🗞 OpenAI's Billions, Google's AI Overhaul, and Anthropic's Claude Testing

🔎 The Latest on the AI Frontier:

• Google I/O 2025: Major AI Upgrades Across Entire Product Line

• OpenAI Raises $11.6B to Fund Massive Texas Data Center Expansion

• Anthropic Prepares for Claude 4 Launch with Internal Testing Underway

• Perplexity's Comet Browser Introduces Integrated AI Features for Enhanced Browsing

• Microsoft AI Security Presentation Accidentally Reveals Walmart's Confidential Plans

• Other news you might find interesting

🚀 Google's I/O 2025 developer conference unveils sweeping AI advancements across its entire product ecosystem.

• Gemini 2.5 introduces "Deep Think" capabilities that enable high-level research techniques, while Gemini 2.5 Flash requires 20-30% fewer tokens and both models now support audio-visual input and native audio dialogue via the Live API.

• Google launched two subscription tiers: Google AI Pro and the premium Google AI Ultra ($250/month), which provides access to the highest rate limits, early access to experimental products like Project Mariner, Veo 3, and Gemini 2.5 Pro with Deep Think mode.

• Android XR is expanding to smart glasses through partnerships with Gentle Monster and Warby Parker, featuring configurable in-lens displays for navigation, message visualization, real-time translation, and voice-activated photo capture, with expected market release later this year.

💻 OpenAI secures massive $11.6 billion funding to expand its largest data center project in Texas.

• The funding will allow OpenAI to expand its Texas data center from two to eight buildings, with each building housing up to 50,000 Nvidia Blackwell chips, bringing the total project funding to $15 billion.

• This expansion marks a crucial step in OpenAI's infrastructure development, helping the company reduce dependence on Microsoft while building capacity for more complex generative AI tasks.

• The data center is being built by Crusoe, a startup that pivoted from cryptocurrency to AI infrastructure and is now considered one of the emerging "neoclouds" specializing in AI cloud services—the same company tapped by Oracle to build the first data center for the $500 billion Stargate project.

🔬 Early testing of next-generation Claude 4 models suggests Anthropic is nearing public launch phase for its most advanced AI systems.

• Anthropic's internal web configurations now reference Claude Sonnet 4 and Claude Opus 4, with early access limited to a group of internal testers called "Ants" and the models currently operating under heavy rate limitations.

• The new models are categorized under the ASL-3 safety tier, Anthropic's internal classification for higher capability models with increased misuse potential, placing them above current ASL-2 models.

• If Anthropic maintains its typical release schedule, Claude 4 could officially arrive around June, potentially featuring more autonomous reasoning capabilities and deeper context retention for users already invested in Claude 3 or Opus tiers.

🌐 Perplexity launches Comet Browser with deeply integrated AI capabilities that transform the traditional browsing experience.

• Comet Browser features multiple AI entry points (homepage input, address bar, and dedicated sidebar) that allow Perplexity to see and interpret tab content, enabling webpage summarization and in-chat screenshot insertion.

• The browser's standout feature is its agentic capabilities—Perplexity can autonomously manage tabs and execute tasks across platforms like Gmail, Google Calendar, and social media when granted permission.

• This marks a strategic shift for Perplexity from a passive Q&A engine toward becoming a full-stack AI agent embedded in daily workflows, directly competing with Microsoft's Edge+Copilot but with deeper control over the browsing environment itself.

🔍 Microsoft's AI security chief accidentally exposes Walmart's confidential plans during a disrupted Build conference session.

• Neta Haiby, Microsoft's head of security for AI, unintentionally shared her screen showing internal Microsoft Teams messages that revealed Walmart's readiness to implement Microsoft's Entra Web and AI Gateway services.

• The exposure occurred after protesters disrupted Haiby's session on AI security practices, with former Microsoft employees Hossam Nasr and Vaniya Agrawal protesting the company's cloud contracts with the Israeli government.

• The leaked messages included a quote from a Walmart AI engineer stating: "Microsoft is WAY ahead of Google with AI security. We are excited to go down this path with you," confirming the retail giant's expanded commitment to Microsoft's AI services beyond their existing Azure OpenAI implementation.

More news you might find interesting:

• MIT Technology Review's analysis reveals AI's growing energy footprint is larger and more complex than previously understood.

• Google begins rolling out Flow, its comprehensive video creation platform, to selected early testers through Google Labs.

• Google invests up to $150M in partnership with Warby Parker to develop AI-powered smart glasses.

• New research highlights potential for large language models to be weaponized as psychological interrogation tools with concerning implications.

• Top AI researchers command unprecedented compensation as tech giants battle for scarce talent in the industry's intensifying arms race.

Have any feedback? Send us an email