Squadbase Weekly: Vibe Coding News #11 Anthropic's $30B Raise, Claude Opus/Sonnet 4.6, v0's Major Overhaul, and GPT-5.3-Codex
From the Squadbase Team: Here's our biweekly roundup of the latest news and insights on vibe coding and data-driven development. This issue also covers a major Squadbase update — we're evolving toward an AI-led dashboard experience.
🚀 Squadbase Goes AI-First
https://www.squadbase.dev/en/blog/toward-an-ai-led-dashboard-building-experience
https://www.squadbase.dev/en/blog/how-to-analyse-ga4-data-with-squadbase
Squadbase just shipped a significant update. We’re transitioning from a tool you operate to an AI partner that leads the way. The new experience features automatic database creation from Excel/CSV files, proactive suggestions for what to analyze next, and post-task next-step recommendations — essentially, AI support across your entire workflow. The goal: get you thinking about what your data means, not how to use the product. We also published a step-by-step guide on how to connect GA4 data to Squadbase. With no SQL or coding required, you can have a professional-grade dashboard up and running in about 15 minutes.
💰 Anthropic Closes $30B Series G at a $380B Valuation
Anthropic has completed a $30 billion Series G round led by GIC and Coatue, bringing its post-money valuation to $380 billion. The company is running at a $14B annual revenue run rate — growing more than 10x year over year for the past three years. Perhaps the most telling data point: Claude Code revenue has surpassed $2.5B (doubling since the start of the year), with enterprise use now accounting for the majority of that figure. Eight of the Fortune 10 are now Claude customers. These numbers tell a clear story — vibe coding isn’t just a developer trend anymore. It’s going enterprise. The Squadbase team uses Claude Code daily, and the boost to our product development velocity has been real. It’s also worth noting that Anthropic has explicitly ruled out an ad-based business model, which is a meaningful differentiator from OpenAI.
🤖 Claude Opus 4.6 and Sonnet 4.6 Are Here
https://www.anthropic.com/news/claude-sonnet-4-6
Anthropic released Claude Opus 4.6 on February 5th, followed by Sonnet 4.6 on February 17th. Opus 4.6 brings agentic team capabilities, PowerPoint support, and a 1M-token context window (beta). It also introduces an inference_geo parameter for data residency control, with US-only inference available at a 1.1x price premium. Sonnet 4.6 is priced the same as Sonnet 4.5. For vibe coding tools, improvements to the underlying model translate directly into a better user experience. Bolt.new was quick to integrate Opus 4.6, giving users the ability to choose their reasoning depth (light vs. deep). Our own team has been testing it too — the agentic team feature is qualitatively different from working with a single agent. The ability to break complex tasks into parallel workstreams is a genuine step change.
⚡ Vercel v0 Gets a Major Overhaul — From Prototype to Production
https://vercel.com/blog/introducing-the-new-v0
v0 has crossed 4 million users, and Vercel just gave it a substantial upgrade. The new sandbox-based runtime lets you import any GitHub repository and automatically pulls in your Vercel environment variables and settings. A built-in Git panel now supports branch creation, PRs, and merge deploys — meaning non-engineers can ship production code through a proper Git workflow. The update also adds SSO support, Snowflake and AWS database integrations, and a Notion connector. What this update really signals is that vibe coding is growing up. It’s no longer just a tool for making demos — it’s becoming a full production deployment platform. Vercel’s official blog called vibe coding “the world’s largest shadow IT problem today,” and they’re right to take it seriously. Bringing it into a governed, secure workflow is exactly the direction the industry needs to move — and it aligns closely with Squadbase’s own focus on safe, enterprise-grade app operations.
🧑💻 OpenAI Announces GPT-5.3-Codex
https://openai.com/index/introducing-gpt-5-3-codex/
On February 5th, OpenAI announced GPT-5.3-Codex, its most capable agentic coding model yet. The model combines the coding performance of GPT-5.2-Codex with the reasoning and general abilities of GPT-5.2, achieving top scores on both SWE-Bench Pro and Terminal-Bench 2.0. It’s also 25% faster — and notably, it’s the first OpenAI model to have been used in its own development. Beyond code generation, GPT-5.3-Codex is designed to function as a general-purpose agent capable of executing tasks across a computer. A real-time variant, GPT-5.3-Codex-Spark, is available as a research preview running on Cerebras hardware, delivering ultra-low latency at over 1,000 tokens per second. If reasoning speeds continue to improve at this pace, the interactive feel of vibe coding could change fundamentally.
🎨 Figma Make Adds Connectors
https://www.figma.com/release-notes/
On February 10th, Figma added connector functionality to Make. You can now pull live data from external services directly into your prototypes using @mentions. Two days later, Figma launched local data hosting, letting you store data from Make and other Figma products in a region of your choosing. Today — February 18th — Figma is scheduled to announce its Q4 2025 earnings. The company already crossed a $1B annual revenue run rate in Q3, with AI investments in Make and MCP servers as key growth drivers. This will be the first time Make’s business impact is reflected in financial numbers. Worth watching.
That’s a wrap for this issue. Thanks for reading Squadbase Weekly — see you in two weeks.