Databricks hits $188B valuation, extending its run as AI’s favorite second act — TechCrunch Databricks has closed a major funding round at a $188 billion valuation, making it one of the most valuable private companies in the world. The company has fully remade its identity around AI, including publishing research on the cost advantages of open-weight models for coding — a pointed signal about where it sees the next wave of enterprise spend.

💰 Funding & Valuations

Databricks hits $188B valuation, extending its run as AI’s favorite second act — TechCrunch Once known as a data-engineering platform, Databricks has completed a transformation into a vertically integrated AI company. The $188B milestone cements its position as the go-to bet for investors who missed the model-layer wave but want exposure to AI infrastructure and tooling.

Neil Rimer thinks the AI money is coming back out — TechCrunch Index Ventures co-founder Neil Rimer argues the historic wealth concentrating in Silicon Valley’s AI boom will eventually be redistributed — voluntarily or through policy intervention.

Vertu wants executives to pay $6,880 for an AI agent — here’s how it actually performs — TechCrunch The luxury phone brand bundles AI workflows with a premium foldable; reviewers found the hardware compelling but the agent experience underwhelming at the price.

🔐 Security & Privacy

Claude Code 2.1.214 — Claude Code Changelog A concentrated security release patching three distinct permission-check vulnerabilities: a bypass affecting Windows PowerShell 5.1 sessions, a failure to catch file-descriptor redirect forms that bash parses differently than the permission analyzer, and a directory-traversal issue where single-segment dir/** allow rules were incorrectly granting write access to matching subdirectories anywhere in the file tree. Developers running Claude Code should update immediately.

TikTok is testing an AI likeness detection tool — The Verge TikTok is piloting an opt-in scanner for select US creators that detects unauthorized AI-generated likenesses and provides a direct reporting pathway. The move mirrors YouTube’s parallel effort and marks a broader shift toward platform-enforced accountability for synthetic media — a space that has largely been self-regulated until now.

The Zoom hack that says, ‘Don’t record me’ — TechCrunch As AI transcription becomes a default feature in video calling, a new technique lets participants actively block recording — raising questions about consent and ownership of the growing archive of AI-summarized workplace conversations.

🤖 Robotics & Hardware

Agility Robotics plants its flag in Tesla’s backyard — TechCrunch Agility Robotics is opening a Digit humanoid training center in Fremont, California — the same city where Tesla’s factory anchors its own robotics ambitions. The choice of location reads as a deliberate statement of competitive intent in the race to deploy humanoids at industrial scale.

AI-driven memory crunch jolts India’s smartphone market — TechCrunch Surging DRAM demand from AI hardware is squeezing supply for consumer devices, driving up component costs and slowing sales in one of the world’s largest smartphone markets — an early sign that AI infrastructure buildout carries real trade-offs for consumer electronics.

🔬 Research & Models

Controlling Reasoning Effort in LLMs — Ahead of AI Sebastian Raschka examines how large language models learn to operate at low, medium, and high reasoning effort — the training mechanics behind budget tokens and extended thinking modes. Useful background for teams evaluating when deep chain-of-thought is worth the latency and cost versus a faster standard pass.

Generated by claude-sonnet-4-6 on 2026-07-18T10:00:00Z