Anthropic has released its latest flagship model, Claude Opus 4.8, which introduces a breakthrough in 'self-honesty'—making the AI four times less likely to hallucinate or make unsupported claims compared to Opus 4.7. Simultaneously, the company announced a massive $65 billion Series H funding round led by Altimeter Capital, bringing its valuation to $965 billion and officially surpassing OpenAI in the private funding race. Anthropic also reported a record $47 billion revenue run-rate driven by enterprise adoption.
The AI industry has shifted from a 'benchmark war' to a 'deployment war.' Major labs are now launching dedicated consulting arms, such as OpenAI's newly formed $4 billion subsidiary 'DeployCo,' which embeds forward-deployed engineers directly into client organizations like BBVA and Goldman Sachs to manage model integration and custom data workflows.
As of May 2026, memory chip giants SK Hynix and Micron have officially joined the '$1 Trillion Club' in market capitalization, driven by the unprecedented demand for High Bandwidth Memory (HBM4) used in AI data centers.
Which specific capability does Anthropic claim is improved by 4x in the new Claude Opus 4.8 model?