Jensen Huang Unveils NVIDIA's Vera Rubin Platform at GTC Taipei
NVIDIA CEO Jensen Huang took the stage at Taipei Music Center today to unveil the company's next-generation AI supercomputing platform: Vera Rubin. Speaking at GTC Taipei during Computex 2026, Huang called it "the most ambitious endeavor in the history of our company."
The Vera Rubin NVL72 connects 36 Vera CPUs and 72 Rubin GPUs unified by sixth-generation NVLink Switch technology, with ConnectX-9 SuperNICs and Spectrum-X Ethernet for scale-out networking, plus BlueField-4 DPUs for data acceleration. NVIDIA claims the platform delivers up to a 10x reduction in inference token cost and a 4x reduction in GPU count needed to train mixture-of-experts models compared to the Blackwell generation.
The platform is now in full production, with cloud partners including AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure set to deploy Vera Rubin-based instances in the second half of 2026. CoreWeave, Lambda, Nebius, and Nscale will follow as NVIDIA Cloud Partners.
Huang also showcased NVIDIA's physical AI edge strategy with the Jetson Thor module, the N1X ARM laptop chip, and DLSS 5 — reinforcing the company's "five-layer cake" approach spanning energy infrastructure to applications. The keynote underscored NVIDIA's vision of AI agents as "the latest breakthrough in computer science."
Anthropic Ships Claude Opus 4.8 with Improved Agentic Capabilities
Just days ago, Anthropic released Claude Opus 4.8, arriving less than two months after Opus 4.7. The new model shows meaningful improvements across coding, agentic tasks, reasoning, and practical knowledge work.
Key benchmark gains include agentic coding scores jumping from 64.3% to 69.2%, multidisciplinary reasoning with tools improving from 54.7% to 57.9%, and knowledge work scores rising from 1753 to 1890. Anthropic describes Opus 4.8 as having "sharper judgement, more honesty about its progress, and the ability to work independently for longer than its predecessors." Pricing remains unchanged from Opus 4.7.
The release follows Anthropic's Code with Claude developer conference series, which toured San Francisco (May 6), London (May 19-21), and heads to Tokyo (June 5-6). The biggest revelation from the events was a deal with SpaceX to allocate all capacity in its Colossus supercluster to Claude. New features shipped include Dreaming, Outcomes, multi-agent orchestration, Claude Finance with 10 pre-built agents, and Add-ins.
Anthropic also expanded its partnership with PwC, rolling out Claude Code and Cowork to the firm's hundreds of thousands of professionals globally.
Pope Leo XIV Calls on the World to 'Disarm' AI in Landmark Encyclical
Pope Leo XIV released his first encyclical on May 25, titled "Magnifica Humanitas" ("Magnificent Humanity"), urging governments and tech leaders to confront artificial intelligence before it weakens human relationships, critical thinking, and peace itself.
"To disarm does not mean rejecting technology, but preventing it from dominating humanity," the Pope wrote. He declared it "not permissible" to entrust irreversible, lethal decisions to AI systems and stated that the traditional "just war" theory is now outdated.
The encyclical addresses a sweeping range of contemporary crises — war, modern slavery, wealth inequality, the erosion of democracy, and the devaluing of human capacities. The Vatican presentation featured testimony from prominent cardinals, theologians, and notably Christopher Olah, co-founder and interpretability lead at Anthropic.
The document represents one of the most significant interventions by a global moral authority on AI governance, arriving as nations continue to grapple with regulatory frameworks for the technology.
Google Gemini 3.5 Flash Outperforms Last Year's Pro Model
Google's I/O 2026 conference delivered Gemini 3.5 Flash, which outscores last year's Gemini 3.1 Pro on key benchmarks while running roughly 4x faster and costing about 25% less. The model achieves 76.2% on Terminal-Bench 2.1, 83.6% on MCP Atlas, and 1,656 Elo on GDPval-AA.
API pricing comes in at $1.50 per million input tokens and $9.00 per million output tokens. The model is available via the Gemini app, AI Mode in Google Search, Gemini API, and the enterprise platform.
Google also confirmed that Gemini 3.5 Pro is being used internally and will roll out externally in June 2026. The company continues to integrate AI deeply across Search, productivity tools, and the new Android XR ecosystem including smart glasses from partners like Gentle Monster and Warby Parker.