There’s a new sheriff in the data center. Nvidia’s quarterly data-center revenue has topped Intel’s for the first time, quadrupling since 1Q20 owing to growing sales of its AI accelerators. Over the same period, Intel’s revenue from the segment sank 40%. Nvidia has benefited from investments begun years ago to adapt its GPU technology to tasks other than real-time graphics. Among these tasks was AI processing, which emerged as an era-defining technology. The company offers a range of AI products, including the fastest chips for AI inference and training. Once a computing juggernaut, Intel is trying to recover from years of substandard execution that led to its decline. At the same time, server processors have come to play a secondary role to AI processors in AI training systems. Both are essential, but the AI chips handle most of the heavy lifting—crunching data and exchanging data with peers—enabling them to capture more value. Even AI inference has largely moved from microprocessors to AI accelerators. Accompanying this shift in technology has been a shift in value; industry revenue from AI chips is rising past that of general-purpose processors. Both Intel and Nvidia have indicated whole-company revenue will grow sequentially this quarter, but Intel’s 4.5% guidance pales next to Nvidia’s rocketing 53% outlook. AMD, meanwhile, is guiding to a flat quarter. Its decline this past quarter ended a growth streak marked by taking market share from Intel. Spending mightily to surpass Nvidia in AI processing, to defend against AMD in server processing, and to catch up with TSMC in fab technology, Intel has been losing money. Even if these investments yield technical success, their revenue impact will take years to be felt. In the near term, Intel may return to growth, but Sheriff Nvidia won’t soon have to tender its badge.