AI News Daily

The Trump administration has approved Anthropic's release of its Mythos AI model to select companies and government agencies following a cybersecurity review. The conditional approval represents a middle ground between full public release and complete restriction, allowing limited deployment while security assessments continue. Source: CNBC — https://cnbc.com/2026/06/26/us-government-anthropic-claude-mythos5-ai.html

Companies are dramatically cutting AI spending as the era of unrestricted token consumption ends. Lindy CEO switched 100% of traffic from Anthropic's Claude to DeepSeek, reducing costs to near zero, while Uber implemented monthly spending caps starting at $1,500 per employee after blowing through its annual AI budget in four months. This shift threatens the growth trajectories of OpenAI and Anthropic ahead of their planned IPOs. Source: CNBC — https://cnbc.com/2026/06/26/openai-anthropic-new-ai-spending-reality-as-users-shift-to-efficiency.html

The Trump administration has requested that OpenAI stagger the release of its new GPT-56 model and approved Anthropic's Mythos model only for select companies and government agencies pending cybersecurity review. Both companies have agreed to limit distribution of their latest models as part of a government security assessment process. Source: AP News — https://apnews.com/article/trump-ai-openai-gpt56-sol-cybersecurity-mythos-065d5398baac7f16c8265c2cb8ba2baa