November Earnings Radar: Can AI Demand Become Deliverable Infrastructure?
November’s useful earnings question reaches beyond whether a company beats a quarterly estimate. For anyone following land and infrastructure, the stronger test is whether spending intentions are becoming funded, deliverable projects, and whether the economics still work as that spending becomes real.
This edition follows seven companies across the chain from computing and networking to electrical equipment, data center delivery and electricity. Read their results together. A promising sales outlook at one end of the chain does not establish a service date at the other.
Where these companies fit in the ecosystem
A data center depends on several connected businesses: power producers supply electricity; utilities deliver service through grid infrastructure; equipment suppliers make the electrical systems; developers and operators provide the facilities; chip and networking companies equip the computers inside; and cloud and AI providers use or sell that computing capacity. Companies can span several roles. This watchlist samples different layers, so their results answer different questions.
A hyperscaler is a large cloud platform; a neocloud is a cloud infrastructure provider often focused on AI computing. Neither term describes every company below. We start each company’s section with what it actually sells and the part of the system its results illuminate.
November 3: Equipment, computing and networking
Eaton: Are orders turning into deliveries?
What it does and where it fits: Eaton sells electrical power-management equipment and systems, including switchgear, uninterruptible power supplies, and power distribution products used in data centers. Its role is in the electrical infrastructure that gets usable power safely through the facility. It is an equipment supplier rather than a cloud service or data center landlord. Equipment orders and deliveries can signal how construction plans are moving toward installation. [8]
Confirmed: third-quarter 2026 results before the New York Stock Exchange opens on Tuesday, November 3; conference call at 11 a.m. Eastern. [1]
Watch order growth, backlog conversion, delivery schedules, segment margins and working capital. Separate demand for data centers from the company’s other end markets where disclosure permits. A larger backlog can represent valuable visibility; it can also leave the delivery question unresolved.
The constructive signal would be healthy demand accompanied by actual shipment growth and cash conversion. The risk signal would be rising inventory or longer delivery schedules without a clear explanation. For a development project, ask whether the specific equipment needed for energization is available on its schedule. A supplier’s aggregate revenue growth does not reserve a transformer or switchgear slot for a particular site.
AMD: Is computing demand translating into executable deployments?
What it does and where it fits: AMD designs processors and accelerated-computing products. Its EPYC server CPUs handle computing work in servers; its Instinct accelerators support demanding AI and high-performance computing workloads. AMD also supplies related networking and software technology. It sits in the computing-equipment layer inside the data center. Its sales can reflect investment in computing capacity, but do not establish the location or readiness of a particular building. [9]
Confirmed: fiscal third-quarter 2026 results after the market closes on November 3; conference call at 5 p.m. Eastern. [2]
Watch the distinction between announced partnerships, contracted demand, product shipments and recognized revenue. Read any accelerator growth alongside product mix, gross margin, customer concentration and forward supply commentary. If management discusses deployment capacity, identify its unit, location, timeframe and remaining conditions before treating it as a physical development forecast.
The opportunity would be evidence of durable adoption supported by deliveries and a clearer revenue path. The risks include customer timing changes, product transitions and an unfavorable cost mix. For land and power, the useful follow-up is where deployments will be hosted and when those locations can actually take service. Do not convert an accelerator sales figure directly into acres or local megawatts.
Arista Networks: Is the network keeping pace?
What it does and where it fits: Arista supplies networking equipment and software, including Ethernet switches that move data among servers and across networks. In an AI cluster, many computers must exchange information quickly to work together. Arista therefore sits in the networking layer connecting the computing equipment. Networking demand can help reveal whether customers are equipping usable clusters rather than simply purchasing individual chips. [10]
Confirmed: third-quarter 2026 results after U.S. markets close on November 3; conference call at 4:30 p.m. Eastern. [3]
Watch management’s separation of AI networking from broader cloud and enterprise activity. Look for customer breadth, delivery timing, gross margin and evidence that evaluation work is becoming sustained production demand. Avoid treating a new product qualification or customer trial as a completed campus deployment.
The opportunity is stronger networking demand supported by repeat spending and delivery execution. The risk is dependence on a limited set of large customers or an uneven deployment cycle. Read this call alongside AMD’s: computing and networking schedules need to be consistent before inferring that an entire cluster will come online on time.
November 5: Data centers and regulated power
Iron Mountain: How much leasing is becoming operating capacity?
What it does and where it fits: Iron Mountain operates information-management businesses as well as data centers. Its colocation business provides facilities where customers place computing equipment and obtain supporting infrastructure such as power, cooling, and connectivity. In this watchlist, its role is the physical data center capacity layer. Leasing and customer commencements help connect demand for computing with the buildings and infrastructure needed to host it; consolidated results also include other businesses. [11]
Confirmed: third-quarter 2026 results before market hours on Thursday, November 5; conference call at 8:30 a.m. Eastern. [4]
Watch signed leasing, customer commencements, development spending, available funding and the timetable for capacity entering service. Keep the data center business distinct from the consolidated company. Where megawatts appear, preserve the company’s definitions and distinguish leased capacity, capacity under construction and operating capacity.
The opportunity would be leases converting to commenced revenue while capital needs remain manageable. The risk would be a widening gap between commitments and delivery, especially if construction or power schedules move. A lease announcement is useful demand evidence, but it is not proof that all required power and permits are in hand.
Duke Energy: Which load forecasts have become customer commitments?
What it does and where it fits: Duke Energy is an energy holding company with electric utilities and natural gas utilities. Its electric businesses generate power and operate transmission and distribution infrastructure in their service territories. For a data center in the relevant territory, this is the utility-service layer: the electrical system and service arrangements connecting the facility to power. Large-load contracts and grid investment can indicate which proposed developments are becoming credible utility demand. [12]
Confirmed: third-quarter 2026 results at 7 a.m. Eastern on November 5; conference call at 10 a.m. Eastern. [5]
Watch the quality of large-load demand rather than the largest forecast number. Look for signed customer agreements, deposits or other financial protections, staged service dates, necessary generation and transmission additions, and the status of regulatory approvals. A utility-wide load projection cannot establish service availability at a specific parcel.
The opportunity is load growth backed by customers willing to make enforceable commitments. The risk is infrastructure spending running ahead of durable demand or approved cost recovery. For communities, the key question is who pays if an expected customer downsizes, delays or leaves. For developers, it is which milestones make a proposed service date credible.
November 6 and 9: The power economics test
Vistra: How much demand is contracted, and at what economic risk?
What it does and where it fits: Vistra operates electricity-generation and retail electricity businesses. Its generating facilities produce power; its retail businesses sell electricity to customers in the markets they serve. It sits on the energy-supply side of the ecosystem. Demand from AI and data centers can influence power contracting and investment, but selling electricity is different from providing the local wires and connection needed to serve a particular parcel. [13]
Confirmed: third-quarter 2026 results discussion on Friday, November 6, during a conference call beginning at 10 a.m. Eastern. [6]
Watch contracted sales, hedging disclosures, cash flow and the spending needed to support any additional supply. Separate operating performance from noncash valuation movements where the company provides a reconciliation. Treat management’s forward scenarios as scenarios, not completed earnings.
The opportunity would be durable contract economics with credible delivery and funding. Risks include outages, unfavorable costs, counterparty exposure and commitments that depend on future approvals. A generation agreement may support a customer’s power strategy without resolving that customer’s local interconnection or distribution constraints.
Constellation Energy: What remains between a power agreement and delivery?
What it does and where it fits: Constellation produces electricity from a generation fleet that includes nuclear and other technologies, and supplies energy and services to customers through its commercial businesses. Its place in this watchlist is the generation-and-energy-contracting layer. Long-term demand from data centers can support power agreements and supply investment. Those agreements must still be distinguished from the local utility infrastructure that physically connects an individual campus. [14]
Confirmed: third-quarter 2026 earnings conference call on Monday, November 9, at 9 a.m. Eastern. The cited notice confirms the call; it does not specify a separate results-release time. [7]
Watch the terms, delivery periods and remaining conditions of any disclosed large-load agreements. Where management discusses incremental capacity, distinguish existing output, planned improvements and projects requiring approvals or construction. Ask how much spending is needed before the associated cash flow begins.
The opportunity is contracted demand that supports economically sound investment. The risks are execution delays, outages, financing costs and unfavorable contract terms. For a local project, a corporate power procurement announcement should lead to a second set of questions about the service territory, grid connection, delivery path and actual start date.
How to read the month as an analyst
Use three tests across every call. First, evidence of demand: does management identify actual customer commitments, or describe a pipeline? Second, evidence of execution: are shipments, construction and customer commencements moving on the stated timetable? Third, evidence of economics: are margins, cash generation and funding consistent with the expansion plan?
A constructive reading would combine stronger customer commitments with timely delivery and sustainable funding. A cautionary reading would combine larger headline backlogs with deferred commencements, rising costs or weaker financial protections. Mixed signals matter too: good equipment orders paired with slower utility service may indicate a delivery bottleneck rather than weak end demand.
These are conditions to test against the releases, filings and calls. They are not predictions of the reported results. A strong operational quarter also does not establish that a stock is attractively valued; that requires a separate assessment of price, expectations, capital structure and risk.
For landowners, developers and community leaders, the most useful outcome is a sharper distinction between interest, commitment and delivery. Company earnings can improve the regional picture. They cannot replace a parcel-specific review of power, zoning, access, water, environmental constraints and project funding.
Sources
[1] Eaton earnings announcement, October 20, 2026.
[2] AMD earnings announcement, October 6, 2026.
[3] Arista Networks earnings announcement, October 6, 2026.
[4] Iron Mountain earnings announcement, October 15, 2026.
[5] Duke Energy earnings announcement, October 6, 2026.
[6] Vistra earnings announcement, September 29, 2026.
[7] Constellation Energy events calendar and third-quarter call notice.
[8] Eaton, Data center solutions: electrical equipment and power-management systems.
[9] AMD, Data Center Solutions: EPYC processors, Instinct accelerators, and related technology.
[10] Arista Networks, Company Overview: networking equipment and software.
[11] Iron Mountain, Colocation Services: data center facilities and customer infrastructure.
[12] Duke Energy, Investor Relations: electric and gas utility business overview.
[13] Vistra, Corporate website: generation and retail electricity business overview.
[14] Constellation Energy, Business Overview: generation and commercial energy businesses.
Related reading
Ready to Move?
We respond within one business day to developers, landowners, and community leaders.