Your gaming data could be the secret to AGI, according to this Bezos-backed startup. When it comes to achieving artificial general intelligence (AGI), large language models justdon'thave what it takes. Models like ChatGPT and Claudeare great at text, butthey'reless skilled at understanding how thingsactually movethrough space and time - an essential skill for producing intelligence that generalizes. That gap, it turns out, might be filledbygaming data.That'sthe bet behind General Intuition, a Bezos-backed, New York-based startupvalued at$2.3 billionthat just closed a $320 million roundwith Coatue, Eric Schmidt, and researchers at MIT and Google DeepMind joining its list of investors. On this episode of TechCrunch'sEquitypodcast, General Intuition CEO Pim de Witte joins Rebecca Bellan to dig into why world models trained on gaming data might be the next big leap in physical AI, how the company spun out of gaming platform Medal TV, and where the ethical red lines are when your models could end upbeing used for defense applications. Listen to the full episode to hear more about: * How eight minutes of real-world data was enough to get a robot navigating an office cold. * Why General Intuition turned down an acquisition offerreportedly fromOpenAI to stay independent, and why havinginvestors whoback your mission is essential to building a generationalcompany. * How the company is trying to get ahead of AI job displacement by building Nerve, a marketplace connecting gamers to data labeling and teleoperations work. Subscribe to Equity onYouTube,Apple Podcasts,Overcast,Spotifyand all the casts. Youalso canfollow Equity onXandThreads, at @EquityPod. Rebecca Bellan is a senior reporter at TechCrunch where she covers the business, policy, and emerging trends shaping artificial intelligence. Her work has also appeared in Forbes, Bloomberg, The Atlantic, The Daily Beast, and other publications. You can contact or verify outreach from Rebecca by emailing [email protected] or via encrypted message at rebeccabellan.491 on Signal. Theresa Loconsolo is an audio producer at TechCrunch focusing on Equity, the network's flagship podcast. Before joining TechCrunch in 2022, she was one of 2 producers at a four-station conglomerate where she wrote, recorded, voiced and edited content, and engineered live performances and interviews from guests like lovelytheband. Theresa is based in New Jersey and holds a bachelors degree in Communication from Monmouth University. You can contact or verify outreach from Theresa by emailing [email protected].

Q3 2026 funding watch: the top 5 AI robotics races and the data centre capex squeeze. As the first half of 2026 closes with a record $18.8 billion in global robotics funding, the third quarter pivots from fundraising to delivery. This forward-looking watch examines the upcoming IPO wave, the delivery scorecard for Tesla, Figure, and 1X, and how the $700 billion hyperscaler data centre buildout is squeezing the robotics supply chain. The closing week and the Q3 pivot. The first half of 2026 closed with a historic $18.8 billion poured into global robotics startups, driven by a massive concentration of capital in general purpose humanoid platforms. However, the closing week of June and early July signaled a distinct pivot. The market is shifting from private venture accumulation to public market validation and supply chain execution. The closing week delivered a rapid succession of IPO movements that set the stage for the third quarter. In China, Unitree Robotics secured registration approval from the China Securities Regulatory Commission for its highly anticipated STAR Market debut, backed by 2025 revenue of 1.699 billion RMB and a 332 percent growth rate. Deep Robotics filed its own prospectus for a 2.5 billion RMB raise on the same exchange. In Hong Kong, Rokae Robotics priced its offering for a July 9 listing, while autonomous driving firm Momenta launched its IPO with $375 million in cornerstone backing from GIC, Fidelity, BlackRock, and Mercedes Benz. In South Korea, dexterous hand developer Tesollo selected underwriters for a KOSDAQ tech special listing, and globally, Agility Robotics moved closer to finalizing its $2.5 billion SPAC merger under the ticker AGLT. The top 5 AI robotics races to watch. As capital seeks liquidity through these public offerings, private investment in Q3 will concentrate around five distinct technology races. Breakthroughs in any of these verticals will likely trigger the next wave of mega rounds. The first race is the general purpose humanoid platform competition. The focus has moved entirely past walking demonstrations to multi purpose autonomous task execution. Companies like Figure AI, Tesla, Unitree, UBTECH, and Apptronik are competing not just on hardware, but on the speed at which their platforms can learn new skills in unstructured environments. The second race centers on robot foundation models and world models. Software companies building universal brains for physical AI are attracting immense capital. Startups like Skild AI, Physical Intelligence, and General Intuition are competing to build the definitive operating system for embodied AI, betting that software will ultimately commoditize the hardware layer. The third race is the battle for the dexterous hand. As humanoids move into complex manipulation tasks, the actuator density and tactile sensing of the end effector have become critical bottlenecks. Specialized companies are raising significant capital to solve the hand problem for the broader industry. The fourth race involves embodied data and simulation infrastructure. Training physical AI requires billions of hours of action labeled data. Companies building the simulation environments, synthetic data pipelines, and teleoperation rigs to feed these models are becoming the essential picks and shovels of the robotics gold rush. The fifth race is deployment and fleet orchestration. Hardware is useless without the software to manage it at scale. Platforms designed to orchestrate heterogeneous fleets of robots across warehouses, factories, and retail environments are seeing increased enterprise adoption and corresponding venture interest. The delivery scorecard: promises meeting reality. The third quarter of 2026 is the critical delivery checkpoint for the aggressive manufacturing targets set over the past eighteen months. The industry is watching closely to see if production promises are matching reality. Tesla began Optimus production in the second quarter of 2026 at its Fremont facility, replacing legacy Model S and X lines. While earlier market expectations pointed to 50,000 units this year, the company has tempered near term targets, stating that initial skills will be limited to simple factory tasks. Figure AI has demonstrated aggressive scaling at its BotQ facility. After announcing a first generation line capable of 12,000 units annually, the company successfully transitioned from prototype to production phase, reportedly scaling Figure 03 output from one robot per day to one robot per hour in under 120 days. In the consumer space, 1X commenced full scale production of its NEO home robot at a 58,000 square foot factory in Hayward, California. The company, which booked 10,000 pre orders in five days last October, maintains its commitment to begin consumer deliveries before the end of 2026, aiming for a 100,000 unit annual capacity by late 2027. In China, AgiBot shipped 5,000 humanoids in the first quarter alone, setting a high bar for domestic deployment volume. The success of Unitree's STAR Market IPO in Q3 will serve as the ultimate financial scorecard for whether these aggressive production volumes translate into sustainable profitability. The data centre capex squeeze. The most significant headwind facing the robotics industry in Q3 2026 is not a lack of demand, but the gravitational pull of hyperscaler artificial intelligence infrastructure. The top five United States technology giants are projected to spend nearly $700 billion on capital expenditures this year, a 75 percent increase from 2025. This massive data centre buildout is reshaping the electronic component supply chain. AI data centres are expected to consume up to 70 percent of global memory chip production in 2026. High bandwidth memory now occupies 23 percent of total DRAM wafer capacity, driving severe shortages and price spikes. Semiconductor lead times reached 40 weeks in March, and power management integrated circuits are expected to remain constrained throughout the year. For robotics manufacturers, this translates directly into supply chain friction. Robots share many of the same component categories as AI servers, including memory, power management chips, multi layer ceramic capacitors, and high density connectors. Robotics procurement teams are now competing directly against trillion dollar hyperscalers for limited cleanroom capacity, driving up bill of materials costs and extending production timelines. However, this capital concentration also provides a tailwind. The massive investment in AI compute is rapidly driving down the cost of inference, making the cloud brains that power embodied AI significantly cheaper to operate. Furthermore, the power constrained data centre construction boom is creating a massive new demand pocket for robotics. From semiconductor fab lifting robots in South Korea to automated inspection systems for hyperscale cooling infrastructure, the AI data centre buildout is simultaneously squeezing the robotics supply chain and creating its most lucrative new customer base.

General Intuition raises $320M to use video game data to train robots. General Intuition is testing world models that will act as training environments for agentic models. Source: General Intuition General Intuition US Inc. this week raised $320 million in Series A funding. The company said it plans to use the financing to build AI models that can perceive, predict, and act in virtual and physical environments. While physical AI has become a dominating topic in robotics, General Intuition claimed that it is taking a unique approach. Instead of gathering hundreds or thousands of hours of real-world data or generating simulated data, the company uses billions of gameplay clips uploaded to Medal, a platform that allows users to post gaming moments. Pim de Witte, founder and CEO of General Intuition, also co-founded Medal. The gaming clips capture humans perceiving an environment and deciding how to move through it. This is what makes it valuable, said the company. Text-based models only provide descriptions of reality, which aren't enough for training physical AI, it said. In addition, the videos come with embedded action labels. These record exactly what button a player presses and when, giving General Intuition more information about how players make decisions. The New York-based company said its Series A brings its valuation to $2.3 billion. The round also brings its total funding to $454 million, which comes after the $134 million it raised in October. General Catalyst led the round, which also included participation from Amazon founder Jeff Bezos and former Google CEO Eric Schmidt. General Intuition quickly gains momentum. Many of the most powerful foundation models are trained on written words. However, General Intuition said human intelligence far exceeds language. It emerged over millennia of interaction and exploration, through the endless cycle of intent, action, and consequences across diverse environments, it asserted. Truly intelligent machines must move from words to worlds, and acquire the capacity to perceive, anticipate, and improvise. They need to obtain a general intuition of reality, said General Intuition. The company explained that its models learn from unique, action-labeled video datasets across countless environments. It said this diversity leads to uniquely capable agentic systems. Since its founding in 2015, General Intuition said it has worked to develop action models that decide what action to take, and world models that predict the outcome of actions. It plans to use the funding to scale its compute capacity and focus on pretraining the next version of its model. General Intuition also hopes to make its API more broadly available this summer, according to TechCrunch. Submit your session idea for the 2026 RoboBusiness