
Sterling Infrastructure Slides to Technical Support as AI Buildout Plays Screened
Swissquote screens five listed US contractors building AI data centers, from EMCOR to Sterling Infrastructure, now testing major technical support after a threefold share price decline.
- By
- Sophie Lindqvist
- Filed
- Channel
- AI & Compute
- Read
- 3 min read
Sterling Infrastructure (STRL) has lost roughly two-thirds of its share price over seven months, and Swissquote says the stock now sits on heavy technical support. The broker's weekly chart shows STRL trading near its 200-week moving average, historically low readings on the RSI and LMACD oscillators, and the Ichimoku Kumo cloud. The claim is technical, not fundamental. No earnings revision or order book update accompanies it.
Swissquote published the screen as part of a broader argument about where AI capital spending actually lands. US semiconductor stocks consolidated between June and November, a phase the broker calls "very healthy" from a technical analysis perspective. The author also revisited earlier work on how a Democratic Congress might affect AI data center construction in the United States. That risk, Swissquote argues, stays contained because President Trump holds veto power.
Which companies made the list?
The selection is deliberately not about hyperscalers. Swissquote focuses on the firms that physically build data centers, including the buildings and the equipment inside them. The full table covers eight major US companies. Five are publicly traded. Three are privately held, and all three private names construct data center buildings directly.
The listed names and their stated roles:
- EMCOR Group (EME) — installs the electrical and mechanical infrastructure data centers need to operate. Swissquote says it benefits directly from demand for power supply, electrical distribution and cooling systems.
- Quanta Services (PWR) — electrical infrastructure, high-voltage transmission lines and substations. Each new data center needs additional grid capacity, which Swissquote calls a major challenge for AI's continued development.
- Comfort Systems USA (FIX) — HVAC and cooling. Server cooling is flagged as a major technical and energy constraint as computing equipment grows more powerful.
- AECOM (ACM) — operates upstream in engineering, design and management of major projects.
- Sterling Infrastructure (STRL) — site preparation, earthworks and the infrastructure required for new facilities.
Swissquote frames these companies as "a genuine barometer of the health of the AI infrastructure construction supply chain." That is an assertion, not a measured correlation. The piece offers no revenue exposure figures, backlog data or margin analysis to back it.
What does the trade actually expose investors to?
The stated appeal is moving beyond conventional exposure to semiconductor makers and hyperscalers such as Microsoft, Amazon and Alphabet. Those giants invest heavily in infrastructure. The contractors, in Swissquote's reading, capture the physical spending that turns those projects into concrete and steel.
The broker itself supplies the counterargument. None of the five listed companies is a pure-play AI data center operator. Their activities are diversified. Valuations also depend on energy infrastructure, commercial buildings and major industrial projects. An investor buying STRL or EME is buying a construction and engineering business first, and an AI theme second.
What could break the thesis?
Swissquote identifies the main risk as a slowdown in hyperscaler investment. Local constraints rank close behind. The piece names access to electricity and water, building permits, and the environmental acceptability of new projects as the operational bottlenecks to monitor. These are real constraints on the US buildout, and they sit outside the contractors' control.
The technical picture on STRL adds a separate layer. A stock that fell by a factor of three in seven months may be finding support. It may also simply be pricing in slower growth. The broker's own chart shows the damage plainly; the support levels tell you where sellers have historically paused, not where demand must reappear.
Swissquote attached its standard disclaimer. The content is for information only, is not investment advice, and does not come from the firm's research department. The views belong to the consultant.
For the sector itself, the screen points to a testable proposition: if hyperscaler capex holds, the contractors that pour the foundations and wire the buildings should show it in backlogs first. If grid connection queues and permitting delays persist, that same backlog will reveal the strain.
Source: tradingview.com
More from Sophie Lindqvist
Show full bio
News editor covering money, funding and market shifts around Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.
1,057 articles
Related articles
google-s-15b-malaysia-build-meets-u-s-data-center-political-risk
Google's $15B Malaysia Build Meets U.S. Data Center Political Risk
doubts-over-a-i-spending-send-tech-stocks-lower-nyt-reports-78d0eda3
Doubts Over A.I. Spending Send Tech Stocks Lower, NYT Reports
nasdaq-and-s-p-500-close-at-record-highs-on-ai-chip-and-nuclear-demand-7f9cb330
Nasdaq and S&P 500 Close at Record Highs on AI Chip and Nuclear Demand
tsmc-earnings-loom-as-ai-demand-litmus-test-for-markets-10de3d1a
TSMC Earnings Loom as AI Demand Litmus Test for Markets



