Hivemind Capital and the University of California, Berkeley have launched darkmatter lab, a programme designed to support frontier technology research before company formation.
The initiative focuses on early-stage academic work in artificial intelligence and blockchain that has not yet moved beyond the university system. darkmatter lab will serve as a core part of Hivemind's upcoming venture fund, while Berkeley is providing access to its innovation and entrepreneurship network through Berkeley SkyDeck.
Google Cloud, Gunderson Dettmer and Goodwin Procter are also participating. They will provide cloud credits and legal support in areas including immigration, intellectual property and incorporation.
Each selected project will receive at least USD $1 million in resources. The package includes research funding from Hivemind, USD $350,000 in compute credits from Google Cloud through the Google for Startups Cloud Program, legal support from Gunderson Dettmer and Goodwin Procter, and operating support through Berkeley SkyDeck.
The programme will begin with projects in AI infrastructure, agentic systems, cybersecurity, compute optimisation, cryptography and decentralised systems. The partners said the model was designed to support researchers at a stage when traditional venture investors often do not participate and corporate funding may come with conditions on intellectual property.
Hivemind argues that this part of the research pipeline has become harder to finance as public support comes under pressure. darkmatter lab was created to help move advanced research from university laboratories into commercial use without requiring researchers to form companies first.
"Too many of the most important breakthroughs never make it out of the lab, or take years to do so," said Emmanuel Vallod, Head of Venture and Research at Hivemind Capital.
He described the gap the programme is intended to address. "The most consequential work happens before a company exists, but that's also when resources are most limited. darkmatter lab is built to support researchers at that moment, with the capital, compute and expertise needed to move from idea to deployment," said Vallod.
Funding gap
The structure reflects a broader debate over how universities and private investors should support commercially relevant research. Deep technology projects often require long development periods, specialist infrastructure and legal work before they are ready for spinout or outside investment, leaving researchers with limited options in the earliest stages.
The issue is especially visible in fields such as AI and blockchain, where computing costs can be high and intellectual property questions can determine who benefits from a breakthrough. By combining funding, cloud access and legal services in one package, the organisers aim to create a path for researchers who are not yet ready to establish a business.
UC Berkeley said the programme aligns with its longstanding role in turning academic work into commercial technology. The university has produced a large number of start-ups and research spinouts, and remains a prominent source of work in computer science, AI and cryptography.
"UC Berkeley has always been where foundational research becomes transformative technology," said Rich Lyons, Chancellor of UC Berkeley.
He said the university's researchers needed support that matched the scale of their work. "Our researchers are working on some of the most important problems in AI and blockchain, and they deserve resources that match the ambition of their work. Hivemind understands the role research plays in driving innovation forward, and darkmatter lab is exactly the kind of partnership universities need right now: patient, research-first capital that keeps our researchers in control of their work," said Lyons.
Partner roles
For Hivemind, the launch also signals how the firm plans to shape its next investment vehicle. Rather than waiting for start-ups to emerge from academic settings, it is moving further upstream to identify and support technical work before formal company structures are in place.
The inclusion of law firms alongside a cloud provider highlights the practical barriers to commercialising university research. Questions around visas, patent ownership, company formation and compute access can all delay progress for academic teams, even when the underlying science is promising.
darkmatter lab is intended to reduce those delays by bundling support around selected projects while researchers still control their work and before external funding structures become more complex.