business · Reuters
Crusoe hits $30.9B valuation as AI infrastructure race intensifies
Confirmed: AI infrastructure provider Crusoe raised $3.9 billion in Series F funding at a $30.9 billion post-money valuation. Crusoe says the money will support expansion of large-scale AI campuses and its modular Spark data-centre units.
Why it matters: The round shows how much capital is flowing into the physical infrastructure needed to train and run large AI systems, alongside investment in models and software.
Our read: Crusoe is pursuing both very large campuses and smaller modular capacity. The funding is evidence of investor demand for AI infrastructure, not proof that any one deployment model will dominate.
Open the source for story 1
technology · Reuters
German court finds Meta liable for fraudulent ads
Confirmed: The Frankfurt Regional Court ruled that Meta can be held directly liable for fraudulent investment ads distributed on Facebook and Instagram because its advertising system actively ranks and targets paid ads rather than merely hosting third-party content. The ruling can be appealed.
Why it matters: The judgment tests how EU hosting protections apply when a platform algorithmically distributes paid advertising, but it is a regional-court decision and does not by itself establish an EU-wide rule.
Our read: If upheld, the decision could increase legal pressure on platforms to police paid-ad delivery systems more closely. Its wider significance will depend on appeals and how other courts treat similar cases.
Open the source for story 2
developer · blog.ferstar.org
ZCode patches Git-history upload behaviour after security report
Confirmed: A security researcher reported that ZCode could upload workspace snapshot data, including Git history, to Z.ai servers during Repo Wiki or indexing activity without sufficiently clear disclosure. Z.ai said the uploads were associated with Repo Wiki generation, said uploaded data was destroyed after processing, and released a patch while committing to a third-party security audit.
Why it matters: The report and vendor response show why developers need explicit controls and clear disclosure when coding tools send repository context to cloud services.
Our read: The exact scope and trigger conditions are disputed, so it is safer to describe this as a transparency and data-handling issue than as proven malicious exfiltration.
Open the source for story 3
ai · Thomas Claburn/The Register
Anthropic adopts AGENTS.md spec for Claude Code
Confirmed: Anthropic added AGENTS.md support to Claude Code 2.1.277. The file can provide repository instructions when no CLAUDE.md is present. OpenAI previously contributed the AGENTS.md format to the Agentic AI Foundation.
Why it matters: Standardisation is rarely exciting, but for developers juggling multiple AI agents, having a common way to tell those agents how to behave is a practical step toward reducing configuration headaches.
Our read: Support from another major coding agent reduces setup friction for teams that use more than one tool and gives the shared convention broader practical reach.
Open the source for story 4
Practical takeaway
If you use AI coding assistants, verify which repository data can be uploaded, which feature triggers it and how long the provider retains it. For platform teams, treat the Frankfurt Meta judgment as a specific, appealable ruling about paid-ad distribution—not a settled Europe-wide rule on algorithmic content.
What to watch next
Watch for any appeal in the Meta case, Z.ai's promised third-party audit and further detail on the scope of ZCode uploads, and whether other coding assistants adopt AGENTS.md.