The following content is taken from Book “The Road to Nvidia: Jen-Hsun Huang and the Making of a Tech Giant,” written by Tae Kim, senior technology writer at Barron’s, was published by WW Norton & Company on December 10. This excerpt is from a chapter by Starboard Value, the activist hedge fund founded by Jeff Smith.
In early 2013, Nvidia shareholders were getting restless. The share price has been essentially flat over four years, and financial performance has been mixed. In the latest quarter ended in January, sales rose 7% on the year, but earnings fell 2%.
Nvidia has a strong balance sheet of about $3 billion in net cash, an important asset when the company has a total market capitalization of $8 billion. However, its growth rate is in the single digits, resulting in a price-to-earnings (P/E) ratio of only 14 times. After recovering Nvidia’s cash on hand, Starboard believes that the company is significantly undervalued and that its core assets have more room to grow. The fund goes on the offensive: The hedge fund accumulated 4.4 million Nvidia shares in the quarter ended June 2013, worth about $62 million, according to SEC 13F filings.
Some Nvidia executives aren’t excited about Starboard as an investor. A senior Nvidia executive said the company’s board is very worried about activist foundations forcing the company to reorganize, form its own board, and have Nvidia cut its investment in CUDA – the kind of radical reinvention it is trying with Darden Year. Another Nvidia executive said Starboard wanted a board seat, but the board refused.
Despite this, the relationship never became too hostile. “I don’t think things have reached what I would call a crisis stage yet. Do you know about DEFCON 1?” an Nvidia executive said, referring to the U.S. military’s warning system for nuclear war. DEFCON 5 means peace, DEFCON 1 means nuclear war is imminent. “Arrived at DEFCON 3.”
The Starboard team met with Jensen and other Nvidia leaders multiple times to discuss strategy. Looking back on the investment years later, Smith said Starboard mainly advocated an aggressive stock repurchase program and de-emphasized non-GPU projects such as mobile phone processors. After the session, no additional pressure was applied to starboard. Hedge funds finally got their wish to buy back. In November 2013, Nvidia made two announcements: a commitment to repurchase $1 billion of stock in fiscal 2015 and an authorization to repurchase an additional $1 billion of stock. Over the next few months, the stock price rose about 20%, and Starboard sold its stake in Nvidia the following March.
The relationship between Nvidia and Starboard has been far from contentious and seemed to work well together during this brief period.
“We were very impressed with Jason,” Smith said.
For his part, Jensen recalls meeting Starboard but not much of what was discussed. Before he knew it, Starboard was no longer an investor. But that doesn’t end Starboard’s impact on the chip industry and Nvidia.
In 1999, several Israeli technology executives founded a company called Mellanox, led by Eyal Waldman, who became its CEO. Mellanox provides high-speed networking products under the “InfiniBand” standard for data centers and supercomputers, and has quickly become an industry leader. Its revenue growth has been impressive, from $500 million in 2012 to $858 million in 2016.
In January 2017, Starboard acquired an 11% stake in Mellanox. It issued a letter criticizing Waldman and his team for their disappointing performance over the past five years. Mellanox shares fell despite a 470% gain in the semiconductor industry index. Its operating profit margin is half that of its peers. “Mellanox has long been one of the worst-performing semiconductor companies,” Starboard wrote in the letter. “The time for marginal changes and marginal improvements is long past.”
After a series of lengthy discussions with the board of directors, Starboard and Mellanox reached a compromise in June 2018. right. Even with those concessions, Starboard still retains the option of launching a proxy fight to replace Waldman. Alternatively, Mellanox could choose to sell itself to a company that can generate a better return on assets than it could as a standalone company. This set the stage for one of the most important deals in the history of the chip industry.
In September 2018, Mellanox received a non-binding takeover bid from an outside firm for $102 per share, almost a one-third premium to its current share price of $76.90. Mellanox is now fully functional. It sought additional bidders from an investment bank, eventually expanding the list of potential buyers to seven.
Another Nvidia executive said Jensen wasn’t considering acquiring Mellanox when it went public. But he quickly saw the strategic importance of the asset, decided Nvidia had to win the auction, and joined the auction house in October.
Eventually, the list was narrowed down to three serious bidders: Nvidia, Intel and Xilinx, which mainly make chips for industrial use. Three potential buyers were locked in a months-long bidding war, with Intel and Xilinx offering the highest bid of about $122.50 per share. Nvidia’s stock price rose slightly to $125 per share. It won the bidding war on March 7, 2019, with an all-cash offer of $6.9 billion.
Days later, Nvidia and Mellanox made the deal public and held a conference call with analysts and investors.
“Let me tell you why this makes sense for Nvidia and why I’m excited about it,” Jensen said. He talked about how the need for high-performance computing will grow – how workloads such as artificial intelligence, scientific computing and data analysis will require huge performance improvements that can only be achieved through GPU-accelerated computing and better networking. He explained how AI applications will eventually require tens of thousands of servers to be connected and work together, and Mellanox’s market-leading networking technology will be critical to making this happen.
“Emerging artificial intelligence and data analytics workloads require optimization of data center scale,” he said. Jensen predicts that computing will move beyond one device and that entire data centers will become computers.
A few years later, Jensen’s vision came to fruition. In May 2024, Nvidia disclosed that the portion of the company formerly known as Mellanox had quarterly revenue of $3.2 billion, a more than sevenfold increase from the last quarter Mellanox reported as a public company in early 2020. Just four years later, the former Mellanox business (which once cost Nvidia a one-time $6.9 billion fee) generates more than $12 billion in annual revenue and is growing at a triple-digit rate.
“Melanox is, frankly, a great thing that activists have thrown at us,” said a senior Nvidia executive. “If you talk to artificial intelligence startups today, Mellanox’s networking technology, InfiniBand, is critical to scaling computing power and making everything work.”
Brian Venturo, co-founder and chief technology officer of CoreWeave, a leading GPU cloud computing provider and Nvidia customer, believes that InfiniBand technology remains the best solution for minimizing latency, controlling network congestion, and enabling efficient execution of workloads.
In some ways, Mellanox is a lucky accident for Nvidia. Jensen wasn’t in control from the start. But once Nvidia discovered and understood the opportunity, it decided to aggressively pursue Mellanox. It’s a great deal, although the outcome depends on Nvidia’s ability to execute on the new business once it becomes part of the company. In these ways, Mellanox is Nvidia’s quintessential achievement: The company pounced when other companies failed to take action, and Mellanox helped propel Nvidia’s dominance in artificial intelligence.
“This will definitely go down in history as one of the best acquisitions ever,” said Jay Puri, head of global field operations at Nvidia. “Jensen realized that data center scale computing required really good, high-performance networking. And Mellanox is the best in the world at this.”
After seeing all that Nvidia has accomplished over the past decade, Starboard Value’s Jeff Smith also has a closing thought.
“We should not be withdrawing from this position at all.”
