Jensen Huang, NVIDIA’s founder, president, and CEO, speaks at a news event in Taipei, Taiwan, on June 4, 2024. Nvidia has sought to create the perception that its interests and the US government’s are one and the same, writes Paul Pillar. (Shutterstock/JamesonWu1972)
Is What’s Good for Nvidia Good for the United States?
The Jensen Huang scandal has revealed just how much national security thinking has conflated itself with corporate bottom lines.
David Cohen, formerly deputy director of the CIA and undersecretary of the Treasury, argues in the current issue of Foreign Affairs that the United States should not only conduct economic espionage but also share the take with US businesses. Cohen wrote his article before former Nvidia CEO Jensen Huang’s story of sex, security, and semiconductors grabbed headlines, but Huang’s situation is relevant to this question in two respects.
On one hand, the case is a reminder of China’s aggressive and extensive operations to collect economic intelligence. One of Cohen’s premises is that because competitors such as China do this, so should the United States.
But suppose some of the intelligence US agencies collected was relevant to Nvidia’s business. Should those agencies share such sensitive information with a firm whose CEO was having an affair with a Chinese intelligence asset?
The idea that strong US businesses benefit the US economy and US interests overall is generally valid. Cohen writes that “US national security now depends on private-sector innovations,” and that “breakthroughs that are crucial to the country’s power and influence are emanating from private companies.”
Yet, it is easy to overstate this case. A breakthrough’s location is not the same as where its economic benefits are reaped, even without government assistance in transferring knowledge. For example, Silicon Valley’s breakthroughs in information technology have benefited Europe’s economy about as much as America’s.
An example involving China and artificial intelligence (AI) concerns AI models developed within China that are relatively inexpensive to train and use. Many US firms are profitably using such models because they are cheaper than American-made alternatives and still accomplish the purposes for which those firms use AI. No espionage was necessary.
Even accepting the general proposition about broader benefits to US interests of helping US businesses, the practical problems of sharing economic intelligence with private-sector firms are too great to make it advisable. Security is certainly one of those problems, even without philandering CEOs. Careful redaction can lessen the dangers of revealing the sources and methods through which the intelligence was collected, but it does not eliminate the security problem. An adversary that learns the United States has somehow penetrated its operations can make adjustments that render further collection of intelligence on that target more difficult.
A related question concerns distribution of such information within a firm. Within the US government, access to intelligence is governed by security clearances and need-to-know principles. Extending similar standards to a firm would involve extensive government intrusion into a company’s internal communication practices. But without those standards, the shared information would come into the hands of many people with no need to know and no security vetting.
Perhaps the biggest practical problem in the proposed sharing of intelligence with private-sector firms is in deciding which firms should be recipients. Some companies would receive that benefit and others would not. The government would in effect be picking winners and losers. It would be practicing a form of industrial policy through intelligence sharing, with all the reasons industrial policy can fail.
Even when the firms in one line of business seem easily identifiable, aiding those firms would involve government favoritism toward them at the expense of other potential entrants into the same business. Such exclusion of newcomers would go against the goal of fostering innovation and future breakthroughs.
Cohen, to his credit, acknowledges that the inevitable favoritism means the intelligence-sharing “could easily be politicized or slip into outright corruption,” but he suggests this problem can be dealt with through “neutral, well-defined, and clearly enforced rules.” The absence of any indication of exactly what those rules would be overlooks the devil-in-the-details reasons the problem probably would not be overcome.
It is remarkable that Cohen makes such a suggestion during a presidency—which has two years to go—that shows little respect for rules and laws, has blatantly politicized the intelligence community, and is the most corrupt administration in memory.
A further difficulty is in defining a US company. Seventy-three years ago, General Motors head Charles Wilson said at his confirmation hearing to be secretary of defense, “For years I thought that what was good for our country was good for General Motors, and vice versa.” However valid that sentiment may have been in Wilson’s day, it became less so during the subsequent half-century, with globalization and the rise of multinational corporations that must consider the policies and actions of multiple countries for them to operate.
This is at least as true of Wilson’s business, the automobile industry, as it is of other businesses. And how is one to classify companies whose roots and most of whose operations are in the United States but have domiciled headquarters in Ireland or the Cayman Islands to avoid US corporate income taxes?
The most fundamental problem in conflating the interests of a private-sector company with the interests of the nation is that the company is motivated mainly by profit. That motive is not a national interest. It can conflict with the national interest.
The Nvidia case illustrates such conflict. The company sought sales in China and had designed some of its chips specifically for Chinese customers. Its profits were thus hurt by the Biden administration’s controls on exports of advanced chips to China. Jensen Huang lobbied the Trump administration to relax such controls, despite any danger to US leadership in artificial intelligence, which the Trump administration itself declared a goal.
This subject is part of a larger problem, which the Trump administration’s corruption has highlighted, of monied interests diverging from the national interest. Huang cultivated a close relationship with President Donald Trump and contributed $25 million to Trump’s ballroom project. Trump reciprocated not only on chip exports (which could boost his personal financial stake in Nvidia) but also by turning a blind eye, despite repeated warnings, to the security implications of Huang’s affair with the Chinese asset. The situation is worthy of the label of scandal.
Export controls have a long history of policy debates involving conflicting considerations of economics, security, and diplomacy. Reasonable arguments have been made at different times for either more stringent or more relaxed export policies. But no policy should be the product of influence bought by a private-sector firm motivated by profit.
Cohen quotes a reported remark by a CIA case officer in the 1990s—“I’m prepared to give my life for my country, but not for a company”—in stating that the intelligence-sharing proposal would have to overcome “deep-rooted cultures of the agencies themselves.”
But what the officer said is not an expression of some recalcitrant organizational culture. It instead reflects the principle that there is a national interest that is the foundation of patriotism, is the defining context for the proper functioning of government agencies, and is decidedly different from more parochial interests of individuals and corporations.
About the Author: Paul Pillar
Paul R. Pillar retired in 2005 from a 28-year career in the US intelligence community, in which his last position was as the national intelligence officer for the Near East and South Asia. Earlier, he served in a variety of analytical and managerial positions, including as chief of analytic units at the CIA, covering portions of the Near East, the Persian Gulf, and South Asia. His most recent book is Beyond the Water’s Edge: How Partisanship Corrupts US Foreign Policy. He is also a contributing editor for The National Interest.
The post Is What’s Good for Nvidia Good for the United States? appeared first on The National Interest.