🗞 OpenAI's Enterprise o3-pro, ChatGPT's Water Impact, and Nvidia's $50T AI Vision

🔎 The Latest on the AI Frontier:

• OpenAI Debuts Enterprise-Focused o3-pro Model with Premium Pricing

• Sam Altman Reveals ChatGPT's Water Usage Per Query Statistics

• Nvidia CEO Projects $50 Trillion Physical AI Market Opportunity

• ChatGPT Suffers Extended 15-Hour Service Outage Across Platforms

• Apple Explains Why Advanced Siri Features Are Delayed Until 2026

• Tesla Plans Austin Robotaxi Launch for June with Safety Focus

• Other news you might find interesting

🤖 OpenAI launched o3-pro, a more powerful AI model designed for enterprises that prioritizes reliability and tool integration over speed, with pricing 10 times higher than the standard o3 model.

• The model costs $20 per input and $80 for outputs compared to o3's reduced pricing of $2 and $8, while taking significantly longer to respond—early users report 2-3 minute response times for simple queries like "Hi, I'm Sam Altman."

• o3-pro offers enhanced tool access including web search, file analysis, visual reasoning, and Python capabilities, with OpenAI evaluations showing superior performance in science, education, programming, business, and writing domains.

• The launch coincides with OpenAI reaching 3 million business users and 50% enterprise growth since February, positioning the model for challenging applications where accuracy matters more than speed in the competitive reasoning model market.

💧 OpenAI CEO Sam Altman claimed in a new blog post that an average ChatGPT query uses approximately 0.000085 gallons of water, equivalent to "roughly one fifteenth of a teaspoon."

• Altman provided the unsourced statistic alongside energy usage claims that each query consumes about 0.34 watt-hours, comparable to what an oven uses in just over one second or a high-efficiency lightbulb uses in a couple of minutes.

• The disclosure comes amid growing scrutiny of AI companies' environmental impact, with researchers forecasting AI could consume more power than Bitcoin mining by year-end and previous studies showing a 100-word GPT-4 email requires "a little more than 1 bottle" of water.

• OpenAI did not respond to requests for comment on how Altman calculated these figures, while industry analysis shows water usage varies significantly based on datacenter location and cooling requirements.

🚗 Nvidia CEO Jensen Huang outlined a $50 trillion market opportunity for physical AI systems spanning factories, transportation, and humanoid robots at the company's GTC Paris event.

• Nvidia's automotive business is projected to reach $5 billion this year with trillion-dollar potential as autonomous vehicles scale, though only 1% of current road vehicles have L2+ capabilities today.

• The company announced its full-stack Drive AV software is now in production starting with Mercedes Benz CLA sedan, with upcoming launches including Volvo's ES 90 sedan and JLR's next-generation fleet in 2026.

• Nvidia is expanding into industrial AI with partnerships including Siemens and Deutsche Telekom to build Germany's first industrial AI cloud featuring 10,000 GPUs, addressing an anticipated 50 million person labor shortage over the next five years.

💻 ChatGPT experienced a major outage on Tuesday that left users unable to access the AI service for over 15 hours.

• The outage began at 3am ET and peaked with nearly 2,000 user reports on Downdetector, affecting ChatGPT's website, desktop apps, mobile apps, and APIs across all platforms.

• OpenAI identified the root cause by 9:07am but required until 6:30pm ET to restore most services, with voice mode still experiencing elevated error rates as of the latest update.

• Users seeking alternatives during AI downtime can turn to Google Gemini, Microsoft Copilot, Anthropic's Claude AI, and Perplexity, with free versions available for most basic generative AI needs.

🍎 Apple executives revealed why the promised new Siri features have been delayed and won't arrive until 2026, citing architectural limitations that failed to meet company standards.

• Apple was simultaneously developing two versions of Siri's underlying architecture, with V1 proving inadequate for customer expectations despite initial confidence it would launch by December 2024 or spring 2025.

• The company confirmed that advanced Siri capabilities including personal context understanding, on-screen awareness, and in-app actions will arrive "in the coming year" meaning after iOS 26 launches in fall 2025.

• Apple emphasized its AI strategy focuses on integrating intelligence into existing apps rather than building standalone chatbots, with new features like Live Translation, Call Screening, and enhanced Visual Intelligence coming to iOS 26.

🚗 Elon Musk announced Tesla's robotaxi service will tentatively launch June 22 in Austin, Texas, with 10-20 Model Y SUVs operating under remote human supervision in a limited area.

• The announcement comes with the caveat "We are being super paranoid about safety, so the date could shift," marking another timeline promise after previous failed predictions including a 2019 claim for a 2020 launch that never materialized.

• Tesla enters a market already dominated by Waymo, which provides hundreds of thousands of paid rides weekly across Phoenix, San Francisco, Los Angeles, and Austin, though recent vandalism incidents in Los Angeles highlight security risks for unmanned vehicles.

• The service faces potential vulnerabilities from hostile actors who could summon empty robotaxis to remote locations for destruction, a concerning scenario given public sentiment divisions around Musk and previous Tesla vandalism incidents.

More news you might find interesting:

• Google Cloud launched G4 virtual machines powered by NVIDIA RTX PRO 6000 Blackwell Server edition, becoming the first cloud provider to offer this advanced GPU combination for large-scale AI inference and simulation workloads.

• Mistral AI launched its new Magistral model family, featuring advanced reasoning capabilities to compete directly with OpenAI's o3 and Anthropic's Claude 4 Opus in the high-reasoning AI market.

• The Trump administration is reportedly developing an AI chatbot accessible through AI.gov, with plans to launch by July 4, 2025, according to code discovered on GitHub and early website versions.

• Schneider Electric and Nvidia signed a global agreement to develop cooling and power management systems for Europe's AI datacenters, targeting support for ultra-high-density 1 MW racks.

• Researchers at Hong Kong Polytechnic University developed VideoMind, a multi-modal AI agent that mimics human thinking processes to analyze and reason about long videos exceeding 15 minutes in duration.

• Google added Gemini AI to Google Forms with a new "Summarize responses" feature that automatically analyzes text-based survey answers and identifies key themes.

• Talen Energy and Amazon signed an expanded nuclear power agreement to supply up to 1,920 megawatts of carbon-free electricity from Pennsylvania's Susquehanna plant to AWS data centers through 2042.

• Google overhauled its internal learning platform "Grow" to focus exclusively on AI training, eliminating over 500,000 diverse course listings that previously included everything from product development to personal finance and Rubik's cube solving.

Have any feedback? Send us an email