AI Market Intelligence Daily

1. Largest moves cluster in Kalshi's model-release markets as short-dated contracts decay toward their deadlines. Anthropic's "release Mythos before Sep 1, 2026" fell 50% (2% to 1%) and Grok 4.7 "before Aug 28" fell 29% (21% to 15%), both with only a few dollars of volume. These are classic time-decay moves: as an expiry approaches with no announced release, the near-dated contract bleeds toward zero while the longer-dated legs hold or gain — Grok 4.7 "before Sep 4" sits at 44% and the "next Mythos-Class model before Oct 1" market at 71%. The pattern signals a stretch of the expected release timeline, not a belief that the releases are dead.

2. The only top-15 story that maps directly onto a moving market is the Grok quality complaint, and it is consistent with the Grok release markets sliding. TechCrunch reports "Grok keeps sending gibberish responses to users," a negative quality signal that supports lower odds of a Grok 4.7 shipping imminently, so the Aug 28 and Aug 31 contracts drifting down to 15% and 20% reads as news-aligned rather than noise. By contrast, nearly every other large move — Anthropic's Mythos and Opus 6 windows, Google's Gemini 3.5 Pro "before Aug 31" at 7% — has no corresponding headline, confirming the report's own takeaway that most of today's price action is trading activity and deadline decay rather than headline-driven sentiment.

3. The OpenAI IPO complex is internally inconsistent and the clearest divergence to flag. "Will OpenAI not IPO by December 31, 2026?" sits at a high-conviction 79%, yet the affirmative "Will OpenAI IPO by December 31, 2026?" is at 18%, a 3-point gap that should be near-exact complement in an efficient market. At the same time the implied valuation is drifting down: the "between $1T and $1.25T" bucket rose 8% (4% to 5%) and "between $1.25T and $1.5T" fell 17% (2% to 2%), while both the sub-$500B and $500B–$750B buckets sit at 0%. The market has effectively ruled out a low-valuation listing and is concentrating mass just under $1T, yet no top-15 story explains a reprice, so this is most likely thin-book repositioning worth watching for volume confirmation.

4. Two remaining markets show price moving against the tone of the news. "Will Ari Weinstein leave OpenAI by December 31, 2026?" fell 24% (46% to 35%) with no negative OpenAI-personnel headline in the feed, while the adjacent Conrad Kramer departure market ticked up 7% (15% to 16%); the opposite directions in the same personnel complex suggest churn in thin books rather than a coherent news narrative. And "What will be the top AI model this month?" is locked at 91% for claude-opus-5-high with no release headlines this week, so the current leader's dominance is being priced as a lock-in — consistent with the near-term release markets decaying and the absence of any new frontier model to displace it.