Perched in Akamai’s omniscient command center high above the NAB 2025 show floor—a rare vantage point that renders even the slickest booths below oddly ugly—Faultline was given an unfiltered download from Jon Alexander, Akamai’s SVP of Products. What’s emerging is a radical reimagining of what a delivery platform can be in the AI era. While the conversation began with VPUs (more on this last week), it quickly segued into Akamai becoming an edge inference engine for AI, bringing compute and machine learning right to the network edge. Succinctly, the goal is to shrink AI response times from seconds to milliseconds—which might be a big deal when trying to personalize a live stream on the fly or serve betting odds the…