Data centers boast green power. They’re quiet on the scrap — and that’s our problem.

The data center boom is rewriting power curves — and burying an uncounted mountain of decommissioned hardware. E‑Scrap News just flagged a telling gap: Iron Mountain’s latest sustainability overview leans hard into energy metrics while saying little about what happens to the servers, batteries, and switchgear that exit their buildings. For operators who live off predictable, well‑documented material flows, that omission isn’t academic. It’s a signal that the sector’s most visible ESG narratives still sidestep the physical waste our industry has to move, secure, and process.

The headline is power; the footnote is hardware

As reported by E‑Scrap News, Iron Mountain’s 2025 Sustainability Overview is “strong on power, quiet on hardware.” That tracks with broader industry messaging: renewable PPAs, PUE improvements, and grid impacts get center stage. What’s missing are basic end‑of‑life facts — annual tonnage of retired IT assets, reuse vs. shred ratios, downstream certifications, and how much lithium or lead from UPS systems is leaving sites. For haulers, MRFs, and ITAD shops, those are the numbers that drive routing, permits, insurance, and pricing. When a top‑tier operator doesn’t quantify hardware flows, it makes capacity planning guesswork just as AI workloads accelerate turnover.

AI shortens refresh cycles — and multiplies risk and opportunity

The AI buildout is pushing dense, high‑value gear into racks faster: GPU trays, liquid‑ready servers, heavier busways, and more complex battery chemistries on the power side. That shifts the profile of outbound material. Expect more mixed‑metal server carcasses with premium fractions, higher volumes of network gear, and a steady stream of lithium‑ion and lead‑acid from backup systems. That means:

• Safety: more lithium on docks raises thermal risk; you need fire‑rated staging, Class D readiness, and DOT‑compliant packaging for transport.
• Security: serial‑level chain of custody, tamper‑evident seals, and auditable handoffs become table stakes, not upsells.
• Throughput: demanufacturing lines must balance value recovery (boards, CPUs, memory, GPUs) with uptime; shredders need maintenance windows that align with after‑hours data center access.

Without transparent reporting from operators, recyclers and haulers are flying blind on volumes and seasonality. That’s how you end up with missed service levels during hot‑swap cycles or underutilized capacity when a refresh slips a quarter.

Contracts, data and margins: tighten the loop or lose the lane

E‑Scrap News’ takeaway — the story is what’s not said — should spur a practical response. If sustainability reports won’t document the waste stream, contracts should. Build data requirements into MSAs: monthly asset counts, weights by category (servers, network, storage, UPS, cables), reuse/redeployment rates, and downstream certification status (R2v3, e‑Stewards, NAID AAA). Tie SLAs to access windows and security clearances so dispatch can plan night and weekend pulls without “gate denied” deadhead miles.

There’s a Scope 3 angle, too. “Waste generated in operations” sits squarely in many corporate greenhouse gas accounts. If data centers want clean ledgers, they need auditable e‑waste reporting from vendors. Haulers and recyclers who can deliver serial‑to‑destination mass balance — with photographic proof at pickup and dock ingress — will win share and defend margin. Those who still bill by the vague “load” will get squeezed when commodity prices dip or when procurement centralizes under risk‑averse ESG teams.