Latest AI (Artificial Intelligence) News
Trump signs executive order for federal review of frontier AI models
President Donald Trump signed an executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security,” creating a voluntary framework for leading AI companies to submit frontier models for federal review before release. Reports say the review window is 30 days, and the White House framed the move as balancing innovation with national-security risk.
Nvidia brings AI directly into personal computers
Nvidia unveiled the RTX Spark superchip, designed to bring advanced AI functions to Windows laptops and desktops. The company says these AI PCs will be able to run agents locally, with new systems from partners including Microsoft and Dell expected later this year.
Anthropic files confidentially for an IPO
Anthropic confidentially filed a draft S-1 with the SEC, moving toward the public markets and reportedly getting ahead of OpenAI in the IPO race. The filing comes after a new funding round that the source says valued the company at roughly $96.5 billion and implied a $47 billion revenue run-rate.
Google Gemini 3.5 Flash reaches general availability
Google’s Gemini 3.5 Flash reached general availability, making the model broadly available for production use. The move was part of a week in which major vendors focused less on standalone models and more on the environments where AI agents run.
OpenAI plans to retire GPT-4.5 from ChatGPT
OpenAI confirmed that GPT-4.5 will be retired from ChatGPT on 27 June, according to the report. The timing places the change alongside other June product shifts as major AI vendors reposition their offerings around newer models and agent workflows.
Microsoft ships its own coding model into the editor
Microsoft used Build to release its coding model directly into the editor, signaling a push to embed AI more deeply into developer workflows. The source describes this as part of a broader industry move away from model-centric marketing toward integrated AI runtimes and control planes.
Big Tech shifts the AI race toward the 'agent runtime'
Across Microsoft Build, Nvidia’s GTC Taipei, and ServiceNow Knowledge, vendors emphasized the governed environment where autonomous agents operate rather than the model alone. The source says the industry is increasingly selling identity, audit, policy, isolation, and spend controls as the core AI product.
Anthropic expands access to its cyber-capable Mythos model
Anthropic expanded access to its cyber-capable “Mythos” model to roughly 150 more organizations across more than 15 countries through Project Glasswing. The expanded rollout reportedly reaches sectors including power, water, healthcare, communications, and hardware.
Nvidia confirms Vera CPU is in full production
Nvidia said its data-center Vera CPU is now in full production, with customers including OpenAI, Anthropic, and SpaceX among the first buyers. The company’s announcements point to continued demand for AI infrastructure at both the data-center and PC levels.
AI governance and infrastructure become the main product story
A broader theme across the week is that vendors are now competing on governance, parallelism, isolation, identity, auditability, and policy controls rather than model benchmarks alone. The report argues that the “control plane” has become the thing being sold as organizations look for safe, scalable ways to deploy AI agents.