10家新公司、9家独角兽，这个新赛道凭什么让硅谷风投疯狂下注？

Core Insights - The current AI investment narrative is characterized by high valuations with minimal products, exemplified by Thinking Machines Lab's valuation of $50 billion despite only having a preliminary API called Tinker [1][6] - A trend of new AI laboratories, referred to as "neolabs," is emerging, with many founders being former top researchers from companies like OpenAI and DeepMind, focusing on innovative AI paradigms rather than immediate commercialization [5][6] - Capital is increasingly betting on the reputations and past achievements of these founders rather than traditional business metrics, leading to astonishing valuations for companies with little to show in terms of products [2][5] Group 1: Neolab Valuations and Trends - Five neolab startups have collectively raised $2.5 billion in funding recently, as traditional giants like OpenAI and Anthropic reach valuations of $183 billion [3] - Neolabs are achieving valuations of $1 billion or more at seed funding stages, with notable examples including Humans& at $4 billion and Safe Superintelligence at $32 billion [2][4] - The founders of these neolabs are primarily individuals who have left major tech companies, bringing significant personal wealth and expertise to their new ventures [5][6] Group 2: Individual Neolab Highlights - Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, is focused on enterprise custom models and voice AI assistant technology, with a valuation of $50 billion anticipated from a funding round of $4-5 billion [11][13] - Safe Superintelligence, co-founded by Ilya Sutskever, aims to develop controllable superintelligence with a reported valuation of $32 billion, emphasizing safety over commercialization [8][10] - Humans&, led by Eric Zelikman, is dedicated to creating emotional AI capable of understanding long-term relationships, achieving a valuation of $4 billion shortly after its inception [14][16] - Periodic Labs, founded by former OpenAI and DeepMind researchers, aims to automate the scientific process, with a focus on low-energy superconductors and materials science [25][27] - Inception Labs, created by Stanford professors, is working on embodied intelligence and long-term learning systems, integrating robotics and sensor technology into manufacturing and logistics [28][30] - General Intuition focuses on spatiotemporal reasoning models using vast amounts of gaming video data, securing $134 million in seed funding [31][32] - Harmonic, co-founded by Vlad Tenev, aims to create a mathematical reasoning system that mimics human intuition, achieving a valuation of approximately $1.45 billion [34] Group 3: Industry Implications - The emergence of neolabs signifies a shift in the AI landscape, moving away from resource-intensive giants to agile, innovative startups led by top-tier researchers [36] - The focus on new paradigms and long-term exploration in AI research is expected to yield breakthroughs that traditional companies may overlook due to their emphasis on immediate results [6][36]