Editorial illustration for Bezos-backed startup raises USD 320 million to build AGI from gaming data
Bezos-Backed Startup Raises $320M for AGI Using Gaming Data
Bezos-backed startup raises USD 320 million to build AGI from gaming data
General Intuition just closed a $320 million funding round at a $2.3 billion valuation, with Jeff Bezos, Eric Schmidt, Coatue, and researchers from MIT and Google DeepMind all writing checks. The New York startup's pitch: forget scraping more text off the internet. The path to artificial general intelligence runs through gaming data instead.
Large language models like ChatGPT and Claude can write essays and pass the bar exam, but they stumble on something a toddler handles without thinking, tracking how objects move through space and time. General Intuition thinks the fix is sitting inside billions of hours of gameplay footage, which happens to be exactly what its predecessor company, Medal TV, was built to collect.
CEO Pim de Witte spun the AI venture out of that gaming platform, and he sat down with TechCrunch's Rebecca Bellan on the Equity podcast to explain the bet. The conversation covers what gaming footage can teach a robot about navigating real rooms, why the company walked away from an acquisition offer reportedly from OpenAI, and where it draws the line on defense uses for its technology.
When it comes to achieving artificial general intelligence (AGI), large language models just don’t have what it takes. Models like ChatGPT and Claude are great at text, but they’re less skilled at understanding how things actually move through space and time — an essential skill for producing intelligence that generalizes. That gap, it turns out, might be filled by gaming data.
Why this matters
General Intuition's pitch rests on a real gap: LLMs read text well but stumble on spatial reasoning, the kind of intuition you build by tracking objects, physics, and movement over time. Gaming footage, with its dense record of cause and effect in three-dimensional space, is a plausible source for that missing training signal. A $2.3 billion valuation and backers like Eric Schmidt plus researchers from MIT and DeepMind signal that serious money believes this too.
But we should be careful about the leap from "useful training data" to "path to AGI." Spatial reasoning is one piece of general intelligence, not the whole puzzle, and plenty of well-funded bets on the next unlock have stalled once the data ran out or didn't generalize past the games it came from. For founders and researchers, the takeaway isn't that gaming data is the answer, it's that data scarcity in LLM training is pushing serious capital toward unconventional sources. Watch whether General Intuition ships models that handle real-world physical tasks, not just game environments, before treating this as validated.
Related Reading
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
From Archives Anthropic Launches Claude Cowork AI Agent for Mobile and Web
From Archives Anthropic Moves Its Claude Agent to Phones as AI Rivals Target Mobile
Common Questions Answered
Why does General Intuition believe gaming data is essential for achieving artificial general intelligence?
General Intuition argues that large language models like ChatGPT and Claude excel at text-based tasks but lack spatial reasoning abilities—understanding how objects move through space and time. Gaming data provides a dense record of cause-and-effect relationships in three-dimensional environments, offering the training signal needed to develop this crucial capability for true AGI.
What is the funding round size and valuation for General Intuition's Series B?
General Intuition closed a $320 million funding round at a $2.3 billion valuation. The round was backed by notable investors including Jeff Bezos, Eric Schmidt, Coatue, and researchers from MIT and Google DeepMind.
What specific limitation do current large language models have that General Intuition aims to address?
Current LLMs like ChatGPT and Claude struggle with spatial reasoning and understanding physical movement through time—tasks that even toddlers can perform intuitively. General Intuition believes gaming data can fill this gap by providing training examples of how objects interact and move in three-dimensional space.
Who are the key investors backing General Intuition's AGI development approach?
General Intuition's funding round includes backing from Jeff Bezos, tech executive Eric Schmidt, venture capital firm Coatue, and researchers from prestigious institutions MIT and Google DeepMind. These high-profile backers signal significant institutional confidence in the startup's gaming-data-based approach to AGI.
Further Reading
- General Intuition raises $320 million to develop AI from gaming - Axios
- General Intuition's $2.3B bet that video games can train AI agents for the real world - TechCrunch
- Jeff Bezos' family office backed five AI startups in June - CNBC
- Bezos-backed AI agents: General Intuition eyes $300M at ~$2B - Instagram
- Jeff Bezos is investing in a new AI startup betting that video games can train AI - X (Twitter)
