AI-Powered Virtual Lab Designs 92 New Nanobodies to Target SARS-CoV-2 Variants

Scale AI, a data-labeling startup, has confirmed a major investment from Meta, boosting the company’s valuation to $29 billion. As part of the deal, Scale’s co-founder and CEO Alexandr Wang will step down and join Meta to support its efforts in building superintelligent AI systems. The investment, reported to be approximately $14.3 billion for a 49% stake, underscores Meta’s strategic move to strengthen its AI capabilities amid growing competition. Scale AI, which provides training data for large language models, has become a critical partner for leading AI labs like OpenAI, Google, and Anthropic. The startup’s recent expansion into hiring top-tier talent—including PhD researchers and senior engineers—reflects rising demand for high-quality data in frontier AI development. Meta’s spokesperson confirmed the partnership, emphasizing its focus on collaborative data production for AI models and highlighting Wang’s role in advancing “superintelligence efforts.” Scale AI stated the funding will be allocated to returning capital to shareholders and accelerating growth, while reaffirming its independence as a separate entity. Wang will retain his position as a board director, ensuring continued oversight. This investment comes as Meta seeks to close gaps in its AI model releases, which have lagged behind rivals. Data from SingalFire also noted that Meta lost 4.3% of its top AI talent to competitors in the past year, signaling the urgency of securing expertise through partnerships. Scale AI previously raised $1 billion in 2023 at a $13.8 billion valuation, with investors including Amazon and Meta. The latest funding reflects the critical role of training data in the AI race and the increasing reliance on specialized startups to drive innovation. The collaboration between Scale and Meta highlights a broader trend: AI tools are evolving beyond answering specific scientific queries to enabling open-ended research. This shift could redefine how interdisciplinary teams tackle complex challenges, from drug discovery to climate modeling. While the article’s abstract focuses on the “Virtual Lab of AI agents,” a separate study details a novel computational pipeline combining ESM, AlphaFold-Multimer, and Rosetta to design 92 SARS-CoV-2 nanobodies. Two of these show enhanced binding to recent variants like JN.1 and KP.3, while retaining effectiveness against the original spike protein. Experimental validation confirms their functional potential, offering promising candidates for further development. The research, led by teams at Stanford University and the Chan Zuckerberg Biohub, demonstrates how AI-human collaboration can accelerate impactful scientific discoveries. By integrating LLMs as principal investigators and leveraging human expertise, the Virtual Lab model addresses limitations in traditional interdisciplinary research, where access to diverse specialists is often constrained. This approach marks a significant step in AI’s role in scientific exploration, moving beyond task-specific applications to support complex, creative problem-solving. The findings suggest that such systems could streamline drug development and other high-stakes research areas, reducing time and resource barriers. The study’s authors include Kyle Swanson, Wesley Wu, Nash L. Bulaong, John E. Pak, and James Zou, with correspondence directed to Pak or Zou. The work highlights the intersection of computational biology and AI, offering a blueprint for future collaborations in addressing global health challenges.