Dell's storage revenue accelerated to 26% growth from 8% the quarter before, its servers and networking business grew 122%, and its AI revenue doubled.
Dell has been treated as a way to own the AI cloud build-out. Amit Daryanani's point is that the parts of the business nobody was watching grew faster than the AI part, and that they earn better margins doing it.
"I think enterprises are actively looking at their token costs and saying, unless I want to go bankrupt, I really can't keep spending this kind of money on frontier tokens."
Daryanani is a senior managing director at Evercore ISI who covers the hardware makers, and he was reading the print live, before the company had taken a single question on its call.
I listened to the full segment so you can skip it.
Here are the 5 takeaways that matter.
👤 Guest: Amit Daryanani, senior managing director at Evercore ISI, who covers Dell and Apple among the hardware makers
📰 Published: 1 September 2026 on YouTube (CNBC Television)
🔴 YouTube | 🟣 Apple Podcasts | 🔗 Episode page | ⏱️ 4 min
Key Takeaways
The story of the quarter is that the non-AI business grew faster than the AI business Servers and networking up 122% against AI revenue up 100% Storage accelerated to 26% growth from 8% the quarter before
Enterprises are moving AI work onto their own hardware because paying per token is too expensive Daryanani's version of the customer's reasoning ends with the phrase "unless I want to go bankrupt"
The buyer is no longer only the AI cloud operator; it is the Fortune 500 That mix carries better margins and better cash generation than the AI cloud business does
The question the market has not answered is whether this rate of growth lasts He is watching durability of growth and the availability of supply chain components
Apple is being held as the opposite side of the AI trade, and effectively as cash When semiconductors and AI stocks fall, he said, Apple tends to rise
1. A shock-and-awe quarter
The host opened on a beat across the board, with the stock up about 8.7% after hours, and asked what stood out. Daryanani's answer was that this is now a pattern rather than a single print.
He framed it as a run of quarters, not a surprise. "I mean, listen, this is the second or third quarter in a row where Dell's almost putting the shock in our numbers in terms of how well AI is doing for them."
The headline was not the part he wanted to talk about. "But I think the part that really should stand out to folks is the diversity of growth over here."
2. Growth is no longer just AI
The comparison that makes the quarter unusual is between the AI line and the traditional one. "While AI is the story and AI revenues are up 100% in the quarter, you actually starting to see the traditional business servers and networking, for example, up 122%."
Storage is where the change of pace is clearest. "And then storage, I would call out accelerated pretty dramatically to up 26% versus up eight last quarter."
The mix is the reason he thinks it matters, because the older business earns more. "So you suddenly see diversity of growth. It's not just AI, the traditional cloud stuff, but it actually started to become more and more enterprise driven, which carries a lot better margin profile, a lot better free cash flow profile for Dell as you go forward."
3. Cheaper to run it in-house
The host asked whether the non-AI demand is customers buying ahead of price increases or a genuine spending cycle. Daryanani allowed a little of the first and then made a much larger claim.
He conceded the pull-forward, and sized it as small. He said he was sure there is an element of buying ahead of price increases, and that it is a small element
The bigger narrative he sees is a cost calculation being made by customers who have run the numbers on paying per token. "I think enterprises are actively looking at their token costs and saying, unless I want to go bankrupt, I really can't keep spending this kind of money on frontier tokens." A frontier model is the largest and newest class of AI model, sold by the API call, so the bill scales with use rather than being fixed
The response is to move the work onto hardware the company owns. "And so I have to repatriate workloads back to on prem and build out my compute my network, my storage infrastructure on premise, because it's way cheaper to do it that way than running everything on a frontier model."
He says he is seeing the same pattern at more than one supplier. Daryanani named Dell, Cisco and Pure Storage as companies where it is showing up
4. What the call has to answer
Asked what he would drill into on the earnings call, given open questions about customer concentration and financing, Daryanani reframed who the customer is before saying what he wants to know.
His correction to the framing is about which kind of AI buyer is spending. "Maybe the nuance should be it's not just AI neo cloud driven. It's actually driven by AI for enterprises." A neocloud is a specialist operator that rents out AI computing capacity, as distinct from a company buying hardware for itself
The buyer list now includes the ordinary large corporate. "There's actually the traditional Fortune 100, Fortune 500 companies."
The numbers are not what he would ask about; their persistence is. "What is the durability of this? Right? Is this a cyclical peak number or is this the start of a new era where you can sustain this kind of growth, let's say durability of growth, supply chain component availability are kind of the things that we would be focused on this call."
5. Apple as the anti-AI trade
The host turned to Apple, noting the change of chief executive, a product launch the following week, and a share price acting as a counterweight to the leveraged AI trades, and asked whether that had inflated the valuation.
Daryanani accepted the framing for the near term and put it more bluntly. "It's almost a anti semiconductor anti AI stock right. So when semis and AI sell off Apple tends to do well because it is the de facto cash position." "I think for a lot of investors it's as good as cash right now."
What decides the stock later is the product, not the hedge. "I think eventually what will matter is how does product innovation look like for Apple."
The test he sets for the incoming chief executive is about turning AI into an advantage inside Apple's existing ecosystem. That means both differentiated new products and scaling AI into the services business
Bonus Insights
The share price reaction was immediate. The host said the stock was moving higher in after-hours trading, up about 8.7%
Daryanani thinks the effect runs past Dell to the whole category. "And really what I think this means is you're seeing the entire hardware stack get commoditized over time."
Daryanani's bottom line is that Dell's quarter is evidence of companies buying their own AI infrastructure rather than renting frontier models by the token, and that the open question is whether that rate of growth can be sustained.
Products, Companies & Tools Mentioned
Dell Technologies (AI revenue up 100%, servers and networking up 122%, storage accelerating to 26% from 8%, with the mix shifting toward higher-margin enterprise buyers)
Cisco and Pure Storage (Named as two other suppliers where Daryanani says the same move of workloads onto customer-owned hardware is visible)
Apple (Held as the opposite side of the AI trade and, in his words, a de facto cash position; the incoming chief executive's test is scaling AI into products and services)
Evercore ISI (Daryanani's firm, whose numbers he says Dell has beaten for two or three quarters running)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: