Runway Gen-4.5 is rolling out, and the text-to-video model grabbed the No. 1 spot on the Video Arena leaderboard for generative models that don’t simultaneously output sound, beating the non-audio versions of Google’s Veo 3 and OpenAI’s Sora 2 Pro. Offering what Runway AI calls “unprecedented visual fidelity” across cinematic and “highly realistic” outputs, it is also built for creative freedom, providing “precise control over every aspect of generation.” The new model is good at understanding physics, cause and effect through casual reasoning, camera movements and human emotion, claims the New York-based startup. Continue reading Runway Gen-4.5 Video Debuts at No. 1 on Video Arena Chart

DeepSeek-V3.2 is now in release, integrating thinking directly into tool-use for the first time, improving its predecessor DeepSeek-V3.2 Experimental. The model supports tool-use in both thinking and non-thinking modes. China-based DeepSeek began disrupting the U.S. AI market in January with the debut of foundation models that rival those from Google and OpenAI that are available for free. The company released internal benchmark scores indicating its new model can compete with OpenAI’s GPT-5 in reasoning benchmarks and agentic tasks. A variation, DeepSeek-V3.2-Speciale, has been released for specialized math and is said to perform comparably to Google’s Gemini 3 Pro. Continue reading DeepSeek Debuts Its V3.2 Reasoning Model in Two Versions

Amazon Web Services is marking its entry into the lucrative business of providing the U.S. government with artificial intelligence and high-performance computing services via purpose-built infrastructure for which it has committed up to $50 billion in construction and equipment. The investment “will add nearly 1.3 GW of compute capacity across AWS Top Secret, AWS Secret, and AWS GovCloud (U.S.) across all classification levels,” Amazon says, adding that the commitment “expands access to AWS’s trusted infrastructure” and advances America’s AI leadership and enables U.S. agencies “to accelerate discovery and decision-making.” Continue reading AWS Earmarking $50 Billion for Government AI Infrastructure

Perplexity AI has launched its Comet browser for mobile, taking on Google in the Android ecosystem. Perplexity says it will soon release an iOS app as well. The AI-powered Comet browser debuted in July on a limited basis for Mac and Windows desktops, followed by a wider release in October. The Comet browser uses Perplexity’s AI-powered search engine to answer questions, either typed or posed conversationally, and lets users mention specific tabs or can focus on open tabs. It can also agentically complete tasks. Perplexity hasn’t disclosed Comet usage figures, but says manufacturers are already requesting to include Comet on-device. Continue reading Perplexity’s AI-Powered Comet Browser Now Live on Android

OpenAI is live with a retail companion it calls “shopping research” that helps consumers find viable purchase options based on detailed requirement parameters. Trained on a version of GPT‑5 mini, it can query on a virtually unlimited list of discovery constraints, sourcing only from “reliable retailers.” “Results are organic and based on publicly available retail sites,” OpenAI says, indicating there are no advertiser-boosted entries — not yet, anyway. Available on iOS, Android and the Web for logged-in ChatGPT users on the Free, Go, Plus, and Pro plans, OpenAI is making nearly unlimited usage available to all plans through the holidays. Continue reading OpenAI Continues Push into Retail with ‘Shopping Research’

TikTok is testing a new tool that aims to help users determine how much AI-generated content they see in their For You feeds. The ByteDance-owned company becomes the first of the giant social media platforms to try putting that control in user hands amidst growing animosity toward “AI slop.” Pinterest added a GenAI user filter in October, while Spotify and Jack Dorsey’s Vine relaunch diVine have implemented system-level filters. TikTok’s AI controls fall within its Manage Topics tool. The platform is also adding more advanced AI labeling technology. Continue reading TikTok Testing User-Controlled Feed Filter for GenAI Content

Nvidia reported stellar results for the recent quarter, logging record revenue of $57 billion, up 62 percent year-over-year and 22 percent from Q2. “Blackwell sales are off the charts, and cloud GPUs are sold out,” Nvidia founder and CEO Jensen Huang told investors, saying “we’ve entered the virtuous cycle of AI,” with “more new foundation model makers, more AI startups across more industries, and in more countries.” The results quieted fears about an AI bubble. But there was some drama as Nvidia disclosed there is “no guarantee” of finalizing a previously announced $100 billion investment in OpenAI. Continue reading Nvidia Reports Record $57B in Revenue, $32B in Profit for Q3

Determined not to be left behind in the race to own the future of AI-assisted online shopping — which has everyone from Walmart, Amazon and OpenAI to TikTok scrambling for digital shelf-space — Google is adding conversational shopping in Google Search. The new shopping features, accessible in AI Mode powered by Gemini, include agentic checkout and the ability to “Let Google Call” local stores to ask about product availability. “Say goodbye to endless filter clicks and picking perfect keywords” and hello to chatting with your search agent like you’re shopping with a friend, Google says, “just in time for this holiday season.” Continue reading Google AI Shopping Adds Chat Search and Agentic Checkout

Google announced the availability of Gemini 3 for select subscribers, with a broader release planned for the coming weeks. The tech giant’s latest AI model will be available via the Gemini app, AI Mode and AI Overviews search tools, and Google’s enterprise products. The model arrives nearly eight months after the launch of Gemini 2.5 as the company looks to compete with OpenAI’s ChatGPT. According to CEO Sundar Pichai, Gemini 3 is the company’s “most intelligent model, that combines all of Gemini’s capabilities together so you can bring any idea to life.” Pichai also announced the Google Antigravity agent platform, that enables developers to code “at a higher, task-oriented level.” Continue reading Google Making AI Search and Coding Smarter with Gemini 3

Google is investing $40 billion to build new data centers in Texas through 2027. Google describes the project as focused on “new cloud and AI infrastructure, including new data center campuses in Armstrong and Haskell Counties,” emphasizing their energy efficiency. The investment includes the creation of a $30 million Google-backed Energy Impact Fund, and one of the Haskell County facilities will be built housed adjacent to a new solar and battery storage plant. At a Dallas-area launch event announcing the project, Alphabet and Google CEO Sundar Pichai said the initiative will create thousands of new jobs. Continue reading Google Commits $40 Billion for New AI Data Centers in Texas

OpenAI is piloting a new group chat experience designed for socialization and work collaboration with ChatGPT, taking aim at Slack and Microsoft Teams. “Whether you’re organizing a group dinner or drafting an outline with coworkers,” OpenAI wants to make ChatGPT the place you do it. Group chats are beginning to roll out on mobile and web for logged-in users on ChatGPT Free, Go, Plus and Pro plans in Japan, New Zealand, South Korea and Taiwan. OpenAI says the pilot is “a small first step toward shared experiences in ChatGPT.” The company plans to collect early user feedback to inform further development. Continue reading OpenAI Pilots ChatGPT Group Chat in Select Foreign Markets

San Francisco-based AI startup Anthropic, founded in 2021, announced plans to invest $50 billion in U.S. computing infrastructure, starting with the creation of data centers with Fluidstack in Texas and New York (with additional plans for future sites). Fluidstack builds and operates high-performance GPU clusters for researchers and AI teams. The new project with Anthropic is expected to create 800 permanent jobs and about 2,400 construction jobs. Sites are scheduled to come online throughout next year. “These facilities are custom built for Anthropic with a focus on maximizing efficiency for our workloads, enabling continued research and development at the frontier,” according to Anthropic. Continue reading Anthropic Plans to Invest $50 Billion in AI Data Centers in U.S.

Microsoft’s new Azure AI data center in Atlanta, Georgia — the second in its Fairwater series — is ushering in a new era of connected AI super clusters that will power what The Register calls “the 100 trillion-parameter models of the near future.” Fairwater Atlanta, which became operative in October, is part of “a new class of data center – one that doesn’t stand alone but joins a dedicated network of sites functioning as an AI superfactory to accelerate AI,” training models “on a scale that has previously been impossible,” according to Microsoft. The new purpose-built data center is connected to the tech giant’s first Fairwater site, located in Mount Pleasant, Wisconsin. Continue reading Microsoft Introduces New Era of Connected AI Super Clusters

Microsoft and Google are making big investments in European AI infrastructure, with the Seattle-based software giant spending more than $10 billion to build a data hub in Portugal, and Alphabet investing $6.35 billion on new construction and expansion of an existing complex in Germany. The moves are the latest in U.S. tech’s push for a global AI footprint. Microsoft is working with Nvidia, the UK startup Nscale and data center building specialist Start Campus. Google is bringing green initiatives to its project in Deutschland, which it estimates will bring 9,000 jobs and nearly $1.2 billion to German GDP. Continue reading Microsoft, Google Spending Heavily on European Data Hubs

SoftBank has cashed out of Nvidia, liquidating its stock holdings for $5.83 billion that it will reportedly use to fund what CNBC describes as “its ‘all in’ bet on ChatGPT maker OpenAI.” The sell-off was disclosed in SoftBank’s most recent earnings statement, which indicated 32.1 million Nvidia shares were sold in October. The Japanese conglomerate also said it sold part of its T-Mobile holdings for $9.17 billion. “Through those options and tools we make sure that we are ready for funding in a very safe manner,” SoftBank CFO Yoshimitsu Goto said at an investor presentation. SoftBank announced a four-to-one stock split yesterday, while Nvidia shares were down 1.7 percent on the sell-off news. Continue reading SoftBank Sells Off Nvidia Shares to Pursue OpenAI, Robotics