A different kind of billionaire: Willenz adds another $50 million exit
The Israeli chip veteran continues to defy convention with precise, and early, dealmaking.
The history of Israel’s semiconductor industry is, in many ways, intertwined with the career of Avigdor Willenz, one of the figures most closely associated with placing the country on the global chip map. His trajectory began in 1993 with the founding of Galileo Technology, which was later listed on Nasdaq and sold in 2001 to Marvell Technology for $2.7 billion, a landmark deal that helped usher in Israel’s era of large-scale tech exits.
In 2015, Willenz recorded another major success with the sale of Annapurna Labs to Amazon Web Services for approximately $370 million. What initially appeared to be a strategic but modest acquisition evolved into a cornerstone of Amazon’s infrastructure. Annapurna became central to AWS’s development operations in Israel, with its server chips now generating revenues estimated in the tens of billions of dollars.
In a letter to investors published earlier this month, Amazon CEO Andy Jassy noted that the company’s chip division, largely built on the Annapurna acquisition,has already generated $20 billion in revenue this year. He added that if the unit operated independently, it could be valued at around $50 billion, placing it in direct competition with industry leaders such as Intel and Nvidia. Jassy described the original acquisition as a move that seemed “crazy” at the time, but ultimately enabled Amazon to reduce reliance on external suppliers and build its own hardware stack.
Willenz’s ability to anticipate technological shifts has been a recurring theme. As an early investor and chairman of Habana Labs, he helped guide the company to its $2 billion sale to Intel in 2019. He was also behind the sale of Leaba Semiconductor to Cisco in 2016 for $320 million, and Banias Labs to Alphawave in 2022 for $240 million.
His influence extends beyond Israel. As an early investor in Astera Labs, Willenz participated in its March 2024 IPO at a valuation of $5.5 billion. The company has since surged to a market value of roughly $27 billion.
The latest addition to this track record comes with the sale of DustPhotonics to Credo Technology Group. Under the terms of the deal, Credo will pay $750 million in cash and $123 million in shares, with an additional earnout of up to approximately $430 million. Willenz, who holds a roughly 6% stake in DustPhotonics, is expected to receive around $50 million from the transaction.
At the same time, Willenz continues to build new ventures. One of the more notable is Element Labs, founded alongside former Habana executives David Dahan and Ran Halutz. The company is developing AI processors focused on inference, the stage where trained models are deployed in real-world applications. Unlike Habana, which focused on large-scale training chips, Element Labs is targeting smaller, local data centers, aiming to reduce bandwidth and energy strain while improving response times. The company has raised $50 million to date.
Despite this long string of successes, Willenz’s strategy has not been without criticism. Some argue that he tends to sell companies too early, before their full value is realized. Annapurna is often cited as the clearest example. Although Habana Labs was sold for a substantial sum, Intel ultimately struggled to translate its technology into a dominant position in the AI chip market.
A similar question now surrounds DustPhotonics. The silicon photonics sector, which integrates optical components directly onto chips to enable faster and more energy-efficient data transfer, is still in its early stages of adoption. With demand for AI infrastructure accelerating, it remains possible that the company’s full upside has yet to be realized.
