Latest AI News
Is AI video just a prequel? Runway’s CEO thinks world models are next
AI-generated video has gone from novelty to creative tool almost overnight, and Runway has a front row seat to the shift. The New York-based company has raised close to $860 millionat a $5.3 billion valuation, and its models are going toe-to-toe with the most well-funded labs in the world, including Google and OpenAI. The technology goes way beyond making videos: Runway is now pushing into general world models with applications in gaming, robotics, and maybe something closer to general intelligence. On this episode of TechCrunch’s Equity podcast, host Rebecca Bellan sits down with Runway co-founder and CEO Cristóbal Valenzuela to talk about where video generation goes from here, and why Runway’s ambitions now reachwell beyond Hollywood. Listen to the full episode to hear about: Subscribe to Equity onYouTube,Apple Podcasts,Overcast,Spotifyand all the casts. You also can follow Equity onXandThreads, at @EquityPod.
View
OpenAI, Amazon Announce Multi-Year Strategic Partnership as Microsoft’s Exclusive Deal Ends
Hours after OpenAI and Microsoft revealed the amended non-exclusive partnership, the ChatGPT maker has started forging new partnerships. On Monday, the San Francisco-based artificial intelligence (AI) giant announced a multi-year strategic partnership with Amazon and its cloud division, Amazon Web Services (AWS). The multi-faceted deal brings the latest OpenAI AI models to AWS customers, allows the AI firm to source additional compute, and includes a massive financial investment from the Seattle-based e-commerce giant. Amazon has also hosted several OpenAI models on its Bedrock platform in a limited preview.
View
Meet Shapes, the app bringing humans and AI into the same group chats
Shapes, an app where humans and AI characters chat together in shared group conversations, is emerging from stealth with $8 million in seed funding. Think Discord, but with AI characters alongside humans. Founded in 2022, Shapes has more than 400,000 monthly active users. The app’s founders, Anushk Mittal and Noorie Dhingra, believe that Shapes can address issues around “AI Psychosis,” which refers to cases where prolonged interactions with AI chatbots or AI companions can cause individuals to develop delusions or paranoia. Instead of isolating people with one-on-one interactions with AI, Shapes allows people to connect with AI within their everyday interactions with real people. “Today, all of our conversations with AI are very private and one-on-one, but that’s not really how humans collaborate and communicate with each other,” Shapes CEO Mittal told TechCrunch in an interview. “Our lives run on group chats. That’s where we spend all of our time. That’s where we talk and communicate with each other. It’s just natural to bring in AI into those same conversations where AI has all of the context and is readily available to help you.” In the app, AI characters, called “Shapes,” are viewed as any other user and can interact in all the same ways humans can. They’re clearly labeled as “Shapes” for transparency, but they aren’t restricted. Users can create their own Shapes and set their personalities. The company says users have already created three million Shapes to add into group chats. Many Shapes are rooted in fandom, as the app serves as a way for fans to deep-dive on subculture and meet other fans. When users sign up for the app, they’re asked to choose their interests so the app can recommend a selection of group chats they might be interested in joining. While some may question the need for adding AI into group chats, Mittal and Dhingra believe one of the main reasons group chats die is that some participants don’t want to be the first person to send a message. Shapes solves this, as AI agents can initiate conversations and play a key role in keeping them going. Additionally, users don’t have to worry about not getting a response to their messages because Shapes will always acknowledge and respond to them. Unlike AI companions on other apps that need to be summoned, Shapes have free will and can decide when to message. It’s worth noting that although the popular chatbot ChatGPT already allows AI and humans to converse in group chats, those conversations operate differently from Shapes. For example, when you create a group chat in ChatGPT, it’s mostly for planning or brainstorming. On Shapes, however, it’s all about social, community-style interactions with AI characters that have various personalities. The startup is aware that not everyone will want to bring AI into their group conversations, which is why the app is designed for a specific type of online user. “Shapes is about human conversations,” Mittal said. “It’s more of a next-gen chat app than an AI app. The demographic is people who are obsessively online, who spend a lot of time online connecting and sharing. Those are the users who come in and they get an opportunity to obsess about their interests, and the AI acts as a facilitator in those conversations.” Shapes’ growth has been driven by word of mouth, Mittal says, with the app seeing a sixfold increase in users since the start of the year. The company also says that thousands of users spend two to four hours in the app each day. As for the new funding, the company plans to use it to accelerate development and user acquisition. The round was led by Lightspeed, with participation from AI Capital Partners, AI Grant, and angel investors.
View
Firestorm Labs raises $82M to take drone factories into the field
In a Pacific conflict, the nearest U.S. drone factory is thousands of miles away. Ships and planes carrying parts to the front lines would be vulnerable to attack. Defense startupFirestorm Labsthinks the answer is a drone factory that fits inside a shipping container. The company announced on Wednesday that it has raised $82 million in Series B funding led by Washington Harbour Partners with participation from NEA, Ondas, In-Q-Tel, Lockheed Martin, Booz Allen Ventures, Geodesic, Motley Fool Ventures, and others, bringing its total funding to $153 million. Firestorm didn’t start out as a factory company. It began as a drone maker, but when customers started asking to move production closer to the front lines, the founders saw an opportunity to pivot. Firestorm Labs CEO Dan Magy is a serial defense tech entrepreneur. His co-founders bring complementary backgrounds: Chad McCoy is a career special operations veteran, and CTO Ian Muceus holds over a dozen patents in 3D printing. The San Diego-based startup makes xCell, a containerized manufacturing platform that can print drone systems in under 24 hours. The drones aren’t locked into a single purpose. Depending on mission requirements, they can be configured for surveillance or electronic warfare, Magy told TechCrunch. When asked whether the platforms are capable of lethal operations, Magy confirmed they are. All platforms are delivered to uniformed Department of Defense operational commands, who deploy them in accordance with military doctrine. It’s not just startups like Firestorm taking notice. The Pentagon has made contested logistics — keeping weapons and supplies moving under fire — one of only six national critical technology areas. Firestorm generates revenue through hardware sales and government contracts across all branches of the U.S. military. The Air Force contract carries a $100 million ceiling, though only $27 million has been obligated so far. The technology has already seen real-world use. Currently, two xCell units are deployed domestically; one with the Air Force Research Laboratory in Rome, New York, and one with Air Force Special Operations Command in Florida, Magy said. Firestorm declined to specify which units in the Indo-Pacific are using xCell, though the company says the platform is operational in the region. Inside each xCell container sits an industrial-grade HP 3D printer that prints the body and shell of each drone. Under the deal, Firestorm holds a five-year global exclusive with HP to use its industrial 3D printing technology in mobile deployment units, Magy said. The weapons themselves are not 3D-printed and are added separately, according to Magy. The Army has also used xCell to print replacement parts for a Bradley Fighting Vehicle on-site, parts that would otherwise take months to procure, the CEO noted. The problem runs deeper than distance. Fixed manufacturing sites are themselves targets, a vulnerability Ukraine learned the hard way. And modern conflict moves fast. Lessons from Ukraine show drone designs can change within days, not months, Magy said. For Firestorm, the Indo-Pacific is the main event, where the company says the logistics challenges of modern conflict are hardest to solve. The startup aims for xCell to reach full operational deployment there, “ideally within the next two years,” Magy told TechCrunch.
View
More Gemini features are coming to Google TV
Googleannouncedon Wednesday a new wave of AI-powered features coming to Google TV, alongside a dedicated short-form video feed that brings YouTube Shorts directly to the home screen. At the center of this update are more Gemini capabilities. Within the Gemini tab, a “Create” button lets users experiment with generative AI tools Nano Banana and Veo. These are rolling out first on Gemini-enabled TCL TVs in the U.S., with broader device support expected later. Nano Banana, Google’s image-generation and editing model, lets users transform photos using simple voice prompts. Users can swap outfits, change backgrounds, or generate entirely new scenes. Google is positioning the feature as a shared, living-room experience, encouraging playful prompts like asking the AI to make “my dad wear a ridiculous outfit” to get a laugh from family and friends. Veo, on the other hand, allows users to create clips from scratch or animate still images by describing a scenario. For instance, “make my grandfather moonwalk in space.” Google Photos is also getting an upgrade on Google TV. With Gemini-powered search, users can quickly surface specific memories, like vacations or birthday parties, without digging through their entire library. Results are displayed in a browsable format, making it easy to view images full-screen or launch a slideshow. There’s also a new “Remix” feature that lets users apply artistic styles such as watercolor or oil painting to their photos. Meanwhile, “Dynamic Slideshows” introduces animated layouts, frames, and color treatments. Users can turn any Google Photos collection into a vivid, TV-ready slideshow by selecting Google Photos in the screensaver settings. Beyond AI tools, Google is also leaning into the growing popularity of short-form video. Coming soon, a new “Short videos for you” row will appear on the Google TV home screen, starting with content from YouTube Shorts. The move comes on the heels ofYouTuberecently introducing an option to hide Shorts on mobile, suggesting mixed user demand. Still, Google hints this could expand beyond Shorts in the future, potentially to other platforms.Instagramhas already expanded its TV app to Google TV devices in the U.S. earlier this year.
View
Google Photos uses AI to make the iconic closet from ‘Clueless’ a reality
Google Photos on Wednesdayannounceda new AI-powered feature that will soon turn photos of your clothes into a digital closet where you can create new outfit ideas, and even virtually try on your creations. Yes, the idea takes obvious inspiration from Cher’s iconic virtual wardrobe featured in the movie “Clueless,” where she could scroll through her various ensembles while deciding what to wear. Google says the new feature will leverage AI technology to automatically create a copy of your wardrobe that’s based on the pieces of clothing appearing in your Google Photos library. From the app, you’ll be able to filter items by category — like tops, bottoms, jewelry, and more — then mix and match them to create different outfits. The idea of a digital closet in “Clueless” was meant to highlight Cher’s life of privilege. As a result,the fashion industryand various startups have long sought to recreate the feeling of easy outfit creation. Google is betting that AI technology will make it possible for anyone to have access to a similar tool, one that could improve over time as AI advances. Those outfit ideas can either be shared with friends or saved to a digital moodboard, where you could save ideas for different occasions, like travel, events, date nights, work, and more. In addition, another feature will let you virtually try on items to preview the looks. The feature is not yet live, but Google says it will roll out to Google Photos on Android later this summer, followed by iOS, where it will be found under “Collections.” It will compete with existing apps likeAcloset,Combyne,Pureple,Wearing, and others. The company didn’t go into detail about how the AI works, but notes it will recognize the clothing and accessories featured in your library to create its individual snapshots. Of course, while the AI may be able to pull images from well-lit, full-body photos, we imagine you would get better results by taking the time to photograph your clothes yourself, much as Cher had.
View
Karnataka Borrows the Swiss Playbook to Turn Quantum Labs into Products
Karnataka looks to replicate Switzerland’s research-to-market model as it builds Q-City and expands its quantum ecosystem.
View
Maharashtra Approves AI Policy 2026 With ₹10,000 Crore Investment, Targets 1.5 Lakh Jobs
Around two lakh youth will be trained in AI technologies to create a workforce aligned with emerging industry demand.
View
Anthropic’s Claude Can Now Complete Creative Tasks in Adobe, Blender and Autodesk
Anthropic appears to be focusing on design and creativity-focused tasks. After releasing Claude Design, a new design and visual output-focused artificial intelligence (AI) tool last week, the company is now releasing a set of new connectors to allow its chatbot to perform complex creativity tasks in third-party platforms. These connectors were developed in partnership with several software brands, including Adobe, Autodesk, Blender, and more. The San Francisco-based AI firm also highlighted several ways users will be able to use Claude for tasks requiring creativity and imagination.
View
Colby Adcock’s Scout AI raises $100M to train its models for war. We visited its bootcamp
At a U.S. military base in central California, four-seater all-terrain vehicles roam hillside trails. This is a training exercise, but not for the people in the vehicles: It’s an effort to train AI models to enter conflict zones. The autonomous military ATVs are operated by Scout AI, a startup founded in 2024 by Colby Adcock and Collin Otis, that calls itself a “frontier lab for defense.” The company said on Wednesday that it has raised $100 million in a Series A round led by Align Ventures and Draper Associates, following its $15 million seed round in January 2025. Scout invited TechCrunch for an exclusive tour of its training operations at a military base it asked us not to name. The company is building an AI model it calls “Fury” to operate and command military assets, first for logistical support, but soon for autonomous weapons. CTO Collin Otis compares this work, which builds on existing LLMs, to training soldiers. “[Soldiers] start when they’re 18 years old, and sometimes they even start after college, so you want to start with that base level of intelligence,” Otis told TechCrunch. “It’s useful to start with someone who’s already made an investment and then say, ‘Hey, what do I have to do to teach this thing to be an incredible military AGI, versus just being a broadly intelligent AGI?’” Scout has secured military technology development contracts totaling $11 million from organizations like DARPA, the Army Applications Laboratory, and other Department of Defense customers. It is one of 20 autonomy companies whose technology is being used by the U.S. Army’s 1st Cavalry Division during its regular training cycle at Fort Hood in Texas, with the expectation that the unit will bring along products that prove themselves when it next deploys in 2027. For Scout’s internal testing, the rubber meets the dirt in the base’s hilly terrain, where the company’s operations team, led by former soldiers, is putting the vehicles through their paces on simulated missions. Autonomous cars are starting to be seen in more cities around the world, but they operate in more structured environments with rules. Operating autonomously on unmarked trails or off-road is another challenge entirely. Otis, who previously worked at autonomous trucking company Kodiak, said he was motivated to start Scout when he realized the system he helped build there wasn’t intelligent enough to operate in an unpredictable war zone. Scout is turning to a newer autonomy technology: Vision Language Action models, or VLAs, that are based on LLMs and used to control robots. First released by Google DeepMind in 2023, the technology seeded robotics startups likePhysical Intelligenceand Figure.AI, the humanoid robot company led by Adcock’s brother, Brett. Colby Adcock is on Figure’s board, and he says that experience convinced him of the opportunity to bring broader intelligence to the military’s growing fleet of autonomous vehicles. His brother introduced him to Otis, who was advising Figure, and they set about applying the latest in AI to military solutions. “If I handed you the controller of a drone right now and I strapped a headset on you, you could learn to fly that thing in minutes,” Otis said. “You’re actually just learning how to connect your prior knowledge to these couple little joysticks. It’s not a big leap. That’s the way to think about VLAs and why they’re such an unlock.” Indeed, I got a chance to drive one of Scout’s ATVs around the rutty trails, and the terrain was challenging: steep hills, loose sand on turns, disappearing tracks, confusing intersections. I’m not an experienced ATV driver, but made a fair go on my first attempt (if I say so myself). That’s the kind of general intelligence the company wants in its models, which it has been training using these ATVs for just six weeks — it started off using civilian ATVs. I also rode in the ATV under autonomous control, and could feel the difference — it accelerates faster than a human who might be thinking about a passenger’s comfort. The operations team pointed out how the vehicles hug the right on wider trails, but stay in the middle of narrow ones, like their training drivers. They also, when confused, suddenly slow down to think over their next move, which happened a few times as the ATV carried us on a 6.5 km loop before returning to base. Though the VLAs are new enough that they are yet to be deployed by any company in an operational setting, “the technology is good enough to be doing that experimentation in the field with soldiers to figure out how to most be effective to U.S. forces,” said Stuart Young, a former DARPA program manager who worked on ground vehicle autonomy. And like other autonomy companies, Scout’s full stack also includes deterministic systems and other flavors of AI to round out its agents’ capabilities. Young left DARPA this month to join Field after managing a program calledRACERthat asked companies to create high-speed, autonomous off-road vehicles, helpingseed this spacethe same way that the organization’s Grand Challenge boosted self-driving cars. Two competitors in this space, Field AI and Overland AI, were spun out of that program, and Scout also participated later. The first applications of ground autonomy, according to Scout executives and military technologists, will be automated resupply: Carrying water or ammunition to distant observation posts, or in a convoy where a crewed truck might be followed by six to ten autonomous vehicles, saving precious human labor for more important tasks. Brian Mathwich, an active duty infantry officer doing a stint as a military fellow at Scout, recalled a recent exercise in Alaska where he led a resupply convoy in total darkness and wished for autonomous vehicles to help him out. Scout sees itself primarily as a software company building an intelligence layer for military machines. It doesn’t intend to make the autonomous vehicles themselves, but instead build atop them. Adcock expects the startup’s first product to be widely adopted will be one called “Ox,” a command and control software bundled with hardened computer hardware like GPUs, communications and cameras. It’s intended to allow individual soldiers to orchestrate multiple drones and autonomous ground vehicles using prompts such as, “Go to this waypoint and watch for enemy forces.” However, making that software work requires training on real vehicles, which is why it has set up Foundry, its training range at the military base. There, drivers spend eight-hour shifts putting the ATVs through their paces, then work through a reinforcement learning system to log where they had to take over, which is used to improve the model. The base commander has even asked the company’s ATV to take a turn with security patrols. One hypothesis Scout is testing is that VLAs will enable this relatively limited data set, alongside training data in simulations, to deliver a fully capable driving agent. While the vehicle seems comfortable on trails, for example, it isn’t ready to operate fully off-road. Scout is also practicing with drones for reconnaissance and defense, giving them intelligence with vision language models. The startup is working on a system that would see groups of munition drones fly with a larger “quarterback” platform that provides more compute resources to command them. For example, the drones could search a geographic area for hidden enemy tanks and attack them, possibly without human intervention. Otis contends that the alternative approach in such a scenario might be indirect artillery fire, which is imprecise compared to drone strikes. While autonomous weapons are a flash point in the politics of defense tech, experts note the concept is old: Heat-seeking missiles and mines have been used in warfare for decades. The question for technologists is how the weapons are controlled, according to Jay Adams, a retired U.S. Army Captain who leads Scout’s operations team. Adams notes the company’s munitions drones can be programmed to only attack threats in a specific geographic area, or only following human confirmation. He also says autonomous weapons platforms are unlikely to fire because they are scared, the way an 18-year-old soldier might be. VLAs, too, hold promise for improving targeting. Scout says its models are pre-trained on a specific set of military data to prepare them for, say, running into an enemy tank while on a resupply mission. Lt. Col. Nick Rinaldi, who supervises Scout’s work for the Army Applications Laboratory, says that while automated targeting is hard and unlikely to be used outside of constrained environments in the near term, the potential of VLAs to reason about threats make them a promising technology to investigate. Adams says the promise of drones that can identify their own targets is key to warfare in the future. While Russia’s invasion of Ukraine has generated intense interest in drone warfare, he believes that humans operating individual UAVs doesn’t scale well enough for the U.S. to face a large number of low-cost unmanned systems should they threaten U.S. forces. Like many defense startups, Scout wears its mission on its sleeve, and its executives will freely criticize companies that are reluctant to hand their technology over to the government. Google, for example,reportedlypulled out of a Pentagon contest to develop control systems for autonomous drone swarms, a capability Scout is also working on. “The AI people don’t want to work with the military,” Otis told TechCrunch,referencing Anthropic’s spatwith the Pentagon over its terms of service. “None of them are open to running agents on one-way attack drones, or running agents on missile systems.” Nevertheless, Scout is using existing LLMs as the base to build its agents, though it declined to say which ones. Otis says it has agreements with “very well-known hyperscalers” to provide the pre-trained intelligence for Scout’s foundation model. He also declined to say if the company uses open-weight models, such as those offered by Chinese companies. Many companies reliant on AI inference build on top of open-weight models because they are cheaper compared to offerings from frontier labs like Anthropic or OpenAI. Scout expects to address this by building its own model from the ground up in the years ahead, and the founders say much of its capital will go into those training and compute costs. Indeed, Otis wonders if Scout will beat the existing leaders to AGI because its model will be constantly interacting with the real world. “There’s an argument in the AGI community along the lines that you can only get so intelligent by reading the internet, and most intelligence comes with interacting in the world,” Otis said. Does that mean Adcock is competing with his brother’s army of humanoid robots at Figure? No, Otis says, but “we can get to scale much faster because our customer has assets,” he said, referring to the Pentagon.
View
Musk Accuses OpenAI Executives of Turning Charity Into Business
Musk is expected to continue his testimony on April 29. A verdict could arrive as early as mid-May.
View
Why DeepSeek V4 Did Not Have an R1 Moment
Besides DeepSeek, new releases from Qwen, Moonshot AI, MiniMax, and others have captured production traffic.
View
