AI tools, automation platforms, and enterprise adoption strategies
Moonshot AI's Kimi K3 undercuts frontier models by 90% on cached tokens before open weights drop in July. Anthropic now controls 40% of enterprise spend.
China-based DeepSeek introduced time-of-day pricing with 75% off-peak discounts on inference APIs. Enterprise buyers now have a hard benchmark for batch workloads and negotiation leverage against OpenAI, Anthropic, and Google.
KPMG signed a global deal to deploy Microsoft 365 Copilot and Agent 365 across 276,000 professionals, marking one of the largest confirmed enterprise AI rollouts in professional services.
Anthropic launched an enterprise marketplace allowing Claude customers to buy third-party apps using existing contract budgets, reducing procurement friction but increasing vendor lock-in.
New EU transparency obligations for AI-generated content take effect August 2, 2026, with fines up to €35 million. Colorado's high-risk AI law adds U.S. compliance layer.
Anthropic's new model compressed a 50-million-line code migration from 60 days to one day before regulators forced its withdrawal. The recall creates new procurement risk for frontier API deployments.
Meta's Hyperion expansion and TSMC's $60–64 billion capex hike signal sustained GPU capacity growth, lowering scarcity premiums and improving availability for enterprise AI workloads through 2028.
Anthropic has overtaken OpenAI in US business adoption, claiming 54% of enterprise coding spend while KPMG-Microsoft push agentic AI into governed production with 66% task success rates.
IBM's July 9 update gives enterprise IT teams multi-agent orchestration, built-in AI spend tracking, and pre-built modernization workflows — shifting agent evaluation toward governance and economics.
The EU pushed high-risk AI compliance to August 2028, the FTC proposed new rules on accuracy suppression, and Singapore issued the first runtime framework for agentic finance.
New deployment frameworks put hard numbers on when to move from API-only to self-hosted LLM infrastructure, with specific token-cost thresholds changing budget and vendor planning.
Google priced Gemini 3.5 Pro's 2-million-token context window at $1.25 per million input tokens, forcing enterprise buyers to remodel TCO for long-document and agentic workflows against OpenAI and Anthropic.