Plus: Claude Duet goes to work, Nvidia ships data centers by truck, and one economist says AI's payoff is zero
Sponsored by
Bring your creative ideas to life with OpenArt. Generate stunning AI images, videos, and animations, or explore powerful creative tools to turn your imagination into reality.
Start creating with OpenArtToday, we've got OpenAI closing in on a $1.25 trillion valuation, a Chinese open-weights model knocking on the frontier's door, Anthropic shipping an agent that reads your inbox before you do, and Nvidia literally trucking data centers to your parking lot.
One year ago this week, in 2025…
Microsoft and OpenAI signed the deal that redrew the map. OpenAI recapitalized into a public benefit corporation, Microsoft's stake landed at roughly 27% (about $135 billion), OpenAI committed to another $250 billion of Azure spend, and Microsoft's model IP rights were extended through 2032. Twelve months later, OpenAI is raising at $1.25 trillion and has confidentially filed to go public.
💰 OpenAI Nears $30B Round at a $1.25 Trillion Valuation
OpenAI is raising up to $30 billion led by UAE investors alongside Coatue, and told investors it wants signatures within the week. It's being framed as an "opportunistic" bridge ahead of an IPO that CFO Sarah Friar has signalled for 2027 — with a side deal for more UAE compute attached.
🏆 Kimi K3.1 Pro Is About to Take the Open-Weights Crown
Moonshot's new 3T-class model went live on October 23 with a 1M-token context window, and Artificial Analysis scored it at 55 — ahead of GPT-6 Astra and Gemini 4 Argon, behind only Claude Opus 5.5. The weights drop under an open license on November 4, which would make it the strongest openly available model on the planet. At $2/$10 per million tokens, it's half the price of Opus.
🤝 Anthropic Launches Claude Duet
Built on Opus 5.5 and shipping as a research preview on iOS, Android and the web, Duet sends a morning brief, watches your connected apps through the day, and runs tasks on its own cloud VM with a built-in browser. It taps 6,000+ connectors — Gmail, Slack, QuickBooks, Docusign, Shopify — and asks permission before it sends, pays, signs or deletes anything. Aimed squarely at Meta's Muse and OpenAI's Dots.
📦 Nvidia Now Ships Data Centers on a Truck
At GTC Berlin, Jensen Huang unveiled DSX POD: seven Vera Rubin racks, cooling, power and networking pre-assembled into a 28,000kg shipping container. Nvidia claims it cuts time-to-first-token from six months to 14 days. Deliveries start in early 2027 via Foxconn, Lenovo, HPE and others — and a second reference architecture targets OpenShell, Nvidia's sandboxed agent toolkit.
⚡ single-agent-loop-kit is Anthropic's migration kit for collapsing your subagent swarm back into one loop — published alongside research showing a single agent loading context on demand ran roughly twice as fast as the multi-agent version.
🔬 AI-Scientist-v3 is Sakana AI's third generation "Agentic Scientist," orchestrating coding agents end to end for automated scientific co-discovery.
Coming in Hot
AI Song Generator, AI Lyrics Generator and audio studio
Accurate AI content detector and plagiarism checker
Vibe code a mobile app in minutes.
Production AI gateway for multi-LLM apps
Turn still photos into lifelike animated videos
Unified LLM API gateway for 100+ AI models
Managed OpenClaw hosting with one-click AI agent deployment
AI presentation maker with deep research and citations
Data-driven SEO content optimization platform
AI search engine that builds interactive apps for you
Submit your AI tool to the #1 AI tools marketplace.
What it does: Lacuna is an AI song generator that turns text prompts and lyrics into original songs with AI vocals, melodies, and instrumentals. It helps creators make music for videos, podcasts, social media, and commercial projects.
Why It's Worth Your Time
Pricing: Lacuna offers a free plan with 120 monthly credits. Paid plans start at $9.50/month (billed annually) for Basic, $24.50/month for Pro, and $49.50/month for Ultra. Annual billing saves 50% compared with monthly billing.
Why it matters: Lacuna helps musicians, content creators, and independent artists turn ideas into finished music without a recording studio. Its editing and mastering tools make it easier to refine AI-generated songs for creative and commercial projects.
What it does: OpenArt is an AI creative platform that lets users generate images and videos, edit visuals, and create stylized content using AI models. Its Suite provides tools for tasks such as image generation, image editing, character creation, and video production, helping creators turn ideas into visual content.
Why It's Worth Your Time
Pricing: OpenArt offers four paid plans, billed monthly or annually, with annual discounts of up to 27%. Starter costs $14/month with 4,000 credits, Plus costs $34/month with 12,000 credits, Pro costs $56/month with 24,000 credits, and Wonder costs $240/month with 106,000 credits. Discounted annual rates are $13, $27, $44, and $175 per month, respectively. New users can try 40 credits during a seven-day trial.
Why it matters: OpenArt brings AI image generation, video creation, editing, and character design into one workspace. Instead of switching between separate tools, creators can develop visuals, produce videos, and refine content in one place. It's useful for marketers, designers, content creators, and filmmakers who want to produce a wider range of AI-generated content.
An Apollo economist argues that despite enormous spend and louder hype, AI has delivered no measurable financial return to corporate America outside the tech sector. Hold that next to a $1.25 trillion funding round and you have this week's entire mood.
🧪 OpenAI Quietly Buys a Biotech Startup
OpenAI acquired four-year-old Orchestra Bio, whose agents handle drug discovery operations, and is folding the team into Rosalind — its life sciences model already used by Moderna and Novo Nordisk. A reminder that the labs can absorb your category as easily as they can compete with it.
🗣️ Block Built a Voice Model Trained Only on Speech
Dessa is full-duplex — it listens and talks at once, backchannels, and reacts to laughter and sarcasm — and it was trained from scratch on human speech with no text and no pretrained base. It runs locally on an M4 Max, faster than real time. Currently waitlist only.
