On Monday, Andy Konwinski, a computer scientist and co-founder of Databricks and Perpelexity, said that his own business, Laude, is establishing a new AI research institute with a $100 million investment.
The Laude Institute functions more as a fund seeking to make grants-like investments than as an AI research center. Konwinski is joined on the board by Jeff Dean, Google’s top scientist, Joelle Pineau, Meta’s vice president of AI Research, and Dave Patterson, a professor at UC Berkeley who has a long list of award-winning research.
The new AI Systems Lab at UC Berkeley will be anchored by the institute’s first major “flagship” grant, which Konwinski revealed will be $3 million year for five years. Ion Stoica, a renowned Berkeley professor and the current head of the Sky Computing Lab, is in charge of this new lab. Additionally, Stoica is a co-founder of the AI big data company Databricks and the startup Anyscale, which is a Python and AI platform built in Berkeley’s lab system.
Together with Stoica, several other renowned experts will be part of the new AI Systems Lab, which is scheduled to open in 2027.
Konwinski stated that the institute was “built by and for computer science researchers” in his blog post announcing it. Our goal is to spur work that not only advances the field but also steers it in the direction of more advantageous results.
That isn’t strictly a jab at OpenAI, which began as an AI research center before its massive commercial side allegedly took over. However, other scholars have also succumbed to the allure of financial gain.
For example, Epoch, a well-known AI researcher, came under fire when it disclosed that OpenAI had helped develop one of its AI benchmarks, which was subsequently used to introduce its new o3 model. The founder of Epoch also started a firm with the contentious goal of using AI agents to replace all human employment worldwide.
Similar to previous AI research groups with business goals, Konwinski has established his institute in a way that transcends boundaries: as a nonprofit with an operating arm that is a public benefit corporation.
He is allocating his research funds into two categories, which he refers to as “Slingshots and Moonshots.” Slingshots are intended for early-stage research that can profit from funding and practical assistance. Moonshots are for “long-horizon labs tackling species-level challenges like AI for scientific discovery, civic discourse, healthcare, and workforce reskilling,” as the name suggests.
For example, his lab has worked with Anthropic’s “terminal-bench,” a Stanford-led metric for how well AI agents perform tasks.
It should be noted that Konwinski’s business, Laude, is more than just a research institute that writes grants. Additionally, in 2024, he co-founded a for-profit business fund. Pete Sonsini, a former NEA VC, is one of the fund’s co-founders. Laude oversaw a $12 million investment in Arcade, a business that develops AI agent infrastructure. It has also covertly supported other startups.
According to a Laude representative, Konwinski has committed $100 million, but he is also seeking and amenable to investment from other accomplished technologists. How did Konwinski get rich enough to secure $100 million for this new venture? In January, Databricks raised a $15.3 billion fundraising round, bringing the company’s valuation to $62 billion. Additionally, Perplexity was valued at $14 billion last month.
Is the world truly in need of another AI study that claims to be “good for humanity” or that has a hazy nonprofit/commercial structure? Yes, and no.
Research on AI has grown more and more confused. For example, there are now several AI benchmarks that demonstrate the superiority of a specific vendor’s model. (For CRMs, Salesforce even has its own LLM benchmark.)
An appealing alternative might be an alliance with people like Konwinski, Dean, and Stoica who support genuinely independent research that has the potential to develop into autonomous, human-helpful commerce in the future.