RuntimeWire

Real-time startup intelligence for the AI economy. Funding, model launches, infra shifts, and founder moves — the signals that matter.

Latest stories — page 19

• Anthropic submits confidential draft S-1, opening door to IPO

  The AI company says SEC review still has to finish before it decides whether to pursue a public listing.

• Mecka AI raised $60M to train robots on human motion data

  Fortune reports the New York startup closed a $25 million Series A and a $35 million follow-on led by Framework Ventures, while CEO Josh Gao says signed contracts point to a $100 million annual run rate.

• Shafiq Joty's team proposes Procedural Memory Distillation for less forgetful LLM training

  PMD is presented as a training-time method that stores lessons from prior model attempts, then uses them to teach the next policy.

• Invisix Raised EUR20M Seed for Soft X-Ray Chip Metrology, Report Says

  The ASML-alumni team is targeting sub-2nm process control, but the named founders, valuation and chipmaker investor remain undisclosed.

• Legacy Labs Starts A Two-Month Retro-Computing Camp Built Around 2009-Era Sysadmin Work

  The LambdaCreate author is trading one-week constraints for a Windows Server 2008 R2 lab built with modern Linux tooling.

• RuntimeWire launches AI Models hub to track pricing, speed, and real usage

  New AI Models hub at runtimewire.com/models tracks pricing, speed, and live usage across major models, with continuous updates and price history views.

• Tripo's VAST previews Project Eden, a bid for persistent AI worlds

  Rather than treating world models as video generators, VAST says Eden separates durable state from rendering so objects, actions and multiplayer views can persist over time.

• NVIDIA Is Assembling the Operating System for Robots

  The Agent Toolkit collection turns robotics, AV, vision AI and industrial digital twin workflows into agent-executable tasks, with early use from TSMC, Pegatron, Siemens and more.

• Andrej Karpathy joins Anthropic to work on frontier LLMs

  The former OpenAI and Tesla researcher is moving into Claude's pretraining orbit and says he plans to resume his AI education work later.

• Jae Park's Unastella raises $24 million to chase Korean orbital launch

  Altos Ventures led the Series B for the pre-revenue Seoul rocket maker, which says it flew UNA EXPRESS-I from Goheung in May 2025.

• ElevenLabs' reported $500M ARR puts government expansion in focus

  Aligned News says the AI audio company has a Greek government partnership, but the ARR and Greece claims remain uncorroborated by public company materials.

• AgileX Robotics shows research platforms at ICRA 2026, backs WBCD Track 3 at Booth 87

  AgileX Robotics says its hardware targets R&D use cases and is built for Cosmos 3 integrations the community is testing this week.

• NVIDIA unveils Vera, a CPU for agents claiming 80% faster task completion vs x86

  Vera extends NVIDIA's AI platform from GPU to CPU and will ship in standalone servers, Vera Rubin systems, and Vera BlueField-4 STX AI storage platforms, per the company.

• NVIDIA Unveils Nemotron 3 Ultra, Targets the Next Wave of AI Agents

  Jensen Huang used his Taipei keynote to position NVIDIA as a full-stack AI platform, installing Nemotron 3 Ultra atop its open model family with a latent MoE trained in NVFP4 and a claimed 5x throughput uplift.

• Fleet AI's Big Bet: Before Agents Run the Office, Send Them to the Gym

  Insider round led by Sequoia, Bain, Menlo, and SVA; RR climbed from $1M six months ago to $63M now and is tracking $160M next quarter as Ouporov also previews a benchmark exposing spatial failures in frontier models.

• Researchers warn Meta's AI Instagram support can be tricked to email password reset links

  Posts on X and a now-removed Hacker News thread describe a prompt that convinces the AI agent to send reset links to attacker emails without identity checks; posters say takeovers are active and urge users to lock down email and 2FA.

• Pixio AI shares fresh look at its creative tools, signaling ongoing development

  In a post on X, the AI creative platform showed work-in-progress tooling aimed at professionals in visual production workflows.

• YC alum-led Tesana shows a 2-day, 39-prompt AI-built game prototype

  The Los Angeles startup posted a vibe-first prototyping clip on X; it says it used muranyi-3 and roughly $90 in tokens. Engine and funding remain undisclosed.

• MiniMax unveils M3, an open-weights model touting coding-agentic gains and 1M context

  In a thread on X, MiniMax cites 59.0% on SWE-Bench Pro and a new sparse attention scheme that scales context to 1M, with an API promo rolling out this week.

• Agent Cookie Lets AI Agents Stay Logged In Across Your Apps

  Open-source, peer-to-peer cookie and token sync pairs two Macs so agents like OpenClaw and Hermes can stay authenticated for carts, orders, and API calls.

• FeatherlessAI raises $20M Series A to build a neutral layer for open-weight AI models

  AMD Ventures and Airbus Ventures co-led the round; FeatherlessAI says it will hire engineers across the stack to keep customers off any single lab, chip, or cloud.

• Insane Mario-like demo game shows the power of Opus 4.8

  In a 4-post thread, Nakajima shared a demo labeled Opus 4.8 and said he will run an AI security session for non-engineers, calling out Claude Code users.

• Spanhove hacks Meta Ray-Ban Display for a Strava-powered ghost race while you run

  The indie builder layered Strava GPX, pace coins, and sprint zones into a browser app that runs directly on the glasses, saying it works without a flagship phone.

• Artisan CEO Jaspar Carmichael-Jack says KC Green dispute is settled, pulls Ava ads using 'This is fine'

  Artisan took down New York and San Francisco ads that riffed on KC Green’s meme; Green says the settlement came together quickly, per TechCrunch, with terms undisclosed.