In October 2019, the Harvard Business Review published its ranking of the world's hundred best-performing CEOs and placed Jensen Huang, of Nvidia, in first place. He had ranked second the previous year. Next came Marc Benioff, of Salesforce, François-Henri Pinault, of Kering, Richard Templeton, of Texas Instruments, and Ignacio Galán, of Iberdrola.
Given what happened afterward, this list became a good test of what rankings can and cannot measure.
How the list was built
The evaluation combined three scores, and only one was financial:
- Financial performance: total shareholder return, the same return adjusted for industry, and the change in market value over the entire tenure.
- Environmental, social, and governance score from Sustainalytics.
- ESG score from CSRHub, which aggregates hundreds of sources.
That's why much more famous names were left out or far from the top: it was a ranking of measured performance, not of fame. And since the financial criterion considered the entire tenure, it rewarded longevity — which helps explain why founders tended to do well.
What the list could see, and what it couldn't
What it saw: Nvidia's stock had multiplied fourteenfold between the end of 2015 and the end of 2018, with a solid environmental and governance score. A well-deserved award based on the numbers that existed.
What it couldn't see is that the decision that would explain the following years had already been made a decade earlier. Nvidia was founded in 1993 making graphics chips for games. In the 2000s, the company began investing billions so those chips could run general-purpose computing, not just imaging — an expensive bet, with no clear market at the time, that became the foundation for training artificial intelligence models. When demand for AI exploded, the infrastructure had already been ready for years.
None of the ranking's three criteria measured this. Shareholder return and ESG score look backward. The bet that decides the next ten years, by definition, hasn't shown up in any indicator yet.
The practical lesson
Rankings measure the past with precision and the future with zero. This applies to CEO lists, vendor awards, and tool comparisons alike. If you use this type of list to make decisions, you're buying performance that has already happened — which is exactly the part that doesn't repeat on its own.
The question that replaces the ranking is always the same: what bet is this company making today that hasn't shown up in its results yet? It's the only one with a chance of saying something about the coming years. The Nvidia story keeps paying off, and the most recent chapter is in Nvidia and Hugging Face.
Within a company, the applied version of this question is choosing where to invest capacity before the market demands it. That's the conversation we have in AI implementation for businesses: which task is worth preparing for now, measured against your own operation, rather than which tool won an award.
Frequently asked questions
Who was named the world’s best CEO in 2019?
Jensen Huang, cofounder and CEO of Nvidia, topped the ranking of the hundred best-performing CEOs published by Harvard Business Review in October 2019. He had come in second the year before. Marc Benioff of Salesforce took second place.
How did Harvard Business Review choose the best CEOs?
By combining three scores. The financial one factored in total shareholder return, that same return adjusted for the industry, and the change in market value over the entire tenure. The other two were environmental, social, and governance ratings from Sustainalytics and CSRHub. That’s why the ranking didn’t line up with the list of most famous CEOs.
Why did Nvidia become so important for artificial intelligence?
Because back in the 2000s, the company invested in making its graphics cards capable of running general-purpose computation, not just image rendering. It was an expensive bet with no clear market at the time, and it ended up becoming the backbone of AI model training. By the time demand exploded, the infrastructure had already been in place for years.
Can you pick a supplier or technology based on a ranking?
A ranking measures the past with precision and the future with nothing at all. Realized returns and reputation scores describe what has already happened, which is exactly the part that doesn’t repeat on its own. The more useful question is what bet the company is making today that hasn’t shown up in its results yet.