Anthropic today launched Web Search on its Claude API—allowing models to autonomously query, refine, and synthesize live web data with enterprise‑grade governance and citations, priced at $10 per 1,000 searches.

Duolingo adopts an AI‑First model—replacing contractor tasks with generative AI, launching 148 courses in under a year, and raising 2025 revenue forecast to nearly $1 billion—while maintaining quality through human oversight and strategic AI partnerships.

Visa Intelligent Commerce empowers AI agents with tokenized payments, personalization, and real-time controls. Discover how this initiative streamlines agentic shopping.

Amazon’s Nova Premier debuts on Bedrock with 1M-token context, text/image/video support, and competitive pricing ($2.50/$12.50 per M tokens). It excels at retrieval and visual tasks and serves as a teacher for distilling smaller Nova variants for edge and cloud deployments.

Microsoft launched three Phi 4 reasoning models—mini (3.8 B), standard (14 B), and plus (14 B reasoning-tuned)—that achieve accuracy on par with DeepSeek R1 and OpenAI’s o3-mini, enabling robust edge and cloud AI deployments under a permissive license.

Colosseum, a DGX SuperPOD powered by Grace Blackwell Superchips and Vertiv’s modular design, launches in Italy 2025 as a sovereign AI data center for finance, healthcare, and public sectors, offering compliance, efficiency, and replicable AI factory blueprints.

At LlamaCon on April 29, 2025, Meta unveiled the Llama API and a standalone AI assistant app—empowering businesses to integrate Llama models with one line of code, retain full model control, and prepare for upcoming paid subscription services.

Palo Alto Networks will purchase Protect AI—valued above $500 million—to accelerate Prisma AIRS, offering integrated model scanning, runtime protection, posture management, and agent security, enabling enterprises to deploy AI securely and at scale.

Microsoft blocked $4 billion in AI-powered fraud, thwarting 1.6 million bot sign-ups hourly. The Cyber Signals report exposes rapid scam evolution—fake stores, deepfake phishing, and job fraud—and recommends adaptive, AI-based defenses.