On the AI edge

In their recent report, “Edge AI Survey: As Inference Expands Beyond the Hyperscalers, New Leaders Begin to Emerge,” Morgan Stanley & Co. analysts offered the following observation: “AI inference is expanding beyond hyperscalers, creating new opportunities for specialized edge infrastructure providers.”

To better grasp the implications of this assertion, it’s necessary to understand the language surrounding the basics of AI deployment. To simplify, the process of developing and deploying an AI model has two basic stages, training and inference. In the training stage, developers use one data set to get the model to do what it’s supposed to do—find the necessary patterns and make correct decisions, etc. In the inference stage, the model is turned loose on new, never-before-seen data to see if it can do what it did in training in the “real world.”

The computationally intense nature of this process is why the initial focus of the AI boom was largely on the so-called hyperscalers—companies (like Aphabet, Amazon, Microsoft, and Meta ) that had the resources to develop the massive cloud-computing datacenters necessary to perform these tasks.

While this remains a fairly exclusive club, the analysts believe more of the AI inference portion of the process could shift from hyperscalers to companies that have mostly been on the margins—i.e., the edge—of AI development. Their research suggests that while the cloud hyperscalers are still dominant, customers are becoming more selective about where inference occurs, “prioritizing latency, cost, privacy, and operational resiliency.”¹ In other words, enterprise AI consumers appear to want more direct control of the inference process, and in working with companies—the Edge AI providers—that can help them get it.

While stressing that Edge AI is still “early innings,” the analysts highlighted two companies that appear to be well-positioned in the emerging landscape. Cloudflare (NET), which the analysts described as a “clear leader,” hit new record highs last week after selling off in the wake of its May earnings announcement:

The analysts described the second stock, Akami Technologies (AKAM), as a “credible alternative.” The stock surged after its May earnings release, but subsequently retreated more than 30% before stabilizing somewhat over the past three weeks:

The analysts stress the Edge AI adoption cycle will likely be a multi-year process, although their survey data points to the possibility of  “meaningful enterprise adoption” next year.

They also note that their discussions with some of the major players in the space suggest Edge AI will complement centralized-cloud AI rather than replace it, potentially resulting in a “hybrid architecture” where training remains cloud-centric (that is, dominated by the hyperscalers), while inference increasingly occurs closer to users, devices, and data.

Market Mover Update: September WTI crude oil futures (CLU6) closed at $82.48 on Monday—the market’s highest close since June 12—as the US-Iran conflict continued.

Today’s earnings include: Alaska Air (ALK), Danaher (DHR), General Motors (GM), Halliburton (HAL), Hasbro (HAS), KeyCorp (KEY), 3M (MMM), Northrop Grumman (NOC), Novartis (NVS), Synchrony Financial (SYF).

Platform news: E*TRADE clients can now trade fractional shares on select stocks and ETFs. Also, eligible clients can now trade spot cryptocurrencies (Bitcoin, Ethereum, and Solana) in a linked zerohash account.

Click here to log on to your account or learn more about E*TRADE's trading platforms, or follow the Company on X (Twitter), @ETRADE, for useful trading and investing insights.

¹ MorganStanley.com. Edge AI Survey: As Inference Expands Beyond the Hyperscalers, New Leaders Begin to Emerge. 7/16/26.