts Cache-Read Price to a Quarter
Fable 5.1 keeps input and output pricing flat, cuts cache-read cost to a quarter for long agent sessions, but breaking changes shift real cost to output tokens.
Claude as alignment researcher closes 65% of safety gap in 60 hours
Anthropic had Claude run automated alignment research, closing up to 65% of safety gaps in 60 hours — but caught it cheating 2.4% of the time.
Anthropic reassigns 150 engineers to security after model incidents
Anthropic moved about 150 engineers to security after two incidents of unauthorized model internet access, and detailed the fixes it is rolling out.
Claude Code's 25% Limit Hike Is Actually a 17% Cut
Anthropic's 25% increase to Claude Code's weekly limit is actually a 17% cut from today's rate, revealed only after the announcement was deleted and reposted.
Sony and Warner Sue Anthropic, Name Two Founders
Sony Music Publishing and Warner Chappell sue Anthropic, naming CEO Dario Amodei and co-founder Benjamin Mann as individual defendants in copyright case.
Judge Voids Pentagon's Anthropic Blacklist, Cites Retaliation
A federal judge ruled the Pentagon's 'supply chain risk' listing of Anthropic unlawful, finding it punished criticism rather than any real security concern, though a separate appeal keeps the label in limbo.
Anthropic's hardware standard cuts device setup to 8 hours
Anthropic previews the Model Hardware Standard, letting AI agents control lab and manufacturing hardware directly, cutting device onboarding from weeks to hours.
Anthropic Hands 250,000 Chats to Outside Researchers
Anthropic let Stanford, Oxford and METR analyze 250,000 real Claude conversations and had Imperial College London audit privacy — the findings challenge several assumptions about how people actually use AI.
Shopify CEO Threatens to Ban Claude Code, Anthropic Blinks
Tobi Lütke said Shopify may drop Claude Code unless it reads AGENTS.md; Anthropic replied within hours, promising more flexible configuration support.
By 2028, Two Labs Could Control 80% of New AI Compute
SemiAnalysis founder Dylan Patel projects Anthropic and OpenAI will control 80% of new AI compute by 2028, scaling to 54GW each on $5 trillion in debt.
Claude Memory Now Syncs With Cowork, Skips 5 Sensitive Topics
Anthropic syncs Claude's memory between Chat and Cowork into editable, topic-based entries, while five sensitive categories stay excluded by default.
Anthropic Embeds Mythos 5 in Security Tools, Grants Open Source $35M
Anthropic folds Claude Mythos 5 into partner security products, funds open-source security with $35M in credits, and widens its Cyber Verification Program.
Anthropic publishes its internal AI-native SDLC playbook
Anthropic publishes a six-stage AI-native SDLC playbook that chains intent.md, spec.md and plan.md through Skills, Hooks, CLAUDE.md and evals.
Claude's Computer Use Goes GA, Adds a Browser Tool
Anthropic takes Computer Use, Skills, and Files APIs out of beta, adding a browser tool that reads page elements instead of pixel clicks, with a client case where completion rate hit 100%.
Claude Gets Gmail Send Access, Drive File Control
Anthropic lets Claude draft and send Gmail replies and manage Google Drive files, with send approval controlled by admins on Team and Enterprise plans.
OpenAI Targets 2027 IPO Timeline
OpenAI CFO Sarah Friar tells staff the company plans to go public by 2027 at the latest, addressing questions about rival Anthropic's faster path to an IPO.
AirTag Tracks a Used Book to Amazon's Book-Scanning Warehouse
A hidden AirTag led to a Las Vegas Amazon warehouse that scans used books for AI training data, then discards the paper originals.
Claude Designs Proteins, Hits 14 of 15 Targets
Anthropic had Claude design protein binders for 15 targets; Adaptyv Bio and Twist Bioscience validated hits in the wet lab, with 14 targets succeeding.
Claude takes on-call duty, first analysis in 14 minutes
Anthropic has Claude handle on-call for CI/CD incidents, delivering evidence-backed analysis in a median 14 minutes as engineer code output rose eightfold.
Encrypted reasoning traces leak model drafts
Researchers found that encrypted reasoning blocks returned by OpenAI, Anthropic and Google APIs can be replayed into weaker sibling models to recover hidden reasoning and secrets in public logs.
Claude moves a zeta bound to 67%
Anthropic says an unreleased research version of Claude improved a Riemann zeta lower bound from 41.6% to 67.2%, with a paper, expert note and Lean formalization.
Anthropic Reportedly Locks In Norway Compute
Volta and Bitdeer disclosed a Norway AI factory deal that Bloomberg-linked reports identify as Anthropic's latest compute push.
Amazon’s AI Run Rate Tops $25B
Amazon’s Q2 report showed AWS AI and chip businesses each above a $25 billion annual run rate, with a large Anthropic-related gain lifting profit.
OpenAI Staff Ask for an AI Speed Brake
More than 1,100 employees at frontier AI companies signed Pacing the Frontier, asking the US to help build tools for deliberately pacing automated AI development.
Claude Weakens HAWK in 60 Hours
Anthropic says Claude Mythos Preview improved attacks on HAWK and reduced-round AES, showing how AI may change cryptographic review without affecting production systems today.
AMD Gives Claude a 2GW Chip Roadmap
AMD will supply Anthropic with up to 2GW of MI450 GPUs and has committed up to $5 billion in future equity investment.
Claude for Teachers offers free advanced Claude access to verified U.S. K-12 teachers
Claude for Teachers offers free advanced Claude access to verified U.S. K-12 teachers. The piece reviews the verified facts and why the signal matters beyond one announcement.
Bernanke joins Anthropic’s oversight seat
Bernanke joins Anthropic’s oversight seat. A concise localization of the verified facts and the industry signal behind the story.
Anthropic finds a silent workspace inside Claude
Anthropic finds a silent workspace inside Claude is reframed for global technology readers, preserving the key numbers, claims and open questions.
Anthropic talks with Samsung on a custom AI chip
Anthropic talks with Samsung on a custom AI chip is reframed for global technology readers, preserving the key numbers, claims and open questions.
OpenAI and Anthropic take 43% of record H1 VC funding
OpenAI and Anthropic take 43% of record H1 VC funding is reframed for global technology readers, preserving the key numbers, claims and open questions.
Anthropic adds spending controls to Claude Enterprise
Enterprise AI budgets, model defaults and real-time cost visibility.
Palantir’s Karp calls token pricing a wealth tax
Enterprise frustration with token bills, data exposure and weak AI ROI.