AI Market Intelligence Daily

1. The most dramatic movement today was the 180% surge in the "Ari Weinstein leaves OpenAI" market (13% → 36%) on Polymarket. This move is notable because no headline directly addresses Weinstein's status — the shift appears driven by trading activity or private signals rather than public news. The simultaneous 18% drop in the Conrad Kramer departure market (41% → 34%) suggests traders may be repositioning around OpenAI's executive stability thesis more broadly: if Weinstein's exit odds are rising, Kramer's are being reassessed downward, possibly reflecting a belief that one departure doesn't necessarily cascade to the other.

2. Grok 4.6 release expectations on Kalshi are the clearest news-correlated cluster. The Aug 9 deadline jumped 153% (17% → 43%) and Aug 10 by 105% (21% → 43%), while the Aug 8 date dipped 16%. This pattern — sharp probability increases on Aug 9-10 paired with a collapse on earlier dates — strongly suggests leaked or anticipated information pointing to a Grok 4.6 release window in the Aug 9-10 range. The Aug 11 market already shows 0% probability, indicating traders have no expectation of a release past that window either. The Grok 4.7 Aug 31 market also rose 48% (21% → 31%), which is consistent with a revised release cadence where 4.6 arrives sooner and 4.7 follows later in the month.

3. Google's Gemini 3.5 Pro market shows a textbook deadline-convergence pattern. The Aug 7 market collapsed 83% (6% → 1%), essentially pricing out a release by tomorrow. Meanwhile the Aug 16 market rose 11% (35% → 39%) and Aug 31 rose 9% (70% → 76%). This redistribution of probability mass from near-term to mid-term deadlines is standard when a product is delayed but not cancelled. Notably, the Jeff Dean exodus from Google (top news story) is not reflected in any Gemini pricing — suggesting the market considers Dean's departure as separate from the product roadmap, or that the news is too recent to have been digested.

4. Two significant divergences between news sentiment and market pricing warrant attention. First, OpenAI's consumer hardware market dropped 20% (38% → 30%) despite news explicitly reporting an AI smart speaker priced at $300-$400. This is a clear bearish bet against bullish news — either traders doubt the product will ship by year-end, or the $300-$400 price point signals a product that may face commercial headwinds. Second, Anthropic is actively hiring an AI chip design team (top news), signaling serious long-term infrastructure investment, yet the Mythos release markets fell across the board (Sep 1: -50%, Nov 1: -33%). This divergence between hiring momentum and product release probability suggests the chip team is viewed as a multi-year play, not an accelerant for near-term model releases.

5. The OpenAI IPO complex remains heavily skewed against a 2026 listing (14% Yes, down 23%), with the "will not IPO by Dec 31" market at 84%. The IPO-day market cap distribution is extremely flat — all buckets sit at 0-4%, indicating the market has essentially given up on an IPO this year and there's no consensus on what a 2027 IPO valuation would look like. The Anthropic IPO-before-2027 market at 66% stands in stark contrast, reflecting the market's belief that Anthropic's path to public markets is clearer and more imminent than OpenAI's.