By Cloud economics are usually built on the assumption that data mostly stays put. AI applications break that assumption, because their data is constantly moving, and the egress bill often shows up only after the architecture is already designed. With real-time customer interactions in the path, that surprise gets expensive fast.
In this interview on TFiR, Danielle Cook, Director of Product Marketing at Akamai, explains how AI is changing workload placement and why egress costs belong in the architecture decision from the start.
Guest: Danielle Cook, Director of Product Marketing at Akamai
Show: The Agentic Enterprise
LinkedIn: Danielle Cook
Watch the discussion on YouTube
Here is what every platform engineer and cloud architect needs to know.
Technical Deep Dive
Q: How do you plan AI workload placement before egress costs surprise you?
Danielle Cook, Director of Product Marketing at Akamai, says teams should design for egress costs upfront instead of discovering the egress bill after the architecture is finished. AI applications keep data constantly moving, so workload placement belongs in the architectural conversation, especially with real-time customer interactions.
“You don’t want to be in a situation where you’re discovering your egress bill after you’ve designed your architecture. You want to make sure you’re accounting for it upfront.” Danielle Cook, Director of Product Marketing, Akamai
Q: Why is AI forcing teams to move compute to the data instead of data to the compute?
Cook explains that teams used to move data to the compute, and AI is increasingly forcing the reverse. The reasons include the size of the data, its sensitivity, regulation, or simply the cost of moving it.
“We used to move data to the compute. Increasingly, however, AI is forcing us to move the compute to the data.” Danielle Cook, Director of Product Marketing, Akamai
Q: What happens when cloud economics and AI architecture make different assumptions about data?
Cook warns that if cloud economics assume data mostly stays put while the AI architecture assumes data is constantly moving, those assumptions collide. The result is an egress bill that is discovered only after the design is done.
“If your cloud economics assumes that data mostly stays put, while your AI architecture assumes that it’s constantly moving, those assumptions are going to collide.” Danielle Cook, Director of Product Marketing, Akamai
Q: What factors decide where an AI workload should run?
According to Cook, workload placement is a decision around latency, data location, regulations, and economics, which includes tokenomics. She argues this decision now belongs in the architectural conversation rather than being left for later.
“Workload placement is becoming a huge decision. And it’s a decision around latency, data location, regulations, and the economics.” Danielle Cook, Director of Product Marketing, Akamai
Resources & Documentation
- Akamai Inference Cloud Platform, the product page for Akamai’s inference cloud platform
***
👇 Click to Read Full Raw Transcript
Swapnil Bhartiya: With token costs, latency, and performance pressures rising, moving AI workloads to the edge is becoming critical, especially since not everyone can run things locally. So what paradigm shifts are you seeing in how technology, hardware, and organizations like Akamai are positioning AI workloads at the edge to cut down on cost and performance impact?
Danielle Cook: So we used to move data to the compute, right? And increasingly, AI is forcing us to move the compute to the data. So whether it’s because of the size of it, sensitivity, regulation, Right? Or simply, it could be just the cost of moving that data. So we’re seeing that workload placement is becoming a huge decision. And it’s a decision around latency, data location, regulations, and the economics, to your point around, you know, tokenomics. And so this absolutely belongs in the architectural conversation that, you know, because AI applications That data, again, is constantly moving. So, if your cloud economics assumes that data is mostly staying put, but that your AI architecture assumes it’s constantly moving, like, those assumptions are going to collide. So, you don’t want to be in a situation where you’re discovering your egress bill, you know, after you’ve designed your architecture, you want to make sure that you’re designing for it upfront. Especially with these real-time customer interactions.
