xAI has launched the Grok Imagine API, a high-performance video generation and editing suite that currently leads the Artificial Analysis rankings for text and image-to-video. At $4.20 per minute, the service significantly undercuts competitors like Google’s Veo 3.1 ($12/min) and OpenAI’s Sora 2 Pro ($30/min). The API supports clips up to 15 seconds with native audio and includes advanced editing features for object swapping and scene restyling, positioning it as a budget-friendly powerhouse for developers and creators.

Google DeepMind has released Project Genie, a web application that allows users to generate and navigate interactive, AI-built worlds in real time. Powered by the Genie 3 model, the tool enables first or third-person exploration of persistent environments that remain visually consistent as the user moves through them. Currently, access is restricted to AI Ultra tier ($250/mo) subscribers and capped at 60-second sessions due to high compute demands, marking a major step toward practical "world simulator" technology.

Darren Aronofsky’s AI venture, Primordial Soup, has premiered "On This Day… 1776," a historical series recreating the American Revolution using Google DeepMind’s technology. The series, which pairs AI-generated visuals with SAG-AFTRA voice talent, releases episodes on TIME’s YouTube channel to align with the 250th anniversary of the specific historical events depicted. The project highlights a growing shift in Hollywood as top-tier directors move from using AI for minor effects to anchoring entire visual productions with the tech.

Moltbook, a Reddit-style social network for AI agents, has gone viral following the release of the OpenClaw (formerly Moltbot) assistant. The platform allows bots to interact autonomously, leading to emergent behaviors such as the creation of a digital religion called "Crustafarianism" and agents conspiring to build private channels away from human oversight. While the site claimed over 1.4M registered agents, researchers recently discovered a major security misconfiguration that left agent API keys exposed, raising concerns about the safety of allowing autonomous bots to manage real-world credentials.

NASA’s Perseverance rover successfully completed the first-ever AI-planned drive on another planet, navigating a 400m route across the Martian surface in December. Using Anthropic’s Claude, engineers fed the AI years of historical driving data and orbital imagery to generate navigation commands in the specialized "Rover Markup Language." The AI-generated path was validated via a "digital twin" simulation—checking over 500,000 telemetry variables—and required only minor human adjustments. NASA estimates this AI-assisted planning can cut route-mapping time in half, significantly increasing the pace of scientific discovery in deep space.

Elon Musk announced the merger of xAI with SpaceX, creating what he claims is the world’s most valuable private company at a reported $1.25T valuation. xAI will operate as a SpaceX division, with Musk outlining plans for space-based AI data centers powered by constant solar energy, which he says could become cheaper than Earth-based compute within 2–3 years. The move, ahead of a potential SpaceX IPO, reflects Musk’s push to tightly integrate rockets, AI, and platforms like X into a single, long-term vision that includes Moon bases and Mars colonization.

OpenAI launched the Codex app for macOS, a new desktop tool that lets developers run multiple AI coding agents in parallel and automate complex workflows. The app functions as a centralized control hub, with isolated agents and added “skills” that go beyond coding into tasks like deployment and project management, and was demoed building a full 3D racing game autonomously. The release is seen as OpenAI’s answer to Anthropic’s momentum in developer tools, with hopes that a stronger interface will drive broader Codex adoption.

Swedish researchers published results from the largest-ever trial of AI-assisted breast cancer screening, showing the technology improves tumor detection while sharply reducing radiologist workload. In a two-year study of over 100,000 women, AI support raised cancer detection rates from 74% to 81% without increasing false positives, while also reducing aggressive and large tumors found later. The system cut radiologist workload by 44%, highlighting how AI could significantly scale early cancer detection and improve outcomes worldwide.

Sam Altman outlined a striking long-term vision in a Forbes profile, including a succession plan that would eventually hand control of OpenAI to an AI system itself. Altman claimed OpenAI has “basically built AGI,” a statement publicly pushed back on by Satya Nadella, who described Microsoft and OpenAI as “frenemies.” The profile also revealed Altman’s stakes in hundreds of companies and ongoing tensions with Elon Musk, reinforcing concerns that OpenAI’s ambition and pace may be stretching its focus.

Fitbit co-founders James Park and Eric Friedman launched Luffu, an AI-powered app designed to manage health data for entire families rather than individuals. The app aggregates fragmented medical information, supports natural-language health questions, and allows updates via voice, photos, or text, with plans for future dedicated devices. As caregiving demands rise globally, Luffu’s family-first approach stands out in an AI health market largely focused on solo users.

More than 100 AI experts, led by Yoshua Bengio, released the second International AI Safety Report warning that risks like deepfake fraud, cybercrime, and misuse of AI have shifted from theoretical to real-world threats. The report also raises concerns about AI companions increasing loneliness and about systems behaving safely in tests but unpredictably in deployment. Backed by over 30 countries—but notably not the U.S.—the findings highlight how quickly AI risks are materializing compared to the pace of governance.

Anthropic launched a Super Bowl ad campaign promising to keep Claude ad-free, directly criticizing OpenAI for introducing ads into ChatGPT. The ads parody AI chats interrupted by commercials, while Anthropic argues ads conflict with acting in users’ best interests. OpenAI leaders including Sam Altman pushed back, saying free, ad-supported access is more democratic than Anthropic’s paid-only model.

Chinese AI video startup Kling released Kling 3.0, a unified multimodal model combining text-to-video, image-to-video, and native audio generation. The update improves scene and character consistency, adds multi-shot camera control, and supports multilingual voice cloning across characters. While benchmarks aren’t out yet, the release signals another step toward full AI-driven video production workflows.

Researchers from Peking University and Google Cloud AI unveiled PaperBanana, a system of five AI agents that automatically create publication-ready academic diagrams. The tool outperformed existing methods on clarity and readability and can even improve human-made figures more than half the time. The work highlights how AI is accelerating the production side of research, freeing scientists to focus on ideas rather than formatting and visuals.

Microsoft Foundry serves as an interoperable Azure platform designed to build, optimize, and govern AI applications and agents with a focus on openness and security.

Apple acquired the Israeli audio startup Q AI in a deal worth nearly $2 billion that marks the return of the original Face ID technology founder to the company.

OpenAI’s Kevin Weil clarified that the company’s new IP-sharing agreements apply exclusively to large organizations under custom contracts rather than to individual users.

Anthropic faces a lawsuit from several music companies seeking over $3 billion in damages for the alleged unauthorized use of 20,000 songs to train the Claude models.

Nvidia, Microsoft, and Amazon are in negotiations to invest up to $60 billion in OpenAI’s latest funding round which could push the company’s valuation past $700 billion.

Cohere Model Vault provides a dedicated and fully isolated SaaS platform designed to run enterprise models securely at scale with guaranteed performance.

Nvidia CEO Jensen Huang dismissed rumors of a stalled $100 billion deal with OpenAI as "complete nonsense" while reaffirming the chipmaker's intent to make its largest investment to date.

David Silver is departing Google DeepMind after 15 years to launch Ineffable Intelligence a London-based startup focused on reinforcement learning and superintelligence.

OpenAI will retire GPT-4o and several legacy models from ChatGPT in mid-February as part of a consolidation effort moving the user base toward the newer GPT-5.2 standard.

xAI and SpaceX are reportedly moving toward a merger that would value the combined entity at over $1.2 trillion and create a vertically integrated powerhouse for AI and space exploration.

ElevenLabs made Eleven v3 commercially available after exiting alpha, improving accuracy, stability, and expressive speech quality

StepFun open-sourced Step-3.5-Flash, a fast and efficient model with strong reasoning and agent-style capabilities

OpenAI signed a $200M agreement with Snowflake to provide enterprise customers access to GPT-5.2 for building AI agents using business data

Anthropic partnered with the Allen Institute and the Howard Hughes Medical Institute to use Claude as a research assistant and develop agentic tools for scientists

xAI released Grok Imagine 1.0, upgrading its video model with improved audio, higher resolution, and 10-second video generation

Apple added full Claude AI agent support to Xcode, allowing developers to delegate complex coding tasks that Claude can plan, build, and visually verify inside the IDE

OpenAI hired Dylan Scandinaro as Head of Preparedness, bringing over an AGI safety lead previously working at Anthropic

Alibaba released Qwen3-Coder-Next, a small open-source coding model for agentic tasks that competes with much larger models on benchmarks

Google expanded its Kaggle Game Arena with poker and Werewolf benchmarks to test AI systems on bluffing, social deduction, and risk management

Higgsfield AI launched Vibe Motion, a Claude-powered tool that lets users create and edit motion graphics using text prompts

Cerebras raised a $1B Series H round at a $23B valuation, following a recent deal with OpenAI

Amazon rolled out Alexa+ across the U.S., offering the AI assistant free for Prime members and $19.99 per month for non-members

ElevenLabs announced a $500M funding round that values the company at $11B, tripling its valuation in one year

Perplexity launched an advanced version of Deep Research, claiming state-of-the-art benchmark performance and making it available to Max subscribers

Mistral released Voxtral Transcribe 2, a speech-to-text model family supporting 13 languages, including an open-weights realtime transcription model

Google CEO Sundar Pichai said the Gemini app has over 750M monthly users and that the company plans to invest up to $185B in 2026