Deep33 Ventures Closes $200M Deep-Tech Fund Targeting AI Infrastructure
Israeli deep-tech venture capital firm Deep33, founded by investor Lior Prosor and U.S. entrepreneur Michael Broukhim, has closed its debut fund at $200 million well above its original $150 million target.
Israel’s Deep33 Bets Big on Physical Tech to Power the AI Era
More than half of the fund’s capital comes from overseas investors, primarily from the U.S. and UK, alongside Israeli institutional investors who, until recently, viewed deep-tech and hardware as too complex and risky to back. The fund’s managers say there has been a notable shift in sentiment among Israeli institutional players, who now increasingly recognise the sector’s potential.
Operating from the U.S., the fund focuses on companies in AI infrastructure, energy, quantum computing, communication components, and advanced physical technologies. Around 70% of investments will target Israeli-founded companies, with the remaining 30% directed at U.S.-based startups.
Deep33’s thesis built around a clear conviction: the next decade of technology will not be won on software alone. The firm argues that the global AI investment cycle now estimated at $1 trillion is being constrained less by software than by physical capacity, including power, compute, and deployment infrastructure.
Co-founder Broukhim, formerly of FabFitFun and an early backer of SpaceX, Stripe, and Hut8, described Deep33 as occupying “a critical and underserved niche at the intersection of Israeli engineering and U.S. national resilience.”
Beyond capital, the firm has built Deep to Market a network connecting portfolio companies with major tech firms, potential customers, and U.S. government officials. Portfolio companies include Particle, QuamCore, Seabridge, and OHR.
The fund’s ultimate ambition is for Israel to establish a position in deep-tech supply chains similar to what Taiwan built in semiconductors a bet that the physical infrastructure of the AI era is both Israel’s next frontier and its biggest opportunity.
