ReportGem ReportGem EN

实时全球研报

Global Sustainability: 2Q‘26 Earnings Preview (from a Decarb & Physical AI Lens)

发布日期: 2026-07-22研究机构: UBS Equities报告页数: 20原文语言: English证据页码: 1

研报英文原文证据摘录

Global Sustainability: 2Q‘26 Earnings Preview (from a Decarb & Physical AI Lens)

Powered by Lab Global22 JulyResearch2026ab Evidence UBS YES

Global Sustainability Equities

Global2Q'26 Earnings Preview (from a Decarb &

Physical AI Lens) Global Sustainability

Shneur Z. Gershuni, CFA

Analyst

shneur.gershuni@ubs.com

The setup for 2Q’26 earnings season +1-212-713 3974

As we enter 2Q'26 earnings season, the debate has shifted from whether AI, Kathryn Graves, PhD

electrification, automation and energy security are investable themes to where the Associate Analyst

physical constraints are becoming binding. In 1Q, the market was still testing whether kathryn.graves@ubs.com

hyperscaler capex, data-center load growth, Physical AI deployment and Energy +1-212-713 4871

Independence had enough duration to support earnings revisions and multiple

expansion. Since then, the evidence has become more concrete: ERCOT is tracking

roughly 410 GW of large-load interconnection requests, with approximately 87% tied Thi

to data centers, while its preliminary long-term forecast points to nearly 368 GW of s is an

demand by 2032 versus an all-time peak of about 85.5 GW in August 2023. The EIA’s ImageMAEO2026 work similarly points to data-center server electricity use becoming a much

larger share of U.S. commercial-sector load over time, with servers rising from an

estimated 7% of commercial electricity consumption in 2025 to 22%-33% by 2050

across cases. The earnings focus is therefore less about proving demand and more about

proving execution. Specifically, we think the focus will be on who has secured power,

who has grid access, who can deliver cooling, who can finance infrastructure and who

can translate backlog into earnings.

What are we listening for during earnings?

本摘录由系统从所标注的 PDF 证据页直接提取并保留英文原文,不做批量翻译;登录后在阅读器切换中文时才按需翻译。

打开研报阅读器