AI compute isn’t one thing. It’s two. Under the umbrella of “AI workloads,” training and inference represent distinct computational worlds with different goals, hardware profiles, and economics. They often get …

For decades, compute has scaled faster than memory. Processors can execute more operations every year, but the speed at which data moves in and out of memory has lagged behind. …

The semiconductor industry is changing quickly, especially as it relates to AI. As AI workloads grow ever more demanding, old monolithic chips are giving way to new chiplet-based designs. But …

Artificial intelligence has reshaped the semiconductor industry, driving an endless chase for better performance and efficiency. But as transistor scaling slows and Moore’s Law fades, the gains from smaller nodes …

The memory wall is more of an issue than ever in AI workloads. How will it be fixed? As AI workloads scale, compute performance is increasing far faster than memory …

Is one kind of AI processor likely to reign supreme? GPUs have long been the workhorse behind much of the AI infrastructure buildup. But as AI needs have grown, specialized …

Edge AI silicon is getting much more capable, but how much will edge AI really handle? Artificial intelligence is moving outside massive data centers. As silicon evolves, AI processing is …

The Nvidia DGX Spark is a so-called “supercomputer” built for AI — but what does that mean, and why would you want one?

Semiconductor packaging is rapidly advancing, but how will it impact AI?

There might be movement from vendors like Nvidia and AMD, but supply issues run a whole lot deeper A few short years ago, it would have been hard to predict …