SPXC · SPX Technologies, Inc. · Stock research

SPX Technologies (NYSE: SPXC) data-center cooling picks-and-shovels. Refreshed 2026-07-19.

SPXC — SPX Technologies, Inc.

Current Thesis

The pullback the June work was waiting for has arrived, and it arrived without a fundamental crack. SPXC closed at $211.73 on 2026-07-17, down roughly 10.3% over the trailing month and about 16% off the $251.08 52-week high set in early July. That drawdown was not driven by anything the company said. Two macro-and-multiple events did the work: a 5.5% single-session drop on 2026-07-01 to $231.72 attributed to valuation reassessment, and a 2026-07-07 industrials-wide flush when Iranian missile strikes on tankers near the Strait of Hormuz spiked crude and revived inflation pricing. Meanwhile the sell-side kept marking targets up Truist to $295 on 2026-07-02, Wolfe Research to $266, a 12-analyst consensus at $272.25 with a Strong Buy skew.

So price now sits directly on the shelf that mattered: the ~$210 area that was the 200-day reclaim and the origin of the May–June breakout from $205 (2026-05-21) to $234 (2026-06-04). Everything about this leg reduces to whether that shelf holds into the 2026-07-30 Q2 print, which is seven trading sessions out. The narrative engine HVAC backlog $755.3M versus $451.3M a year prior, a data-center cooling growth guide hiked from +50% to +70% on the 2026-04-30 call is intact and untested since April. The tape is the variable, and it is testing support into a binary.

Bullish and bearish views on SPX Technologies, Inc.

The model's bull view on SPX Technologies, Inc. (SPXC), in brief: Q1 2026 (reported 2026-04-30): Adj EPS $1.69 vs $1.56 consensus, +22.5% YoY; revenue $566.8M vs $558.5M est, +17.4% YoY. The bear view: Lower high, then a break of trend. The $251.08 print in early July was; the subsequent slide to $211.73 gives up the entire June advance and leaves price at the last defended shelf with momentum pointed down. Earnings sit 7 trading sessions out (2026-07-30, after close, call… Both cases follow in full.

Bull Case

• Q1 2026 (reported 2026-04-30): Adj EPS $1.69 vs $1.56 consensus, +22.5% YoY; revenue $566.8M vs $558.5M est, +17.4% YoY. FY26 guide raised to Adj EPS $7.75–$8.15, revenue $2.575–$2.645B, Adj EBITDA $600–$625M.
• HVAC backlog $755.3M at Q1 vs $451.3M a year earlier +67% reported, +38% organic with management attributing the organic component primarily to data-center demand. Segment revenue +22% YoY.
• Data-center cooling growth guide raised from +50% to +70% YoY on the 2026-04-30 call. That is a within-quarter acceleration of the specific line the multiple is priced on, and it has not been revisited publicly since.
• Marley OlympusMAX Fluid Cooler launched 2026-04-29 modular dry and adiabatic platform aimed at high-density data-center and industrial loads. Shares rose 5.2% on 2026-06-18 when the supporting HVAC capacity expansion was confirmed.
• Capacity is funded and physical: >$100M of manufacturing investment reiterated at the Wells Fargo Industrials & Materials and William Blair conferences on 2026-06-09, anchored by the 459,000 sq ft Madison, AL facility. Framed runway of ~$550M incremental data-center revenue by 2028 against a $3.2B 2029 revenue target.
• Target revisions still rising into the drawdown: Truist $261 → $295 (2026-07-02), Wolfe to $266, JPMorgan Overweight $260 → $270 (2026-05-06), Wells Fargo $225 → $240 (2026-05-01). Analysts marking up while price falls 10% is a divergence that usually resolves one way or the other quickly.
• Forward multiple has compressed with the tape: 25.78x forward versus 40.54x trailing at the 2026-07-17 close. The re-rate is no longer being paid for at the highs.

Bear Case

• Lower high, then a break of trend. The $251.08 print in early July was; the subsequent slide to $211.73 gives up the entire June advance and leaves price at the last defended shelf with momentum pointed down.
• Earnings sit 7 trading sessions out (2026-07-30, after close, call 4:45pm ET). A name that has already surrendered 16% from its high going into a print where the whole thesis rests on one guide line is a binary, and entering a binary on a broken tape is the wrong side of the distribution.
• Valuation still full on the conservative lens: GuruFocus GF Value $178.38 against $211.73 price ~19% above 'fair' even after the drawdown, versus ~30% above in early July. The gap between the $272–$295 target work and the ~$178 intrinsic work remains wide enough that both camps can be wrong.
• Margin drag is live and unresolved. HVAC margins carry start-up costs on the TN, KS and AL capacity ramps, flagged on the 2026-04-30 call. Growth is being bought with near-term margin, and any ramp slippage lands on the exact segment carrying the story.
• Second-order exposure cuts both ways. The 2026-07-07 selloff had nothing to do with SPXC and everything to do with crude and industrials beta a reminder that this name gets marked with the group regardless of its own backlog.
• Segment leadership change: Eric Kaled succeeds the retiring John Swann atop Detection & Measurement effective 2026-08-31 (announced 2026-06-22). Minor relative to HVAC, but it removes continuity in the non-narrative half of the business.

Setup & Price Structure

Price structure has flipped from breakout-extension to shelf-test. The May–June leg ran $205 (2026-05-21) to $251.08 (early July), roughly +22%, on the OlympusMAX and capacity-expansion news flow. The reversal from that high has been orderly rather than panicked no gap-down on company news, two identifiable macro/multiple sessions doing most of the damage but it has been complete, retracing the full advance.

The $205–$212 band is the structural pivot. It is the 2026-05-21 breakout origin, the reclaimed 200-day area, and now the level price is sitting on. A hold here with the print as the catalyst is the constructive path: the shelf survives, the Q2 guide confirms the +70% data-center line, and the $234 → $251 zone comes back into play with $272 consensus above it. The alternative is mechanical a weekly close below $205 fails the breakout entirely and reverts the name to the $171–$205 range it spent the prior year in, with no support of consequence until the mid-$180s where GF Value sits.

Practically, this is a support test into a binary event, which is a poor risk shape for fresh commitment. There is no confirmation available before 2026-07-30 that the shelf will hold, and the print itself can invalidate it in a single session. The disciplined read is to let the print resolve rather than pay for the guess: either the shelf holds through the number and a clean higher low forms, or it does not and the range is the destination. Beta 1.27 means macro can move this independent of anything the company reports.

Theme state: MATURING. The data-center cooling narrative is well past discovery JPMorgan initiated Overweight on the thesis in March 2026, the trade press has run repeated "is the data center story just getting started" pieces, and the sell-side cluster is fully built out at $266–$295. That is not saturation, but it is no longer early, and the marginal buyer now needs a fundamental confirmation rather than a narrative one.

Catalyst Calendar (next 30 days)

• 2026-07-30 (confirmed): Q2 2026 results after U.S. market close; conference call 4:45pm ET with webcast and slides. The gradeable items are (a) whether the +70% FY26 data-center cooling growth guide is reaffirmed, raised or trimmed; (b) HVAC backlog against the $755.3M Q1 mark; (c) HVAC margin commentary on the TN/KS/AL ramp drag; (d) any move to the $7.75–$8.15 FY26 Adj EPS guide.
• 2026-08-31 (announced 2026-06-22): Eric Kaled formally succeeds John Swann as Detection & Measurement segment leader. Scheduled and non-binary; relevant only if paired with segment guidance changes on the July print.
• Ongoing, no fixed date: hyperscaler capex commentary from peer prints and cooling-adjacent guides (VRT, ETN). SPXC re-rates with this group independent of its own backlog, and any peer cooling guide cut lands on SPXC the same session.
• No FDA, regulatory, index, or lock-up events in the window.

What Would Change Our Mind

The thesis breaks on a weekly close below $205 — that fails the 2026-05-21 breakout origin and the reclaimed 200-day, and returns the name to the $171–$205 range where the data-center re-rate had not yet been priced. Below that shelf there is no structural argument left, only a valuation argument that GuruFocus already scores at ~$178.

Secondarily, the 2026-07-30 print carries a fundamental invalidation independent of price: any trim to the +70% FY26 data-center cooling growth guide, HVAC backlog printing below the $755.3M Q1 level, or margin commentary showing the capacity ramp drag widening rather than converging. Any of those breaks the acceleration case even if $205 holds.

The reverse also applies. A Q2 print that reaffirms or raises the data-center line with backlog above $755.3M, followed by a reclaim of the $234 pivot, re-establishes the leg with the $251.08 high and the $272.25 consensus above it. A rejection of the shelf that resolves higher without ever losing $205 on a weekly basis is the highest-quality version of this setup and the one worth waiting for.

Correlation Notes

• Data-center cooling / thermal peers: VRT, ETN, MOD, TT. SPXC trades as the smaller-cap, higher-beta expression of this basket. Peer guide cuts or hyperscaler capex digestion transmit within a session and dominate SPXC's own news flow.
• Hyperscaler capex: MSFT, AMZN, GOOGL, META. Backlog conversion is a lagging function of their build commitments; the correlation runs with a multi-quarter lag on fundamentals but a same-day lag on the multiple.
• Industrials beta: the 2026-07-07 session crude spiking on the Strait of Hormuz tanker attacks, industrials selling off showed SPXC marking with the group at beta 1.27 regardless of company-specific news. Macro-driven drawdowns in this name are not thesis signals and should not be graded as such.
• Theme-tagging caveat: SPXC appears on grid/power screens under a "Transformers" label, but the operative revenue driver is data-center cooling HVAC (Marley OlympusMAX), not power transformers. Tagging it with the transformer complex (HUBB, POWL, EMR) mis-specifies the correlation set the cooling names are the correct read-through.