NVIDIA's $150B Taiwan Investment and Anthropic's Claude Opus 4.8 Lead AI Advancements

NVIDIA announces a massive $150 billion investment in Taiwan, solidifying the island as the epicenter of AI technology, while Anthropic launches its latest AI model, Claude Opus 4.8, with enhanced capabilities.

NVIDIA Invests Heavily in Taiwan

NVIDIA is investing $150 billion annually to ensure Taiwan remains at the forefront of the AI revolution. The company plans to establish a new headquarters in Taiwan, expanding its partnership with TSMC and leveraging advanced packaging technology not yet available in the US. This move aims to boost NVIDIA's alliances with other local partners and further expand the AI ecosystem.

Anthropic Unveils Claude Opus 4.8

Anthropic introduces Claude Opus 4.8, an upgraded version of its AI model. The new release builds on Opus 4.7, featuring sharper judgment, more honesty about its own progress, and the ability to work independently for longer periods. Claude Opus 4.8 is now available at the same price point as its predecessor.

AI Companies Find Product-Market Fit

Both Anthropic and OpenAI have found their product-market fit, leading to aggressive pricing for their APIs. These companies are now charging over $200 per month per user, which helps cover costs more effectively than the previous $10 to $20 per month per user. Coding agents, in particular, drive this increased spending.

MLXcel Goes Open Source

MLXcel, an inference engine built for Apple Silicon, is now open source under the Apache 2.0 license. On M5 Max, MLXcel reaches up to 2.70x the prefill median of mlx-lm and matches it on decode. The supported-models list includes over 70 text architectures and 22 vision-language models (VLMs). By going open source, MLXcel aims to democratize AI inference, making it accessible beyond a few GPU vendors and cloud providers.

Research Reveals Secrets to Better MoE Models

A study of 2,000 training runs uncovers key design choices for better mixture-of-experts (MoE) models. MoE models keep most of the model inactive, only activating a small slice to handle each input. The research identifies critical factors that improve the efficiency and performance of these models, helping developers get the architecture right before committing significant computational resources.

Cognition Raises $1B for AI Coding Agent Devin

Cognition raises $1 billion at a $26 billion valuation for its AI coding agent, Devin. The valuation has more than doubled from $10.2 billion just eight months ago, and the run-rate revenue has grown from $37 million to $492 million in twelve months. Devin automates coding tasks, writing, testing, and opening pull requests without human intervention, making it a valuable tool for developers.

References

• AI News Briefs BULLETIN BOARD for May 2026 | Radical Data Science