AI News from August 14, 2026

Tencent reported weaker second-quarter profit as AI spending accelerated, putting pressure on investors to judge whether its infrastructure bet can produce returns.

Cognition is reportedly discussing a new funding round at a $40B valuation, testing investor confidence in Devin’s enterprise revenue growth.

Meta’s Glimmer release supports Zuckerberg’s open AI message, but its split model strategy raises questions about access, power, and control.

Z.ai says GLM-5.3 approaches Anthropic’s Mythos 5 in cybersecurity tests, but limited reporting leaves the benchmark and deployment picture unclear.

American Banker examines whether AI agents could disintermediate banks, alongside a $1.3 billion RIA’s succession model and its industry implications.

Upstage’s Solar Pro 4 is reported as the first Korean LLM listed on OpenRouter, widening visibility while leaving key performance and access details unconfirmed.

Zhipu AI has released GLM-5.3, touting stronger open-weights coding and cybersecurity performance while postponing its weights release for reviews.

Z.ai is positioning a new AI model against Anthropic and OpenAI in software development, intensifying competition for coding tools and enterprise users.

Databricks raised $5 billion at a $190 billion valuation after demand exceeded its plan, funding costly AI research, cloud commitments, and acquisitions.

Block’s open-source Goose offers local AI coding agents without subscriptions, challenging Claude Code’s premium pricing while exposing trade-offs in quality and hardware.

Recent coverage highlights the security gap around open-weight and open-source AI, pushing builders and enterprises to control models across the full lifecycle.

Reports that Anthropic AI agents launched a simulated virtual war are renewing questions about autonomous behavior, oversight, and agentic AI safety.