Here are recent rough notes from the site.
Captive insurer affiliates to private credit/equity are big contributors to the parent organizations, as I have pointed out many times. As ever, this creates a duration mismatch that doens't matter until it does, as we've seen to a limited degree in the recent retail race to liquodty causing lockups at Blue Owl and others.
https://www.ft.com/content/b6be87a9-0abf-4950-9060-2159aa547f3d
- Inference chiip maker Cerebras filles confidentially for an IPO (The Information)
- Insurance is the lifeblood of priatate credit and thus, increasingly, of AI capex (FT)
- IBM share fall 12% on ANthropic tools for defunct COBOL (CNBC)
- Anthropic accuses Chinese AI companies of launching concerted distillation attacks to clone models (Anthropic(
Using the first post-earnings weekly expiry (Fri, Feb 27, 2026) and the nearest ATM strike, Nvidia options are currently pricing about a ±6.2% move into earnings (Wed, Feb 25, 2026). I would take the over.
LLMs are a security crisis, according to a major new report from the Office of the Director of National Intelligence (ODNI) on AI . Not sure how this slipped out in all the administration AI bonhomie, but it is a good report about rapidly growing tail risks to ... everything.
Trojans in Artificial Intelligence: Final Report
https://arxiv.org/pdf/2602.07152
