October 26, 2026 — today in AI.
1. Anthropic's IPO Plans: Record Losses, AI Risks, and Growth Potential in the S-1
Anthropic's S-1 filing lays out record losses and the risk factors it must disclose publicly, after Microsoft's $5 billion investment in a round that valued the lab around $350 billion. For practitioners, the filing is the first audited look at Anthropic's unit economics and the platform dependencies behind Claude's API pricing and availability.
2. Moonshot's Kimi K3.1 Pro Launches on Kimi.com and API, with Full Weights Due by November 4
Moonshot released kimi-k3.1-pro and kimi-k3.1-pro-highspeed on Kimi.com, the mobile apps, Kimi Work 3.1.0+, Kimi Code, and the Kimi API, and says full model weights will ship by November 4, 2026. For builders, this is a near-term chance to evaluate a frontier-class open-weight model for self-hosting, with the timing of the weight release the main thing to watch.
3. AI Agents Are Starting to Set the Prices You Pay Online
Salesforce now prices its Agentforce platform per conversation, with 2026 list rates starting at $2 per interaction. This moves agent costs from seat licenses to usage-based metering, so teams building on agent platforms need to model per-interaction cost against the human-handled baseline.
4. The Agentic Shift: Enterprise AI Agents in 2026
Uber has deployed ChatGPT Work to its entire 30,000-person workforce, and Anthropic has launched Claude Duet to compete in agentic workflows. Two major vendors are now competing directly for enterprise agent seats, which signals that buyers are moving from pilots to company-wide rollouts.
5. AI News Today, October 26: Top Stories
Amazon's CEO Andy Jassy put 2026 capital expenditure at roughly $220 billion and said the company 'could invest a lot more,' describing demand from labs and enterprises as insatiable. Sustained hyperscaler capex at this level keeps GPU and data center capacity tight, which affects inference pricing and availability for teams planning around cloud capacity.