Technology News
Daily Tech Reader - Technology Edition
AI Software & Platforms
AI Software & Platforms
Anthropic Signs $11.6 Billion Cloud Deal With Akamai — CPU Bet Could Grow to $20 Billion
Anthropic has committed $11.6 billion over seven years to Akamai Technologies for cloud infrastructure and software. The deal centers on CPUs rather than GPUs — a deliberate bet on general-purpose compute as AI agents take on more tasks. It is the largest contract in Akamai's history. Akamai issued Anthropic a warrant for 7.7 million shares of common stock, with roughly 2% of shares vesting when the agreement starts. Akamai projects $5.5 billion in capital spending tied to the deal and expects $150 million to $300 million in related revenue next year. The contract can expand to $20 billion under certain conditions. Akamai shares surged as much as 16% Friday on the news.
Sources: Bloomberg · TechCrunch · CNBC · 24/7 Wall St.
Microsoft Overhauls Copilot With Home, Code and Autopilot — Shifts to Usage-Based Pricing
Microsoft unveiled a broad redesign of Copilot on Friday, reorganizing the product around three new tabs: Home, Code and Autopilot. Home combines chat and Cowork and pulls Word, Excel and PowerPoint into a single interface. Code lets non-technical employees build apps, dashboards and workflows in plain language using GitHub Copilot technology. Autopilot — previously called Scout — is a persistent agent designed to monitor projects and act without waiting for a prompt. Microsoft also shifted pricing, moving to flat per-user fees for Chat and usage-based billing for Copilot's more advanced agentic features. Microsoft shares rose more than 4% on the announcement, with Oppenheimer raising its price target to $570.
Sources: VentureBeat · GeekWire · Gurufocus · Softonic
OpenAI CEO Sam Altman Tells UN Security Council AI Solved a Millennium Prize Problem
OpenAI CEO Sam Altman addressed the United Nations Security Council on Friday, using the platform to describe the pace of AI progress. He cited OpenAI's recent solution to the Navier-Stokes existence and smoothness problem — one of the six remaining Millennium Prize Problems — as evidence of AI's capacity to produce knowledge that has never existed before. An internal OpenAI multi-agent system solved the problem in early September after a 88-hour run using more than 10,000 agents. OpenAI has said it does not intend to claim the $1 million Clay Mathematics Institute prize for the result. Altman also called for AI governance to remain in the hands of democratic governments.
Sources: OpenAI · ABC News · BitDigest
AI Devices & Hardware
ChatGPT Voice Gets Plugin Support, GPT-6 Model Choice and Work Integration
OpenAI expanded ChatGPT Voice on Friday with three major upgrades. Voice now supports plugins and connected apps — including email, calendar and Slack — on web, iOS and Android. Users can select from GPT-6 Astra, Sol and Luna as the underlying model for any voice conversation. Voice is also now available inside ChatGPT Work on web and mobile, allowing users to create documents, presentations and spreadsheets by speaking. If a task isn't finished when a call ends, it carries on in text. Free and Go users get plugin access during voice conversations. The changes ship globally as of today.
Sources: OpenAI · TechCrunch · AI Weekly
NVIDIA Releases Open Nemotron 3 Model for Real-Time Speaker Diarization
NVIDIA released Nemotron 3, an open model designed for real-time speaker diarization — the process of identifying and separating individual voices in audio. The model was published on Hugging Face. Speaker diarization is a foundational capability for meeting transcription, call center automation and any agentic system that needs to track who said what in a conversation. NVIDIA's move into open model releases for audio tasks mirrors a broader industry push to build infrastructure-level tools that AI developers can run and customize without cloud dependencies.
Sources: Hugging Face · BitDigest
Microsoft Opens Maryland Quantum Research Center With University of Maryland
Microsoft opened a quantum research center in Maryland in partnership with the University of Maryland. The facility focuses on applied quantum computing and is part of Microsoft's broader push to develop quantum hardware alongside its AI infrastructure. The center will support joint research into quantum error correction and materials science. Microsoft has been working toward a scalable quantum system for years. The Maryland center reflects a growing pattern of academic-industry partnerships as quantum computing moves from theoretical benchmarks toward practical application.
Sources: Microsoft · BitDigest
Enterprise & Cloud
D.C. Circuit Court Upholds Pentagon Blacklist of Anthropic in 2-1 Ruling
A federal appeals court in Washington upheld the Pentagon's designation of Anthropic as a national security supply chain risk on Friday. The 2-1 ruling by the U.S. Court of Appeals for the District of Columbia Circuit rejected Anthropic's argument that the Defense Department's ban on its Claude models was arbitrary and unconstitutional. The conflict began in February when Anthropic refused a Pentagon request to remove restrictions on the use of Claude for fully autonomous lethal weapons and mass domestic surveillance. The DoD formally designated Anthropic a supply chain risk in March — the first time that designation was ever applied to a U.S. company. Anthropic said it "respectfully disagrees" with the decision and is weighing further options, including a Supreme Court appeal. A California court's separate ruling blocking the ban remains in effect for non-Pentagon federal agencies.
Sources: CNBC · Reuters · CNN · Washington Post · France24
Salesforce Unveils AIforce and Headless 360 at Dreamforce — Agents Get Their Own Enterprise Layer
Salesforce used Dreamforce 2026 to announce two major additions to its platform. AIforce is a new AI interface layer that gives agents a way to understand and act on business processes. Headless 360 lets users access Salesforce software without its traditional user interface, enabling AI agents to operate directly against back-end data and workflows. Salesforce described both as part of an Enterprise AI Harness that combines data, business knowledge and controls into a composable architecture. The announcements reinforce a broader trend: enterprise software vendors are rebuilding their stacks around agents rather than human users clicking through screens.
Sources: AI Agents News · Dataiku
Regulation & Policy
Trump and Speaker Johnson Set Sept. 29 AI Meeting With Tech CEOs as Safety Debate Intensifies
President Trump and House Speaker Mike Johnson are set to meet with tech CEOs on Tuesday, September 29, to discuss artificial intelligence. The meeting follows a wave of public warnings from AI lab leaders urging a slowdown in frontier development. Trump has dismissed those warnings as a "hoax" and announced plans last week to create an "AI Force" and appoint a new AI czar. Speaker Johnson told reporters the session will include a discussion of company responsibility for safety and the government's potential role. The list of attendees has not been released. The House is in recess until after the November midterms, limiting the legislative window for any AI regulation to advance.
Sources: ABC News · The Next Web · Reuters
White House Asks OpenAI and Anthropic to Keep New Models From U.K. Testers Until U.S. Review
The White House has asked OpenAI and Anthropic to withhold new frontier models from the U.K. AI Security Institute until the U.S. government has the opportunity to review them first. The request came from the Office of the National Cyber Director. Politico reported that Anthropic appeared to have complied for Claude Mythos 5.1, while OpenAI had not publicly addressed the request as of Friday. The AI Security Institute conducts pre-release safety probes focused on cybersecurity risks and misuse potential. The policy effectively gives the U.S. government first access ahead of Washington's closest foreign partner.
Sources: Politico · The Neuron · BitDigest
Media & Entertainment
Meta Muse Tops App Store Charts — 2.8 Million Downloads in Two Weeks Since Launch
Meta's consumer AI app, Muse, overtook ChatGPT as the top free app on Apple's App Store and Google Play in the U.S. and Canada this week. The app recorded 2.8 million downloads within two weeks of launch, with average daily download growth of 55% in the first ten days. Muse is Meta's direct push into the standalone AI assistant market. Its early traction puts it ahead of where rivals were at comparable points in their rollouts. Meta has been aggressive about shipping AI directly to consumers rather than relying entirely on distribution through its existing social platforms.
Sources: Reuters · FTI Strategic Communications
On the Horizon
OpenAI DevDay Arrives Monday — Creator Tools, Agent APIs and GPT-6 Cyber Preview Expected
OpenAI DevDay lands Monday, September 29. The company's newly formed Creator Product team — led by Patreon co-founder Sam Yam — is expected to present tools for creators and their communities. Developer sessions will cover the broader platform roadmap. Fortune reported that OpenAI is also expected to preview GPT-6 Cyber within days of DevDay, adding a specialized cybersecurity-focused model to the GPT-6 family. The event is the same day as the scheduled White House AI meeting with Trump and tech CEOs.
Sources: The Verge · Fortune · BitDigest
Trump-Johnson-Tech CEO AI Meeting Set for Tuesday — Governance, Safety and "AI Force" Agenda
Tuesday's White House meeting between President Trump, Speaker Johnson and tech industry leaders arrives at a pivotal moment. AI lab CEOs have issued increasingly urgent public warnings about frontier AI risks. Trump has called those concerns overblown and signaled he wants to move faster, not slower, on American AI development. Whether the meeting produces any concrete policy framework, or remains an alignment exercise before the midterms, will set the tone for AI governance heading into 2027.
Sources: ABC News · Reuters · The Next Web