Jensen Huang had barely left Nvidia before people began demanding that he return. A petition circulated, employees reportedly voiced their objections, and prominent technology executives offered their support. Elon Musk reportedly joined the campaign, ensuring that an already complicated corporate governance dispute would not suffer from a shortage of public attention. By Saturday, November 14, the petition had reportedly approached 100,000 signatures, including people identifying themselves as Nvidia employees, shareholders, and admirers of the company’s former chief executive. Some affiliations remained unverified, but the message was unmistakable. They wanted Jensen Huang back.

It was an extraordinary demonstration of loyalty to a man who had spent thirty-three years building one of the world’s most valuable companies. Huang’s supporters pointed to his achievements, his relationships, and the enterprise he had created. Some questioned whether Nvidia’s board had acted prematurely. Others worried about what his absence might mean for the company’s future. Their concerns were not necessarily unreasonable, although enthusiasm for reinstatement appeared to be running somewhat ahead of the answers.

The controversy followed revelations involving multiple extramarital relationships, including one with a Chinese businesswoman suspected by some American intelligence officials of having connections to Beijing. Those suspicions remained unresolved, and Huang maintained that he had not known of the alleged intelligence connection. Reuters subsequently reported, citing people familiar with the matter, that Huang had misled Nvidia’s top lawyer about the nature of their relationship. If substantiated, that account would raise questions about whether those responsible for protecting the company had received accurate information. A corporate lawyer can assess an uncomfortable situation. It is considerably harder to assess a situation that nobody has accurately described.

None of this establishes that Huang knowingly participated in espionage or resolves whether Nvidia’s board handled his departure appropriately. An allegation is not a finding, and public embarrassment is not a substitute for due process. But public debates about controversial departures rarely remain confined to the evidence. They become arguments about character, gratitude, and whether years of extraordinary accomplishment should count for something when the final chapter becomes uncomfortable.

Of course they should count for something. The more interesting question is what.

I have spent much of my professional life in hospitality, where a manager’s career can be reduced to a single personnel decision with remarkable efficiency. Years of exceptional sales, employees developed, difficult shifts managed, and customers retained can suddenly become secondary to the circumstances surrounding a departure. The organization moves forward, the position is filled, and the person who once seemed essential becomes a name on an old schedule. Corporate America has developed elaborate systems for documenting performance, although it can become remarkably efficient at forgetting what those documents once celebrated.

That does not mean the decision was wrong. Sometimes a departure is justified. Sometimes it is not. Personnel decisions can involve confidential information, competing accounts, and obligations invisible to the employees who worked alongside the person involved. A popular manager may have earned genuine loyalty while also having made decisions that deserve serious scrutiny. Both things can be true, even if they make for an unsatisfying conclusion at the next team meeting.

But the achievements remain. The employees who were developed do not become undeveloped. The customers who were served do not become unserved. The difficult shifts do not somehow become easier in retrospect because the manager who handled them is no longer employed. Years of work do not disappear because the final chapter was uncomfortable. Nor do those accomplishments answer every question about how that chapter ended.

Corporate America, which has developed an impressive vocabulary for performance management, remains surprisingly uncomfortable with the possibility that two seemingly contradictory things can be true. Either the departing leader was extraordinary and therefore deserves another chance, or the departure was justified and everything that preceded it becomes suspect. Neither interpretation is particularly useful.

Huang’s supporters are right to recognize what he accomplished. He helped transform Nvidia from a graphics-chip manufacturer into a central force in artificial intelligence, anticipated technological developments that competitors underestimated, and generated extraordinary shareholder wealth. Those achievements are part of the historical record, not privileges that Nvidia’s board can revoke. A resignation does not retroactively diminish the technology his company developed or the opportunities it created for thousands of employees.

But accomplishment is not absolution.

Employees may remember generosity that outsiders never witnessed. Investors may associate years of extraordinary returns with Huang’s judgment. Customers may worry that losing an influential executive will disrupt important relationships. These are legitimate experiences and concerns, not evidence that his supporters are incapable of independent thought. Loyalty becomes more complicated when it is asked to do the work of evidence. A petition can establish that people want someone reinstated. It cannot establish whether allegations are true, whether a board acted fairly, or whether reinstatement would serve the institution’s interests. Popularity is useful in many professions. It is a less reliable method of conducting an investigation.

I understand the emotional force behind such campaigns. In hospitality, a good general manager can become much more than the person responsible for food costs and labor targets. Employees remember who defended them when a customer became abusive, who adjusted a schedule during a family emergency, and who recognized their potential before anyone else did. They remember the manager who stayed late to help clean up after a difficult shift rather than disappearing into the office to review yesterday’s sales. Those relationships are real, and the loyalty they inspire is often deserved.

But employees do not experience a manager’s entire professional life. They experience the portion that intersects with their own. The person who protected one employee may have treated another unfairly. The manager who inspired extraordinary performance may also have exercised questionable judgment. None of this makes the gratitude dishonest. It makes the assessment incomplete.

The reverse is equally important. Organizations make mistakes. Executives can act hastily, investigations can be inadequate, and internal politics can influence decisions subsequently described in the reassuring language of policy enforcement. The existence of a corporate process is not proof that the process was fair. Anyone who has spent enough time in management knows that a decision can be documented with extraordinary precision and still deserve to be questioned.

The petitioners may be wrong about reinstatement while raising legitimate concerns about the process. Nvidia’s board may have been justified in accepting Huang’s resignation while still owing shareholders a credible explanation of its governance decisions. Unfortunately, nuance is rarely the preferred format for an online petition. There is generally a box for your signature, not a section inviting a thoughtful assessment of competing fiduciary obligations.

Musk’s reported intervention introduced a particularly revealing argument. He suggested that important Nvidia supply-chain arrangements depended on personal understandings with Huang rather than written contracts, warning that some deals could be lost without him. Musk did not identify the arrangements, and his prediction was not evidence that any particular deal had been lost. Nevertheless, it was an extraordinary defense. One of the world’s most sophisticated technology companies was apparently being praised for maintaining relationships so personal that their continued existence might depend on one man. Corporate lawyers, who generally prefer their multibillion-dollar commercial arrangements to survive changes in management, may have found the endorsement less reassuring than intended.

The argument exposed a distinction between institutional loyalty and personal loyalty. Employees may believe that defending a founder is equivalent to defending the company he created. Investors may assume that protecting the executive responsible for extraordinary returns is the same as protecting their investment. But the interests of an individual and an institution do not always align, even when their histories have been inseparable for decades.

On November 12, Nvidia announced that longtime chief financial officer Colette Kress would serve as interim chief executive while retaining her financial responsibilities. Harvey Jones became the company’s first board chair, and directors initiated a search for a permanent successor. The company emphasized continuity and its ability to honor customer commitments. The reinstatement campaign introduced an awkward complication. Nvidia’s new leadership was attempting to establish confidence in the company’s future while some supporters were demanding a return to its past. It is difficult enough to assume responsibility for a multitrillion-dollar enterprise without having the previous occupant’s admirers conduct an informal referendum on whether you should be there.

On Tuesday, November 17, that tension will become particularly visible. Nvidia is scheduled to release its third-quarter fiscal 2027 results at approximately 1:20 p.m. Pacific, followed by an earnings call at 2 p.m. Analysts will examine revenue, margins, customer demand, and the company’s plans for artificial intelligence infrastructure. They will also assess how its leadership communicates without Huang, whose public appearances had become an important part of Nvidia’s relationship with investors.

The financial questions are substantial. Nvidia’s customer-financing arrangements have attracted scrutiny, particularly regarding guarantees and other commitments intended to support demand for its products. Kress, who reportedly questioned aspects of those arrangements before Huang’s departure, must now help explain the company’s financial position. Financial exposure does not become less consequential because the person who approved a strategy was unusually charismatic.

But Tuesday’s call presents another problem, one that no earnings presentation can resolve. If Nvidia reports extraordinary results, Huang’s supporters may credit the founder whose decisions helped produce them. If the numbers disappoint, they may regard the results as evidence of what the company has lost. Either interpretation risks overlooking the thousands of people whose work produced those results, as well as the delay between corporate decisions and their financial consequences. A quarter does not begin when a chief executive takes office, nor does the preceding chief executive’s influence disappear when the company announces a resignation.

The irony is difficult to resist. Nvidia helped convince the world that artificial intelligence could perform increasingly sophisticated work, potentially transforming entire professions. Yet some of the industry’s most influential figures appear deeply concerned about the possibility of one particular human being being replaced. Apparently, technological disruption is considerably more appealing when it happens to someone else’s job.

A strong earnings report will not establish that Huang’s departure was justified. A disappointing one will not establish that it was a mistake. Wall Street has never been especially reluctant to explain a complicated financial development through the actions of one recognizable individual, but Nvidia’s performance will require a more serious assessment than whether each quarterly result supports a preferred account of its former chief executive.

I have seen how memory can complicate these judgments. After a respected manager leaves a restaurant, employees sometimes compare every subsequent decision with how things used to be done. The former manager becomes a standard against which ordinary decisions are measured, often with the considerable advantage of no longer being present to make unpopular ones. Difficult shifts fade, disagreements become less memorable, and moments of generosity acquire greater significance. Memory can be a remarkably accommodating performance evaluator.

But institutions can be equally selective. A company may celebrate a manager’s achievements for years, then describe the departure in language so carefully neutral that one might reasonably wonder whether the person had ever contributed anything at all. The recognition disappears from the conversation, the position is filled, and everyone is encouraged to focus on the future. Discretion may be necessary, particularly when personnel matters involve confidential information. It should not require pretending that a person’s professional history began and ended with the final decision.

Neither nostalgia nor institutional convenience provides a complete account of a career.

A fair assessment requires separating questions that public debate tends to combine. What did Huang accomplish? What conduct is supported by the available evidence? What risks did that conduct create? What did Nvidia’s directors know, and when did they know it? Was the process fair? And what decision now best serves the company? These questions are related, but they are not interchangeable. An extraordinary answer to the first does not eliminate the need to answer the others.

The same standard applies to the board. Directors do not become infallible simply because they possess fiduciary responsibilities. If Nvidia’s board acted appropriately, it should be able to defend the integrity of its process within the limits of confidentiality and legal obligations. If new information undermines that process, directors should be willing to reconsider their conclusions. Independent judgment requires more than resisting public pressure. It also requires recognizing when one’s own judgment may have been wrong.

But perhaps the petition is not entirely about reinstatement. Nvidia has already begun moving forward, and reversing its leadership transition would carry considerable institutional consequences. Some supporters may recognize that Huang is unlikely to return while still believing his departure was handled unfairly. Others may simply refuse to accept that the circumstances of his exit should become the final judgment on thirty-three years of achievement.

There is something deeply human about that impulse. People do not stop remembering kindness because a company announces a resignation. They do not forget opportunities they were given, lessons they learned, or difficult moments when someone stood beside them. Those memories remain meaningful even when the person who inspired them has made serious mistakes. Gratitude does not require innocence, and disappointment does not require forgetting.

I have always believed that professional accomplishments should survive the circumstances of a person’s departure. Years of service, difficult decisions, employees developed, and contributions to an organization’s success do not become meaningless because employment ends under unfavorable circumstances. Those achievements belong to the record. The person who performed the work does not cease to have performed it simply because the organization has decided that the relationship should end.

Nor should those achievements become a permanent defense against accountability. Recognition is not restoration, and neither judgment should erase the other.

That distinction briefly played out at the San Jose Denny’s where Huang and his co-founders discussed Nvidia. After the restaurant removed a plaque commemorating that origin story, Denny’s said it planned to reinstall it to recognize the conversations that helped launch the company. The plaque’s return would preserve a piece of Nvidia’s history without settling the questions surrounding Huang’s departure.

Perhaps the petitioners never expected Jensen Huang to return. Perhaps they simply wanted the world to remember what he had given them before it decided what to make of his departure. There is something admirable in that impulse, even when the proposed remedy is difficult to defend.

A career deserves to be remembered in full.

Sometimes, asking for someone back is simply another way of insisting that what they gave was worth remembering.