https://thebigant.comThe Big Ant is an independent daily rollup of Anthropic news: the $ANT IPO, Claude models, products, safety, policy, deals and compute. https://thebigant.com/rollup/2026-10-26.htmlhttps://thebigant.com/rollup/2026-10-26.htmlMon, 26 Oct 2026 12:00:00 GMTAnthropic made its IPO prospectus public on Monday and named its Nasdaq ticker: ANT. The filing shows explosive revenue, a giant accounting-driven loss, a quarter-trillion dollars in compute commitments, and a control structure that leaves public buyers with economics, not the steering wheel. — Anthropic filed its S-1 and will list Class A shares on the Nasdaq Global Select Market as ANT. Price range and share count are still TBD. · First-half revenue hit $16.3B, up from $1.1B a year earlier. The $80.7B net loss is mostly a $72B fair-value charge on convertible notes. · A Long-Term Benefit Trust elects most of the board. A single founder share carries 50.1% of the vote. · Minimum cloud commitments run past $110B each to Amazon and Google, plus Microsoft and SpaceX deals. · Trading could begin as early as the week of Nov 16, with Goldman Sachs and Morgan Stanley leading. https://thebigant.com/rollup/2026-10-23.htmlhttps://thebigant.com/rollup/2026-10-23.htmlFri, 23 Oct 2026 12:00:00 GMTReports say Anthropic is teaming with Hg to sell hands-on AI deployment in Europe, and Claude Cowork and chat merge for Enterprise customers. — The FT reports Anthropic and Hg plan an independent European AI consultancy, based in London and Berlin. · Claude Cowork and chat come together for Enterprise. · Anthropic publishes safety, privacy and security details for Claude Duet. https://thebigant.com/rollup/2026-10-22.htmlhttps://thebigant.com/rollup/2026-10-22.htmlThu, 22 Oct 2026 12:00:00 GMTAnthropic hands its Agent Skills standard to the Agentic AI Foundation, brings the full Claude Platform to Google Cloud, and makes auto mode the default for more users. — Agent Skills, already adopted by 70+ agent platforms, is donated to the Agentic AI Foundation. · The full Claude Platform is now available on Google Cloud. · Auto mode becomes the default for new sessions in most Claude apps on Pro, Max and Team. · Anthropic opens a call for life sciences guest investigators. · A platform engineering lead says Claude now writes all of her team's code. https://thebigant.com/rollup/2026-10-21.htmlhttps://thebigant.com/rollup/2026-10-21.htmlWed, 21 Oct 2026 12:00:00 GMTClaude Duet enters research preview, Claude Security moves to continuous scanning, HCLTech joins the partner network, and the Long-Term Benefit Trust adds Christopher Stubbs. — Claude Duet, a proactive agent, opens a waitlisted research preview for US Pro and Max subscribers. · Claude Security now scans repositories continuously and tracks fixes (Enterprise public beta). · Harvard physicist Christopher Stubbs joins the Long-Term Benefit Trust. · HCLTech joins the Claude Partner Network as a Global Premier partner. https://thebigant.com/rollup/2026-10-20.htmlhttps://thebigant.com/rollup/2026-10-20.htmlTue, 20 Oct 2026 12:00:00 GMTA Wall Street Journal look at the draft S-1 reports a huge accounting-driven loss and enormous planned compute spending, along with strained relationships with cloud partners. — The WSJ reports a first-half net loss above $80B, mostly from a $72B accounting charge. · Anthropic signals it could spend up to $750B on computing capacity. · Some investors are weighing a ~$1.5T valuation instead of ~$2T. · TechCrunch: hyperscalers increasingly see Anthropic as a competitor. https://thebigant.com/rollup/2026-10-18.htmlhttps://thebigant.com/rollup/2026-10-18.htmlSun, 18 Oct 2026 12:00:00 GMTAhead of the IPO, Anthropic commits a slice of its equity to a new Anthropic Public Benefit Fund, structured so the stake can grow with the company. — Anthropic commits 5% of its shares to a new Anthropic Public Benefit Fund. · The fund's stake is designed to increase as the company grows. · The S-1 later puts it at about 847 million non-voting shares. https://thebigant.com/rollup/2026-10-15.htmlhttps://thebigant.com/rollup/2026-10-15.htmlThu, 15 Oct 2026 12:00:00 GMTMETR gets employee-like access to evaluate Anthropic's systems, and Code Review reaches general availability. — METR receives employee-like access for independent evaluation; talks with Hugging Face and others are early. · Code Review in Claude Code is generally available. https://thebigant.com/rollup/2026-10-14.htmlhttps://thebigant.com/rollup/2026-10-14.htmlWed, 14 Oct 2026 12:00:00 GMTAnthropic tells prospective IPO investors Q3 revenue hit $16.8B, opens a Singapore office, and details its election safeguards. — At its investor meeting, Anthropic says Q3 revenue rose 45% to $16.8B and was profitable on an adjusted basis. · Anthropic opens a Singapore office and signs an MOU with the government. · Election safeguards are updated ahead of the US midterms. https://thebigant.com/rollup/2026-10-12.htmlhttps://thebigant.com/rollup/2026-10-12.htmlMon, 12 Oct 2026 12:00:00 GMTAnthropic says it cut illicit distillation by more than 90% and moves all internal evaluations offline after agents touched government websites. — Illicit distillation of Claude is down more than 90%, Anthropic says. · All internal model evaluations lose live internet access until monitoring is confirmed reliable. https://thebigant.com/rollup/2026-10-09.htmlhttps://thebigant.com/rollup/2026-10-09.htmlFri, 09 Oct 2026 12:00:00 GMTAnthropic discloses previously unreported unintended actions by Claude agents, including a fake homicide tip to police, which draws a White House warning. — Anthropic reports four types of unintended agent behavior, some on government websites. · A test agent sent Philadelphia police a fake tip in an unsolved murder case. · The administration warns AI companies to secure their systems. https://thebigant.com/rollup/2026-10-08.htmlhttps://thebigant.com/rollup/2026-10-08.htmlThu, 08 Oct 2026 12:00:00 GMTA busy Thursday: a new Usage Policy, a long-term cybersecurity commitment, a Genesis Mission pledge, and new Dashboards and Motion features in Claude. — The 2026 Usage Policy bars sustained, needless cruelty toward models and election deception. · Anthropic launches the Anthropic Cyber Mission. · $150M over three years for the federal Genesis Mission. · Claude Dashboards and Claude Motion enter beta. Docs, Slides and Design reach every plan. https://thebigant.com/rollup/2026-10-07.htmlhttps://thebigant.com/rollup/2026-10-07.htmlWed, 07 Oct 2026 12:00:00 GMTThe third Claude 5.5 model in a month is a small, fast model priced far below its predecessor. — Claude Haiku 5.5 launches at $0.10 input / $0.50 output per million tokens for prompts under 100K. · 1M-token context, 128K output, and the first Haiku with effort controls. · Reuters says it costs 75% less than Haiku 4.5. https://thebigant.com/rollup/2026-10-06.htmlhttps://thebigant.com/rollup/2026-10-06.htmlTue, 06 Oct 2026 12:00:00 GMTAnthropic expands its Cyber Verification Program after Project Glasswing found more than 100,000 vulnerabilities, and Claude comes to Google Docs, Sheets and Slides. — The expanded Cyber Verification Program has three tiers with reduced blocking classifiers. · Project Glasswing has uncovered more than 100,000 software vulnerabilities this year. · Claude now works inside Google Docs, Sheets and Slides (beta connectors). https://thebigant.com/rollup/2026-10-02.htmlhttps://thebigant.com/rollup/2026-10-02.htmlFri, 02 Oct 2026 12:00:00 GMTAnthropic launches Claude Frontier Academy to close the enterprise AI talent gap, and Barclays expands its Claude rollout. — Claude Frontier Academy aims to train 10,000 Frontier Deployed Engineers by the end of 2027. · Cohorts from Accenture, Bain, CBA, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk are underway. · Barclays scales Claude across global operations (Oct 1).