xAI Acquires Hotshot: Elon Musk’s AI Firm Joins the Generative Video Race
Nvidia’s Breakthroughs at GTC 2025: Next-Gen Chips, Grid AI, Robotics Innovation, and More
Google Launches AI Health Tools: Search Expansions, Medical APIs, and Drug Discovery Models
Plus: 3 Additional Resources to Explore
xAI Acquires Hotshot: Elon Musk’s AI Firm Joins the Generative Video Race
Elon Musk’s xAI has acquired Hotshot, a San Francisco-based startup specializing in generative AI video models. Known for building tools like Hotshot-XL and Hotshot Act One, the small team has developed multiple text-to-video foundation models that rival OpenAI’s Sora and Google’s Veo.
The move signals xAI’s ambitions to enter the AI video generation space, with plans to integrate such capabilities into its Grok platform. Hotshot CEO Aakash Sastry confirmed the acquisition on X, noting the team will scale its work using xAI’s massive compute cluster, Colossus.
Hotshot will shut down its current product by March 30. While terms of the deal remain undisclosed, the acquisition adds generative video expertise to xAI’s growing suite of AI capabilities.
Tech Crunch Article: Elon Musk’s AI company, xAI, acquires a generative AI video startup
Nvidia’s Breakthroughs at GTC 2025: Next-Gen Chips, Grid AI, Robotics Innovation, and More
At GTC 2025, Nvidia unveiled a wave of hardware and AI advancements, headlined by the launch of its new flagship Blackwell Ultra GPU series and the preview of its next-gen Rubin architecture, set to power future AI workloads. CEO Jensen Huang also introduced two “personal AI supercomputers”, aiming to bring powerful, localized AI capabilities to individual users for the first time.
In infrastructure, Nvidia announced the formation of the Open Power AI Consortium in partnership with EPRI and major utilities to address the growing strain AI places on the electrical grid. The consortium will develop and open source domain-specific AI models to help optimize energy distribution as electricity demand is projected to rise sharply.
In quantum computing, Nvidia is pivoting toward long-term investment. Following previous comments that caused quantum stocks to tumble, Huang launched Quantum Day and announced a new research center in Boston focused on quantum architectures and algorithms, reaffirming Nvidia’s interest in the space despite skepticism about its near-term viability.
In robotics, Nvidia revealed Newton, a real-world physics simulation engine co-developed with Google DeepMind and Disney Research to enable lifelike robotic motion. Disney will use Newton in its upcoming BDX entertainment droids, one of which made an onstage debut during Huang’s keynote.
Nvidia also confirmed a nine-figure acquisition of synthetic data startup Gretel, expanded partnerships in autonomous driving with GM and startups like Gatik, and backed a stealth robotics company, Generalist AI, founded by a former DeepMind researcher.
Learn More:
Nvidia Newsroom:Nvidia to Build Accelerated Quantum Computing Research Center
Nvidia Blog:EPRI, Nvidia and Collaborators Launch Open Power AI Consortium to Transform the Future of Energy
Nvidia Newsroom:Nvidia Announces DGX Spark and DGX Station Personal AI Computers
Nvidia Developer:Announcing Newton, an Open-Source Physics Engine for Robotics Simulation
Tech Crunch Article:Nvidia GTC 2025 live updates: Blackwell Ultra, GM partnerships, and two ‘personal AI supercomputers’
Google Launches AI Health Tools: Search Expansions, Medical APIs, and Drug Discovery Models
Google has announced a wave of healthcare-focused updates across Search, Android, and AI, aiming to deepen its role in digital health. On Search, the company has enhanced knowledge panels to include thousands more health topics, while also expanding support to Spanish, Portuguese, and Japanese on mobile. A new feature, “What People Suggest,” uses AI to surface user-generated insights from forums to complement expert content, keeping users within the Google ecosystem rather than turning to Reddit or similar sources.
On Android, Google launched global medical records APIs for its Health Connect platform. These allow apps to read and write data like allergies, medications, and lab results in FHIR format, integrating doctor-provided information with everyday health metrics across over 50 data types.
Google also revealed that its Pixel Watch 3 will gain FDA-cleared Loss of Pulse Detection by the end of March, automatically contacting emergency services if a user becomes unresponsive.
Finally, Google is entering the AI-driven drug discovery race with TxGemma, a new open model series set to launch soon, following its Gemini-based healthcare AI initiatives. These announcements reflect Google’s broader strategy to embed AI into the healthcare value chain, from consumer search to clinical-grade applications.
Learn More:
Google Blog:6 health AI updates we shared at The Check Up
TechCrunch Article:Google launches new healthcare-related features for Search, Android
Beyond The Key Updates: More To Explore
Claude Adds Web Search for Real-Time, Source-Cited Answers Anthropic Article | Read time: 1 min
Claude can now search the web, providing users with real-time insights and source-cited answers. This feature boosts the accuracy of Claude's responses for tasks that benefit from the latest data, such as sales outreach, financial modeling, academic research, and product comparisons. Available in feature preview for paid U.S. users, the update enables Claude 3.7 Sonnet to fetch and integrate current web content into conversations. Broader rollout is expected soon.
Meta’s Llama Models Surpass 1 Billion Downloads as Open AI Push Continues TechCrunch Article | Read time: 3 min
Meta CEO Mark Zuckerberg announced that Llama, Meta’s family of open AI models, has reached 1 billion downloads—up 53% since December 2024. Llama powers Meta AI across Facebook, Instagram, and WhatsApp, and is available under a free but commercially limited license. Despite legal and regulatory setbacks, companies like Spotify, AT&T, and DoorDash are using Llama in production. Meta plans to release new “reasoning” and multimodal models this year, with agent-like capabilities. Zuckerberg aims for Llama to become the leading and most widely used open-source AI model by the end of 2025, with more to be revealed at Meta’s upcoming LlamaCon developer conference.
Elon Musk’s AI venture, xAI, has introduced image generation to its API with the new ‘grok-2-image-1212’ model, offering up to 10 images per request at $0.07 per image. While lacking features like image quality or style customization, it undercuts competitors like Ideogram ($0.08/image) and rivals Black Forest Labs ($0.05/image). The launch follows reports that xAI is seeking a $10B funding round, which could push its valuation to $75B. The company has also acquired a generative AI video startup and is expanding its Memphis-based data center, signalling deeper moves into multimedia AI services.
Thanks for reading The AI Weekly Briefing! Subscribe for free to receive new posts and support my work.
