A growing wave of prominent AI researchers and executives are departing major technology companies including OpenAI, Google DeepMind, and Meta to found independent, research-focused laboratories dubbed “neolabs.” High-profile founders include Ilya Sutskever (Safe Superintelligence), Yann LeCun (AMI), Jerry Tworek (Core Automation), David Silver (Ineffable Intelligence), and Jeff Dean (Discovery Loop).
These newly formed ventures have secured massive financial backing despite lacking commercial products. Safe Superintelligence raised $5 billion from NVIDIA in July, while industrial AI lab Prometheus secured $12 billion at a $41 billion valuation in June. Flapping Airplanes also closed a $1 billion Series A round in October. Investors are backing these ventures under the assumption that current scaling paradigms—relying primarily on larger language models trained on web data—may be insufficient to achieve artificial general intelligence.
The trend underscores a fundamental split in technical strategy across the AI industry. While executives at incumbent firms claim current trajectories remain on track, neolab founders are pursuing alternative bets focused on how AI systems perceive, reason, compute, and self-improve outside standard transformer scaling.
Why it matters
Venture capital is pouring into pre-product frontier research labs, creating intense competition for elite AI talent.
Signals top technical talent believes LLM pre-training on web data may be hitting diminishing returns for AGI.
Incumbents risk losing key research leaders while simultaneously funding their offshoots to maintain strategic exposure.
Source: the-decoder.com



