Last night, Jensen Huang did what he had to do and stepped down as CEO of Nvidia. Bad news for shareholders? Not necessarily. The Board reacted quickly and appointed CFO Colette Kress as interim CEO, but interim is the important word. The real job still lies ahead: finding the person who can lead Nvidia into its next era.
And that raises an important question. Does Nvidia really need another Jensen Huang? Another visionary technologist in a black leather jacket, with the same charisma, instincts and appetite for enormous bets? Or would trying to recreate Jensen be exactly the wrong thing to do?
History suggests that successful succession is rarely about finding someone who resembles the leader who just left. It is about identifying the company’s next problem and finding someone whose character, experience and instincts fit that problem. If a great CEO has done his job properly, he should leave behind more than products and profits. He should leave a culture, capabilities and exceptional people who remain after he has gone. A leader may not have been quite as great as we thought if the company he leaves behind turns out to be smaller than he was.
Apple provides perhaps the clearest example. Steve Jobs was obsessive, intuitive, demanding and uncompromising. His genius was product creation: seeing what consumers might want before they knew it themselves and forcing an organisation to make it happen. Tim Cook was almost his opposite. Quiet, analytical and operational, he understood inventories, suppliers, manufacturing, margins and scale. Most importantly, he never made the mistake of trying to become Steve Jobs. By the time Cook took over, Apple no longer needed someone to prove that it could create extraordinary products. Jobs had already solved that problem. Apple now had to turn those products into one of the most efficient commercial machines ever built. Cook’s obsession with supply chains, operational efficiency and capital discipline fitted that challenge remarkably well. His different personality was not something Apple had to compensate for. It was part of the reason the succession worked.
Microsoft offers a more complicated version of the same lesson. Bill Gates built the company around technical dominance, intellectual intensity and fierce competition. Steve Ballmer inherited an immensely profitable business and was a formidable salesman and operator, but Microsoft’s real problem was already beginning to change. Computing was moving away from the PC, mobile had arrived and cloud computing was emerging. Protecting the existing machine was profitable, but increasingly insufficient. Satya Nadella eventually brought a very different temperament. He was quieter, more collaborative and much less interested in defending Windows as the centre of Microsoft’s universe. His experience in cloud computing fitted the technological transition the company faced, while his leadership style helped change a culture known for internal rivalry and defensiveness. Microsoft did not regain its strategic momentum by rediscovering Bill Gates. It did so by finding someone better suited to a problem Bill Gates himself had never needed to solve.
Uber makes the point even more sharply. Travis Kalanick’s aggression, impatience and willingness to challenge almost every established rule helped Uber force its way into cities across the world. Those characteristics were extraordinarily useful when the company’s central problem was expansion at almost any cost. Eventually, however, some of the qualities that had helped create Uber became part of the problem threatening it. Dara Khosrowshahi arrived with almost the opposite personality. More measured, diplomatic and institutional, he did not need to prove that Uber could conquer another city. He needed to prove that Uber could become a sustainable company. The business had changed, the problem had changed and the characteristics required from its leader had changed with it. Sometimes the best successor is not the person who continues the founder’s style. It is the person willing to leave it behind.
There is a broader point here. We tend to attribute an extraordinary company’s success to the person standing at the front, partly because that person is the one we can see. Yet the New York Philharmonic offers a useful comparison. A great conductor can shape interpretation, timing and intensity and lift a performance from good to exceptional. But even the greatest conductor in the world will struggle in front of mediocre musicians. His greatness depends in no small part on the extraordinary professionals sitting in front of him. Nvidia still has that orchestra. Its engineers do not disappear because Huang has gone. Neither do decades of accumulated knowledge, customer relationships, product development, technological expertise and company culture suddenly walk out of the building with him. In fact, Huang’s greatest achievement may ultimately prove not to be that Nvidia depended on him, but that he built an organisation capable of continuing without him.
That is why the Board should resist the temptation to search for Jensen 2.0. Huang solved the problems of creation, technological conviction and market formation. He saw accelerated computing long before most people understood its importance, helped turn CUDA into the centre of an ecosystem and positioned Nvidia at the heart of the AI infrastructure boom. His willingness to make extraordinary technological and strategic bets was exactly what Nvidia needed.
The next phase will demand something different. The AI industry is moving from the excitement of building capacity towards the much less romantic question of what all that capacity will ultimately earn. Nvidia’s largest customers are investing enormous amounts in AI infrastructure while some of them are simultaneously developing their own chips. Governments increasingly regard advanced semiconductors as strategic assets. Nvidia itself has become deeply involved in an ecosystem of customers, suppliers and AI companies in which technology, capital and commercial relationships increasingly overlap. Then there is perhaps the hardest problem of all. Nvidia must preserve the willingness to make extraordinary technological bets while making those bets less dependent on one extraordinary individual. That may require a leader who is technically credible enough to command the respect of some of the world’s best engineers, financially disciplined enough to decide where ecosystem building ends and subsidising demand begins, politically sophisticated enough to operate in a world where semiconductors have become geopolitics, and independent enough to challenge the organisation Huang created rather than simply preserve it.
Most of all, the next CEO needs the confidence not to become Jensen Huang. Tim Cook succeeded partly because he remained Tim Cook. Satya Nadella did not try to become Bill Gates. Dara Khosrowshahi certainly did not try to become Travis Kalanick. Each inherited a company shaped by a powerful predecessor, and each had to understand that yesterday’s solution was not necessarily the answer to tomorrow’s problem.
Perhaps that is the final test of truly great leadership. Not whether people wonder how the company can possibly survive without you, but whether you have left behind the people, culture and capabilities that allow it to do exactly that.
Jensen Huang built the company the AI revolution needed. Nvidia’s Board now has to find the leader the company itself needs. If it succeeds, Huang’s departure certainly is not the end of Nvidia’s extraordinary story. It could be the beginning of its next one.
