Breitbart Business Digest: It's a Huang-derful Life

Jensen Huang Had to Go—and Now He’s Gone For all the worrying about what Jensen Huang’s departure means for Nvidia, we suspect the people with the most to lose are in China. They have just lost one of their most effective lobbyists in Washington, and finding a replacement may be harder than finding someone to run a chip company. Huang’s fixation on China was always a little puzzling given how many customers elsewhere were waiting to buy his chips. Most businesses would envy the problem of having more demand than they could satisfy. Huang somehow made it sound as though being shut out of China would leave Nvidia with nowhere to turn. There was a deeper story that Huang sold to the Trump administration. He argued that selling chips to China would keep its artificial intelligence industry dependent on American technology. Denying it access would encourage Beijing to develop alternatives, eventually threatening America’s technological leadership. With this argument, Huang managed to elevate his pursuit of Chinese revenue into something resembling a patriotic duty. This argument was always questionable. There is an old line about capitalists selling the communists the rope used to hang them. It’s hardly reassuring to the fellow on the gallows that he made the rope. What’s more, the trouble with the thesis that selling China American chips would keep it dependent on American technology is that that Beijing had no intention of abandoning its own chip industry just because Nvidia remained willing to supply it. Buying American chips while developing Chinese replacements was a perfectly sensible strategy. Huang’s preferred policy would have made the transition more comfortable for China, which is an unusual selling point for an American national security strategy. We’ve seen this play out over and over again. American banks learned this lesson a decade ago. They invested heavily in expanding operations in China, seeing it as a wide-open market for Western finance, only to learn that China’s principal program was to learn from the U.S. and then run its financial operations without American influence. History suggests that China’s strategy was to buy American technology only to copy it and buy time to develop Han chips instead of Huang chips. In the near term, the fall of Huang is a setback for China. As Catherine Thorbecke notes in Bloomberg Opinion, China’s enthusiasm for technological independence ran ahead of its ability to deliver it. Huawei has acknowledged that it cannot meet domestic demand. Losing access to Nvidia before Chinese substitutes are ready would hurt, whatever Beijing might say about welcoming the opportunity. Nvidia After China While Nvidia famously had no real succession plan for Huang, it isn’t crippled without him. There is an effective research and operating bench beneath the famous founder. The software, engineering, and customer relationships will mostly survive his departure. We would be more worried if customers were buying chips because they liked Huang’s speeches or the way he looked in his leather jacket. One big risk is that we do not really know how much of Nvidia’s sales were ultimately destined for China. Some Wall Street analysts estimate that direct sales to China account for around seven percent of the company’s revenue, but that is almost certainly an underestimate of Nvidia’s China exposure. Many sales to non-U.S. entities were likely fronts for sales to China. Some analysts think perhaps a third of foreign sales were rooted in demand from China. There is a more interesting concern for American investors about what follows Huang. Nvidia has helped finance companies that buy its equipment, making its chief executive something of an impresario for the entire AI buildout. His successor will likely be a little more risk-averse and thus less willing to arrange these ambitious deals, slowing American AI. Perhaps some slowing would be healthy. A gentler expansion of AI investment could reduce the risk of a hard landing. There is no requirement that spending accelerate indefinitely for useful computing capacity to keep growing.