In a Latitude Media interview, Amazon's energy chief laid out a sophisticated site-selection strategy — picking locations where new load 'benefits the grid,' pledging full cost coverage, and touting a 52% water-efficiency gain. The claims are worth understanding, because the playbook is coming to a town near you.
A Latitude Media interview published July 17 gives an unusually detailed look at how Amazon thinks about where to build data centers. Brandon Oyer, who leads energy and water strategy for AWS, describes a site-selection philosophy that has evolved from "where can we plug in renewables?" to "where does the grid actually need us?" — a framing that, if taken at face value, turns the developer from a burden on the grid into a benefactor.
The interview is worth reading in full, because the specific claims Amazon makes are exactly the ones your planning board will hear when an AWS-affiliated LLC shows up at a hearing. Here's what was said, what checks out, and what's missing.
Oyer says Amazon invested in transmission-system modeling experts with 20+ years of experience and shifted its strategy from finding spots to connect renewables to analyzing where additional grid load would prove most beneficial. This approach, he says, led to data center developments in Jackson, Mississippi and South Bend, Indiana.
What this means in practice: Amazon is choosing locations where utilities have excess capacity or aging infrastructure that needs investment to justify upgrades. From the grid operator's perspective, a large anchor tenant can accelerate infrastructure modernization that benefits everyone. From the community's perspective, the question is whether that benefit is contractually guaranteed or just implied.
The community question: "Beneficial to the grid" and "beneficial to residents" are not the same thing. A utility that gains a 200 MW anchor customer may improve system economics while still socializing transmission costs. Ask: Who pays for the new substation? Is the rate benefit in writing, or in a press release?
This is the most consequential section for communities. Oyer makes three specific claims:
Amazon covers the full cost of electricity consumption and infrastructure improvements — transmission lines, substations, grid upgrades — without passing expenses to other ratepayers.
Amazon signed the White House Ratepayer Protection Pledge (March 2026), committing to fully cover AI data center electricity production costs.
Evidence from two states:
What checks out: The White House pledge is real and public. Rate reductions in Indiana Michigan Power's territory have been reported. Entergy's Mississippi grid investment has been announced.
What's missing: These are developer-sourced claims presented without independent utility or regulatory confirmation. The critical distinction — which the interview doesn't address — is between direct costs (the data center's own interconnection and consumption) and indirect, system-wide costs (capacity-market impacts, transmission overbuild, peak-load socialization). As we detailed in our utility-bill explainer, PJM's independent market monitor found data centers were responsible for 63% of an 833% capacity-price increase — even though each individual facility paid its direct costs in full.
A data center can simultaneously (a) pay more than its direct cost-to-serve and (b) drive up system-wide capacity costs that land on everyone else's bill. Amazon's claims address (a). Your electric bill reflects (b).
What to demand: Don't accept "we pay our own way" as the full answer. Ask your PUC or utility commission: Has the utility filed testimony confirming that this facility's load will not increase residential capacity charges? That's a testable, auditable claim — and it's the one that matters.
Oyer cites Amazon's carbon-free energy portfolio:
These are large numbers. They represent contracted or announced capacity — not necessarily operational generation at the time and place the data center is consuming power. The distinction between a Power Purchase Agreement signed today and electrons flowing to a facility tomorrow is one the industry routinely elides.
Amazon also touts its $500 million investment in X-energy for small modular reactors (SMRs), starting at 80 MW per unit, partnered with Energy Northwest for offtake commitments. Oyer frames this as potentially establishing an industry model for nuclear deployment without ratepayer financial risk.
Community note: SMRs remain pre-commercial. No SMR has delivered power to a U.S. data center. The X-energy investment is a bet on future technology, not a current-state energy source. When a developer tells your planning board the facility will be powered by nuclear, ask: When? And what powers it until then?
The water section contains the most specific operational data:
Regional water replenishment projects cited:
| Region | Investment | Claimed impact |
|---|---|---|
| Mississippi | Precision irrigation tech | Reduce ag water withdrawals by 150M gallons/year |
| Mexico | Infrastructure repair | Replenish 2.5B liters/year |
| Oregon | $235M surface water supply | Amazon uses 5% of capacity; 95% serves region |
Amazon also says it prioritizes switching from potable to recycled water (treated sewage) where possible.
What's credible: Cold-climate and arid-climate designs genuinely differ in water intensity. A facility in Northern Indiana will use less evaporative cooling than one in Phoenix. The 52% WUE improvement is plausible given the industry trend toward air-cooled and hybrid designs.
What's missing: "Water positive" is a corporate-defined metric, not a regulatory standard. It typically means the company funds water-replenishment projects somewhere that offset the volume consumed somewhere else. If your aquifer drops, the fact that Amazon funded an irrigation project in Mississippi doesn't refill it. The relevant question is always local: What is the net water impact on this watershed, this aquifer, this municipal supply?
What to demand: Facility-level water-use disclosures (gallons per day, source, and discharge), not global averages. If the developer cites "water positive," ask: Where is the replenishment, and does it benefit this community's water supply?
The most important thing about this interview is what it doesn't contain: no critical questions, no independent verification, and no community voices offering counterpoints. This is Amazon's narrative, presented on Amazon's terms.
That's not a criticism of the journalism — corporate profile interviews serve a purpose. But communities should recognize the format for what it is: a pre-packaged site-selection pitch. Every claim Oyer makes is the claim your planning board will hear, often verbatim, when the development agreement comes up for a vote.
Oyer's closing line captures the tone: "From the top of Amazon leadership, all the way to the people who are building and monitoring our substations, we're all aligned on the mission: build responsibly, build economically, and strive to be the most responsible partner."
That's a corporate value statement. Your job is to convert it into contractual commitments — CBAs, rate guarantees, water-use caps, decommissioning bonds — before the vote, not after.
If Amazon (or any hyperscaler) is siting in your community, this interview is a useful decoder ring for the pitch you're about to receive:
They'll say they benefit the grid. Ask your utility to confirm, in filed testimony, that the facility will not increase residential rates or capacity charges.
They'll cite clean energy commitments. Ask what percentage is operational today at this facility, versus contracted for future delivery elsewhere.
They'll tout water efficiency. Ask for facility-level daily water consumption, source, and whether "water positive" credits benefit your local watershed.
They'll promise jobs and investment. Ask for a binding community benefit agreement with annual reporting, clawback provisions if commitments aren't met, and a decommissioning bond.
The hyperscalers aren't wrong that well-sited data centers can benefit a community. But "can benefit" and "will benefit" are separated by a contract. Get it in writing.
Model the numbers on the 🛡️ Negotiation Toolkit tab, check the operator's footprint on the 🏢 Data Centers tab, and review Amazon's environmental data on the 📊 Corporate Profiles tab.