Key points
- The Decoder categorized 45 new AI labs by their specific research bets and mapped limitations of today's systems they aim to overcome.
- High-profile founders including Ilya Sutskever and Yann LeCun left major tech companies to start independent labs with significant venture backing.
The Decoder published Frontier Radar #5, a report that sorts 45 new AI labs by their research bets and analyzes the limitations of current models they seek to address.reported The report profiles 27 labs in detail across six research categories, while briefly covering nearly 20 others.reportedA report by The Decoder examines why top AI researchers are leaving big tech companies to start their own labs with billions of dollars.The Decoder’s Frontier Radar #5 categorizes 45 neolabs by their research bets, highlighting a strategic divergence from the current AI development recipe. It examines origins, valuations, and publicly available work from these entities, noting that founders are leaving established companies to pursue new directions.
Evidence: reported
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their betsaifleet-decoder
Evidence: reported
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their betsaifleet-decoder
What it does
The report maps the specific limitations of today's AI systems that these new labs are trying to overcome. It identifies the research bets being made by founders who departed from OpenAI, Google DeepMind, and Meta. The analysis looks at what is already visible from their work and explains why Nvidia consistently benefits from these shifts in the industry landscape.
How to get it
The Decoder publishes this Frontier Radar series six times a year as a newsletter and exclusively on its site for subscribers. Issue #5 follows previous reports covering agentic AI, AI impact on productivity, the token economy, and the global AI competitive landscape involving China and Europe.reported Access to the full detailed profiles is restricted to paying subscribers.
Limits
Many of these labs have raised billions of dollars but have no product to show yet. Safe Superintelligence, co-founded by former OpenAI chief scientist Ilya Sutskever, has not released a single model or paper despite reported investments.reported
Evidence: reported
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their betsaifleet-decoder
The takeaway
Founders are betting that the current recipe for AI development alone will not be sufficient for future progress. They are building independent labs to pursue alternative research directions outside the major tech companies.
Building on this
Data science teams should note that the Decoder report indicates a divergence in research strategies among top AI talent. Developers can use the categorized research bets to identify emerging techniques that address current model limitations. Teams can monitor the work of these 45 labs for early signals of new architectural approaches or training methods that differ from the standard recipes used by OpenAI, Google, and Meta.reported
Evidence: reported
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their betsaifleet-decoder
Why it matters
- The report provides a structured view of 45 labs, helping teams identify which alternative research bets might yield practical improvements over current models.
- Founders leaving major labs suggests that the current dominant AI development path has perceived limits, signaling potential shifts in industry standards.
- The lack of products from well-funded labs like Safe Superintelligence indicates that breakthroughs may take time, affecting expectations for near-term tooling.
Topics: AI research, neolabs, AGI, The Decoder, Nvidia
Updated
Then / now / next
AI development was concentrated within major tech companies like OpenAI, Google, and Meta.
Founders from these companies are launching independent labs with billions in funding to pursue alternative research bets.
Teams should watch for the first models or papers released by these labs to see if their bets yield practical results.
Story timeline
- Then AI development was concentrated within major tech companies like OpenAI, Google, and Meta.
- 06 Nov 2026 Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their bets
- Now Founders from these companies are launching independent labs with billions in funding to pursue alternative research bets.
- Next Teams should watch for the first models or papers released by these labs to see if their bets yield practical results.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
Which major companies are founders leaving to start these new labs?
Founders are leaving OpenAI, Google DeepMind, and Meta to establish their own independent research laboratories.
Has Safe Superintelligence released any models or papers?
No, the report states that the lab has not released a single model or paper despite receiving investment.
What definition of AGI is used in the report?
The report notes that there is no agreed-upon definition of AGI, though executives claim it has arrived or is imminent.
Connected stories
Reading receipt
- The Decoder categorized 45 new AI labs by their specific research bets and mapped limitations of today's systems they aim to overcome.
- High-profile founders including Ilya Sutskever and Yann LeCun left major tech companies to start independent labs with significant venture backing.
Still open: What specific technical limitations are these labs targeting with their undefined research bets?
Still open: Will the lack of public models from these well-funded labs change as they progress?