Latest AI News

Apple's persistent struggles with artificial intelligence are jeopardizing its product ecosystem, according to a recent Bloomberg report. Despite high-profile hires like John Giannandrea from Google in 2018, Apple has failed to match competitors in AI development, leading to leadership reshuffling with Vision Pro creator Mike Rockwell taking over Siri development. These ongoing challenges threaten not only the iPhone's market position but also Apple's ambitious plans for robotics and other futuristic products.

Stanford's 2025 AI Index Report reveals the cost of querying high-end AI models has dropped dramatically from $20 to just $0.07 per million tokens in 18 months. While the United States maintains leadership in AI development, China is rapidly closing the performance gap, with the difference between top models shrinking from 9.26% to 1.70% in just one year. The comprehensive report also highlights a concerning 56.4% increase in harmful AI incidents, underscoring the need for stronger responsible AI practices.

Google has launched the AI Futures Fund, a strategic initiative designed to invest in and collaborate with promising AI startups at various growth stages. The program offers early access to Google DeepMind's advanced AI models, including Gemini, alongside technical expertise and equity funding. Unlike traditional accelerators, the fund evaluates opportunities on a rolling basis, positioning Google to compete with Microsoft and Amazon in the race to shape AI's future.

Leading tech executives are increasingly highlighting their personal use of their companies' AI products as evidence of quality and safety. Microsoft's Nadella, Google's Pichai, and Meta's Zuckerberg have all emphasized how they're using their own AI tools internally, with some claiming that 20-30% of their companies' code is now AI-generated. This trend of 'eating their own AI dogfood' serves both as a marketing strategy and as a practical approach to testing and improving their technologies.

Vistra Corp. has agreed to acquire seven natural gas power plants from Lotus Infrastructure Partners for $1.9 billion, positioning itself to meet surging electricity demands from AI data centers. The deal includes approximately 2,600 megawatts of generation capacity across key U.S. markets, with five combined cycle gas turbine facilities and two combustion turbine facilities. This strategic acquisition comes as the U.S. Energy Information Administration forecasts record electricity consumption in 2025 and 2026, largely driven by power-hungry AI technologies.

Two major AI breakthroughs have emerged in May 2025, with Google's AMIE bringing advanced medical imaging interpretation to healthcare and Alibaba's Qwen3 narrowing the gap with leading US AI firms. Google's multimodal AMIE can intelligently analyze visual medical information, often outperforming human doctors in diagnostic accuracy. Meanwhile, Alibaba's Qwen3 offers groundbreaking multilingual capabilities across 119 languages and introduces hybrid reasoning that combines traditional LLM functions with advanced dynamic reasoning.

Brazilian star Raphinha has embraced cutting-edge AI technology to enhance his training regimen following Barcelona's domestic treble success. The innovative approach combines virtual reality simulations with performance analytics to create personalized training programs. This technological advancement comes as Barcelona prepares to extend Raphinha's contract until 2029, rewarding his breakthrough season that saw him score 31 goals across all competitions.

Microsoft's GitHub has launched a new autonomous coding agent that can independently handle programming tasks when assigned issues through GitHub. The agent works in a secure environment powered by GitHub Actions, creating draft pull requests that developers can review and modify. Available to Copilot Pro+ and Enterprise subscribers, this new capability aims to free developers from routine coding tasks while maintaining human oversight of the final code.

Microsoft has integrated xAI's Grok 3 and Grok 3 mini models into its Azure AI Foundry platform, expanding its AI model offerings beyond OpenAI. The Grok models will be hosted and billed directly by Microsoft, providing enterprise-grade reliability guarantees. This move comes amid Microsoft's strategy to diversify its AI model partnerships, now offering over 1,900 models from various providers.

Chinese tech giant Xiaomi has committed to investing $6.9 billion over the next decade to develop its own semiconductor chips, starting from 2025. The company is set to unveil its advanced Xring O1 processor, manufactured using cutting-edge 3-nanometer technology, at a major event on May 22. This strategic move represents Xiaomi's ambitious return to in-house chip development amid intensifying global competition in AI semiconductor technology.

At Microsoft Build 2025, the company unveiled Microsoft 365 Copilot Tuning, a groundbreaking low-code capability that allows organizations to customize AI models using their own data without requiring data science expertise. Microsoft also introduced multi-agent orchestration, enabling multiple AI agents to collaborate on complex tasks with human oversight. These innovations mark a significant advancement in Microsoft's vision for human-agent collaboration, with the company projecting businesses will deploy 1.3 billion AI agents by 2028.

Dell Technologies showcased its next-generation AI infrastructure featuring NVIDIA's Blackwell Ultra chips at Dell Technologies World 2025 on May 19. The company introduced new PowerEdge servers supporting up to 256 NVIDIA Blackwell Ultra GPUs per rack, delivering up to four times faster large language model training. Dell's AI factories are demonstrating significant ROI with productivity gains of 20-40%, contributing to AI's projected $15 trillion global economic impact by 2030.

Italy's data protection authority has imposed a €5 million ($5.64 million) fine on Luka Inc., the developer of AI chatbot Replika, for violating data protection regulations. The San Francisco-based company, which offers customized AI avatars marketed as emotional companions, was found to lack proper legal basis for processing user data and failed to implement age verification systems to protect minors. This action follows a temporary ban on the service in Italy since February 2023.

Honda Motor announced on May 20, 2025, that it's scaling back its electric vehicle investment by 30% to 7 trillion yen ($48.4 billion) due to slowing demand, abandoning its previous target of 30% EV sales by 2030. The company will instead focus on hybrid vehicles, planning to launch 13 next-generation hybrid models globally by 2031. Despite this shift, Honda remains committed to its Honda 0 Series EVs featuring the advanced ASIMO OS and AI-driven driver assistance systems.

Honda Motor Co. has announced a significant advancement in its partnership with Renesas Electronics to develop a high-performance system-on-chip (SoC) for its upcoming Honda 0 Series electric vehicles. The AI-powered chip will deliver industry-leading performance of 2,000 TOPS while supporting Honda's ASIMO OS and advanced driver assistance systems. This development represents a major step in Honda's goal to become the first automaker to enable eyes-off driving in all situations.

At Computex 2025, NVIDIA unveiled NVLink Fusion, a strategic move that opens its previously closed ecosystem to allow third-party CPUs and AI accelerators to connect directly with NVIDIA GPUs. This shift represents NVIDIA's evolution from a proprietary full-stack provider to an open infrastructure enabler, positioning the company at the center of global AI development even as competitors build custom solutions. CEO Jensen Huang described this as part of a 'tectonic shift' requiring fundamental rearchitecting of data centers worldwide.

NVIDIA has unveiled DGX Cloud Lepton, an AI platform that connects developers with tens of thousands of GPUs through a global network of cloud providers. The marketplace addresses the critical challenge of securing reliable, high-performance GPU resources by unifying access to cloud AI services across the NVIDIA compute ecosystem. This initiative aims to bridge the gap between overwhelming AI computing demand and limited GPU supply, enabling developers to find and utilize the right resources for their specific needs.

NVIDIA and Foxconn are partnering with Taiwan's government to build a state-of-the-art AI supercomputer featuring 10,000 NVIDIA Blackwell GPUs. The project, announced by NVIDIA CEO Jensen Huang at Computex 2025, will deliver advanced AI infrastructure to researchers, startups, and industries including TSMC. Alongside this initiative, NVIDIA unveiled plans for a new Taiwan headquarters called 'Constellation' in Taipei's Beitou-Shilin Technology Park.

Jensen Huang announced NVIDIA's next-generation GB300 AI platform at Computex 2025, set for Q3 release with significant performance improvements over its predecessor. The Blackwell Ultra-based system delivers 1.5x more AI performance than the GB200 NVL72 and features enhanced memory capacity designed specifically for AI reasoning workloads. This launch reinforces NVIDIA's strategy to dominate the rapidly expanding AI infrastructure market that Huang predicts will reach trillion-dollar scale.

At its annual I/O conference on May 20, 2025, Google introduced a premium 'AI Ultra Plan' subscription priced at $249.99 monthly, offering users the highest usage limits and early access to experimental AI tools. The plan includes access to Project Mariner, a browser extension that automates web tasks, and Deep Think, an enhanced reasoning version of Gemini. Google also showcased Android XR smart glasses featuring real-time translation capabilities, developed in partnership with Samsung.