From hype to hard limits: what Coreweave, Dell, and Intuit just told investors

• CoreWeave shows AI demand is real, but financing and delivery discipline matter as much as growth.

• Dell shows AI hardware demand is strong, yet memory and repricing speed can gate results.

• Intuit shows AI raises “front door” competition, pushing marketing and support costs higher.

AI stops being a cool demo the day it meets a budget, a parts list, and a customer who can switch. That shift is happening now.

Three earnings reports on 26 February 2026 tell one simple story: AI is not one industry. It is a supply chain. It runs from capital (who funds the build), to components (who gets the parts), to customer relationships (who stays the default choice).

Coreweave is the AI construction site

CoreWeave is a “neo-cloud” provider that rents out specialised computing for artificial intelligence, largely built around Nvidia chips. Think of it as a purpose-built power plant for AI workloads.

Its quarter makes one point very clear: demand is not the problem. Cost is the problem, and cost shows up before the profit does.

CoreWeave says it expects capital expenditure (capex, the money spent on chips and data centres) of 30 billion to 35 billion USD in 2026, up from 14.9 billion USD in 2025. It also reports a revenue backlog of 66.8 billion USD at 31 December 2025. That is a big “to-do list” of contracted work.

But a backlog is not cash in the bank. It is a promise that still requires hardware, power, and time. CoreWeave also says its adjusted operating income margin drops to 6% in the December quarter, from 16% a year earlier. Translation: building fast is expensive, and the bill arrives on schedule.

This is why the stock falls after hours despite a revenue beat. Investors do not only ask “how fast are you growing?” They also ask “how are you paying for it, and what happens if money gets pricier?”

If CoreWeave is the construction site, Dell is the truck that brings the materials. And one material matters more than many investors expect: memory.

Dell shows AI can be gated by memory and repricing speed

Dell sells servers, storage, networking gear, and personal computers. In the AI build-out, it sits in the “picks and shovels” layer, shipping the boxes that go into data centres.

Dell’s results show AI demand is strong, but hardware is not a pure volume story. It is a parts story and a pricing story.

Dell says memory chip costs rise as the AI infrastructure build-out tightens supply. Memory here means chips that store data close to the processor, such as DRAM (dynamic random-access memory). AI servers need a lot of it, and shortages can quickly turn “strong demand” into “slow deliveries” or “lower margins”.

Dell’s answer is blunt: reprice. Management says it raises prices for servers and personal computers to offset rising component costs. It also describes initial “sticker shock” from customers, followed by continued buying once customers realise supply matters more than comfort.

The numbers show why investors like that message. Dell reports fourth-quarter revenue of 33.4 billion USD and earnings per share of 3.89 USD. It also guides to roughly 50 billion USD of AI-optimised server revenue for the next fiscal year.

That combination matters. A company can only benefit from a shortage if it can source parts and reprice quickly enough to protect margins.

Now move up the stack: ok, the AI build is real and the parts can be scarce. Next comes the layer investors actually live in every day. Software. Does AI replace it, or does it change how it gets sold?

Intuit shows disruption starts at the “front door”

Intuit is a financial software platform behind TurboTax, QuickBooks, and Credit Karma. It is not a typical software-as-a-service (SaaS) company, but it lives on the same core asset: an ongoing customer relationship.

Intuit’s quarter is the cleanest “disruption reality check” of the three. AI does not automatically kill incumbents. But it can raise the cost of defending the front door: distribution, support, and trust.

Intuit forecasts third-quarter adjusted earnings per share of 12.45 to 12.51 USD, below Bloomberg expectations, as it plans higher marketing spend to attract customers during the United States tax season. This is where AI changes the game.

If AI tools make switching easier, customers may try new assistants that promise a faster answer or a cheaper workflow. Incumbents then have to spend more to stay top of mind, and to keep service quality high when demand spikes.

Intuit also offers the “moat extender” angle. It says it has multi-year deals with Anthropic and OpenAI, and management says more than 3 million clients engage with its AI agents. That suggests AI can strengthen the platform if it keeps ownership of the customer relationship.

So the disruption risk may not show up first as collapsing revenue. It may show up as higher selling and support costs, and a noisier fight for attention.

Risks that are easy to miss

These constraints can worsen quickly.

If financing conditions tighten, balance-sheet-heavy builders like CoreWeave feel it first. If component tightness shifts from memory to another bottleneck, hardware margins can move fast even when demand looks great. If software competition intensifies, you may see the pressure in customer acquisition costs and support intensity well before you see it in reported revenue.

Investor playbook

• For AI infrastructure: does backlog turn into cash flow, or into more debt and capex first?

• For hardware: does pricing keep up with component inflation, and do deliveries stay on schedule?

• For software: who owns the front door, the platform or the AI assistant sitting between platform and user?

• Across the stack: watch constraint signals, capex plans, component availability, and customer acquisition costs.

The cloud meets the concrete

AI is a supply chain, and this week you can hear it creak in three places.

CoreWeave tells you the build-out is alive, but it is capital hungry, and investors price the funding risk immediately. Dell tells you demand can still be gated by parts, especially memory, and winners are the ones who can reprice without losing the order. Intuit tells you AI does not end software, but it makes the customer relationship more expensive to defend. AI may live in the cloud, however, the constraints are very much on the ground.

This material is marketing content and should not be regarded as investment advice. Trading financial instruments carries risks and historic performance is not a guarantee of future results.

The instrument(s) referenced in this content may be issued by a partner, from whom Saxo receives promotional fees, payment or retrocessions. While Saxo may receive compensation from these partnerships, all content is created with the aim of providing clients with valuable information and options.