The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Before investing, work out how the company makes money from AI data centers, what it must spend to deliver that revenue, and whether its projects are operating or still planned. Announced demand, signed contracts, connected power, recognized revenue, and profitable customer use are different milestones. Comparing companies without separating them can give a misleading picture.
Start with the business model
“AI data center company” covers businesses with different revenue sources, assets, and capital requirements. A cloud platform sells computing services alongside other products; a data center operator provides facilities and power; an AI cloud provider sells access to computing capacity; an equipment supplier sells components or systems used to build infrastructure. Their margins, investment needs, and measures of demand are not directly interchangeable.
| Business type | What to establish | Why it matters |
|---|---|---|
| Cloud platform | How much AI-related revenue is identified separately, and what infrastructure spending supports it? | Company-wide cloud results may not reveal whether AI services alone cover their costs. |
| Data center operator | Which facilities are leased or operated, how much capacity is occupied, and who pays for power and equipment? | Fixed lease or facility costs can continue even when move-ins or sales prices disappoint. |
| AI cloud provider | How much computing capacity is delivered and used, and what hardware, power, and financing obligations support it? | Planned capacity does not earn revenue until it is available and customers use it. |
| Equipment supplier | What portion of sales depends on data center buildouts, and what commitments or partner exposures accompany that demand? | Orders and supply commitments can create execution and financial exposure without proving end-customer profitability. |
For each company, trace the path from investment to cash: spending and obligations, construction or equipment delivery, customer deployment, recognized revenue, operating costs, and cash collection. Use the company’s own definitions, and mark any company-wide figures that are not specific to AI.
Recommended Free Tools
Check whether demand is real and economically useful
Separate four stages in company statements: expected or announced demand, contracted capacity, revenue recognized in financial statements, and capacity actually being used by customers. A contract can support demand, but it does not on its own establish full deployment, collection, renewal, or a profit.
- Look for utilization, delivered capacity, customer move-ins, and revenue—not only bookings, pipeline, reservations, or announced agreements.
- Ask whether revenue from used capacity covers power, depreciation, equipment, labor, and financing costs.
- Check whether contracts have cancellation rights, minimum-use terms, or other conditions that affect how much revenue is dependable.
- Consider whether demand is concentrated in a few customers and whether those customers have the funding and infrastructure to take delivery.
Microsoft warns that returns on AI investment depend on customer demand and monetization; if demand is misestimated, infrastructure could be underused or impaired. GDS has reported utilization-related movements and impairment associated with lower sales prices and slower move-in. Those disclosures illustrate risks to investigate, not a prediction that another company will have the same outcome.
Verify power and project readiness
A power agreement or capacity target is not the same as power connected to a site, a completed data center, or operating capacity available to customers. For each major project, check what stage has actually been reached.
- Site: Is land or a facility secured, and are required permits in place?
- Utility: Is power contracted, physically connected, and available in the quantity and timeframe the project needs?
- Build: What construction milestones remain, including shell completion, electrical systems, and cooling?
- Delivery: When is equipment expected to arrive, and when does management expect customer-ready capacity?
- Operation: Is capacity live, occupied, and generating recognized revenue?
Compare milestones with subsequent company updates and filings, rather than treating a target date as proof of completion. Microsoft identifies power availability, cost, delays, and outages as risks to expansion. Nebius’s 2026 company update distinguished contracted power from connected-power targets; those targets are management statements, not completed operating results.
Measure capital needs and obligations against cash generation
AI data center growth can require substantial spending before a site produces revenue. Compare capital expenditure with operating cash generation, but also read the notes for financing needs and commitments that may not appear as current-period capex.
- Review debt maturities, interest costs, leases, equipment purchase commitments, guarantees, and other partner-related obligations.
- Check whether financing is committed and available when construction or equipment payments fall due.
- Consider whether obligations are fixed while project delivery or customer revenue remains uncertain.
- Distinguish a company’s disclosed commitments from a sector-wide estimate or a measure of total industry investment.
NVIDIA reported $279 billion in supply and capacity commitments as of July 26, 2026, in its Form 10-Q for the quarter ended that date, and described guarantees and partner-related obligations. This is NVIDIA’s company-reported figure for its disclosed commitments; it is not a measure of sector-wide capex or a forecast of industry investment.
Test customer and counterparty risk
For large customer agreements, look beyond the headline value. A useful review asks who is obligated to pay, what must be delivered first, how long the commitment lasts, and what happens if either party cannot perform.
- Measure customer concentration using revenue or contracted capacity, and identify whether a small number of counterparties drive the growth plan.
- Read disclosed contract length, delivery schedules, cancellation provisions, and customer funding conditions.
- Check whether the counterparty can obtain the land, power, facilities, and capital needed to use the capacity.
- Track delivery milestones separately from total contract value and expected revenue.
NVIDIA describes risks if customers and partners lack capital or infrastructure. Nebius reports arrangements with Microsoft and Meta alongside delivery milestones; those agreements should be evaluated on their disclosed terms and progress, not treated as proof that every planned project is complete or profitable.
Examine costs, margins, and equipment life
Revenue growth alone does not show whether a project earns an adequate return. Check whether power, cooling, depreciation, compute, labor, and financing costs are rising, and whether customer pricing can adjust when costs change. Fixed or falling customer prices can squeeze margins if operating costs increase.
Also consider how quickly the equipment may need replacing or upgrading as AI architectures change. A long-lived facility can still depend on computing equipment whose economics or usefulness shift faster. Microsoft identifies uncertainty in AI service costs and the risk of margin pressure from higher costs or competition. GDS has discussed utility-cost increases and asset impairment in its own China-focused business context; its results are an operator-specific example, not an industry benchmark.
| GDS-reported measure | Reported result | Qualification |
|---|---|---|
| Net revenue | RMB 11,432.3 million in 2025, up 10.8% from 2024 | GDS company result; not an industry growth rate. |
| Utility costs | RMB 3,995.3 million in 2025, up 18.9% from 2024 | GDS company result; not a typical cost level for all operators. |
| Long-lived asset impairment losses | RMB 1,561.2 million in 2025 | GDS said these losses mainly related to lower sales prices and slower move-in at certain data centers with fixed lease terms. |
Assess suppliers and execution capacity
Projects depend on more than land and funding. Check exposure to a limited number of suppliers for electrical equipment, cooling, servers, networking, and construction. Long lead times or few replacement options can delay a project or raise its cost.
- Identify which critical items have long delivery timelines and whether alternatives are qualified.
- Check whether the company or its customers bear the risk of price increases, shortages, or late delivery.
- Compare announced capacity with the company’s record of completing projects on schedule.
- Look for dependencies among suppliers, customers, and financing partners that could compound a disruption.
Applied Digital’s filing describes long-lead equipment and reliance on a limited number of vendors. Microsoft also reports supply constraints affecting components such as semiconductors, networking, power, and cooling equipment. These disclosures make supplier exposure a due-diligence question; they do not establish that every company faces identical constraints.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Compare companies using consistent measures
Use the same reporting period and definitions wherever possible. Not every company reports AI-only revenue, utilization, or capacity, so label unavailable or company-wide measures rather than treating them as directly comparable.
Best Value
- Family farms not data design for people against AI server farms, data center expansion, rural land buyouts, corporate agriculture, and industrial tech development replacing farmland and open space. Rural conservation and anti data center message.
- AI protest design for farmers, land conservation supporters, anti AI activists, sustainability groups, environmental advocates, rural communities, and people opposing server farm construction, power grid strain, and farmland destruction.
- Lightweight, Classic fit, Double-needle sleeve and bottom hem
| Comparison axis | What to record |
|---|---|
| Revenue | Recognized AI or data center revenue and growth; state whether figures are AI-specific or company-wide. |
| Use and delivery | Utilization, delivered capacity, or other clearly defined operating measures; keep planned capacity separate. |
| Profitability | Gross or operating margin and, where disclosed, the costs included in the measure. |
| Investment and cash | Capex, operating cash flow, debt, leases, guarantees, and purchase or supply commitments. |
| Power and projects | Contracted versus connected power, project milestones, and delivery record. |
| Commercial and supply risk | Customer concentration and contract terms, plus critical supplier exposure and alternatives. |
Do not treat bookings, a capacity pipeline, management targets, or contracted power as equivalents of recognized revenue, used capacity, or operating cash flow. If a company does not report an AI-specific measure, say so; do not infer it from an unrelated company-wide figure.
Use company disclosures as evidence, not as a verdict
Company filings and updates can show what management reports, what risks it acknowledges, and which milestones it expects. They cannot by themselves establish future returns or a fair value for a stock. For example, NVIDIA’s Form 10-Q for the quarter ended July 26, 2026 states: “The availability of land, power, shell, and capital is crucial to support the buildout of a full data center inclusive of NVIDIA AI infrastructure by our customers and partners, and any shortage of these or other necessary resources could impact our future revenue and financial performance.” It is a company risk disclosure, not independent confirmation of a particular project’s readiness.
Use the latest filing and update available when making a decision: project plans, reported results, and market conditions can change. Treat company forward-looking targets as targets until later disclosures show what was delivered.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
