Harvard AI lab seeks $100M
A new Harvard AI lab is reportedly seeking $100 million to fund research aimed at enabling humans to 'remember everything,' signalling institutional investment interest in memory‑augmentation technologies. The fundraising ask was highlighted in social posts drawing attention to the lab’s ambitious research goals and potential implications for human‑computer memory systems. (x.com)
A Harvard-linked startup led by neuroscientist Gabriel Kreiman is in talks to raise about $100 million for artificial-intelligence tools designed to help people “remember everything.” (bloomberg.com) The company is Engramme, which came out of stealth in March and says its goal is to give users “searchless, promptless recall” of people, conversations, and places from their own lives. Bloomberg reported the fundraising talks on April 10, 2026, citing people familiar with the matter. (engramme.com, bloomberg.com) Kreiman’s Harvard work helps explain the pitch. Harvard pages describe his lab as focused on artificial intelligence, computational neuroscience, and memory, including how the brain encodes and retrieves experiences. (mbb.harvard.edu, biophysics.fas.harvard.edu) The underlying idea is memory augmentation: software that builds a structured record of your life from conversations, documents, photos, and location history, then surfaces the relevant piece when you need it. Harvard’s Institute for Quantitative Social Science described that approach last year in a profile of the founders’ earlier venture, Memorious. (iq.harvard.edu) That profile said the founders interviewed more than 50 potential users, including older adults with memory-loss concerns, project managers, and developers, before building a beta iPhone application and an enterprise application programming interface. (iq.harvard.edu) Engramme now says it is building “AI to augment human memory,” and its founding team includes Kreiman, Harvard-trained computer scientist Spandan Madan, Harvard-trained researcher Will Xiao, and former Google designer Emma Nguyen. (engramme.com) The company is also trying to separate itself from general-purpose chatbots. Its website says the product is powered by “novel AI models of human memory,” while a recent product report said Engramme had opened a memory application programming interface beta for developers and teams. (engramme.com, testingcatalog.com) The pitch lands as technology companies push further into “personal context” products that store more of a user’s digital history so software can answer with better recall. Engramme’s version takes that idea to its limit: not just helping a chatbot remember a thread, but trying to turn memory itself into a product. (forbes.com, engramme.com) That also raises the obvious questions: what gets recorded, where it is stored, who can access it, and how a system built to preserve every conversation handles consent and error. Engramme’s public materials emphasize the recall goal; fuller answers on governance and privacy may matter as much as the science if the $100 million round comes together. (engramme.com, bloomberg.com)
