SPXC · SPX Technologies, Inc. · Stock research

Data-center cooling picks-and-shovels. Refreshed 2026-08-16.

Current Thesis

The binary flagged in the July note resolved, and it resolved in the company's favour. On 2026-07-30 SPX Technologies reported Q2 adjusted EPS of $2.02 against $1.85 consensus on revenue of $679.0M against $640.1M, raised the FY26 adjusted EPS guide from $7.75–$8.15 to $8.20–$8.60 (consensus $8.03) and the sales guide from $2.575–$2.645B to $2.705–$2.765B, put HVAC backlog at $919M (+59% organic YoY), lifted the 2026 data-center revenue expectation to $430M and raised the eventual full-production data-center capacity target to $1.1B from $750M, reachable in the second half of 2028. That is the strongest disclosure of this cycle.

What the tape did with it is the more interesting datapoint. Coverage of the call records a 2026-07-30 regular-session close of $199.30 (+6.21%) and $212.86 after hours (+6.8%), with shares advancing again on 2026-07-31. Two weeks later the 2026-08-14 close was $217.32 — 11.8% below the 52-week high of $246.41, with RSI(14) at 49.0 and a three-month return of 8.1%. A $350M upgrade to the stated capacity target, a $0.45 lift to the EPS midpoint and a closed acquisition bought roughly one gap and then flat drift. The operating story is accelerating; what an investor is buying at these levels is the gap between that and a price that has not reclaimed its July high.

The narrative is maturing. Not accelerating — the best print of the cycle on 2026-07-30 failed to produce a new high through 2026-08-14. Not saturated — new company-specific facts are still arriving (Neptronic closed in Q3, capacity target raised) and sell-side targets are still being marked up (Truist to $301, 2026-08-03). Not dead — the structure that broke ahead of the print was reclaimed on it.

Bullish and bearish views on SPX Technologies, Inc.

The model's bull view on SPX Technologies, Inc. (SPXC), in brief: HVAC backlog $919M at Q2 end, +59% organic YoY (2026-07-30 call), against $755.3M reported at Q1 on 2026-04-30 — the order book is building faster than revenue is converting it. The bear view: The price response is the problem. A beat, a raise and a capacity-target increase on 2026-07-30 left the 2026-08-14 close at $217.32, still 11.8% under the 52-week high of $246.41 with RSI at the midline. When better news buys less price, the marginal bid is thinner than the… Both cases follow in full.

Bull Case

• HVAC backlog $919M at Q2 end, +59% organic YoY (2026-07-30 call), against $755.3M reported at Q1 on 2026-04-30 — the order book is building faster than revenue is converting it.
• 2026 data-center revenue expectation set at $430M and full-production data-center capacity raised to $1.1B from $750M, management attributing the throughput gain to lean and flow work on OlympusMAX and Everest lines (2026-07-30 call). The capacity ceiling moved, which is what a picks-and-shovels multiple keys on.
• FY26 adjusted EPS guide $8.20–$8.60 vs $8.03 consensus and sales $2.705–$2.765B vs $2.614B consensus (2026-07-30). Adjusted EBITDA growth of 27% at the midpoint; consolidated segment income +23% to $167.1M in Q2.
• HVAC revenue +27.6% YoY (+18.9% organic) in Q2; D&M segment margin +610bps to roughly 26.5%, which is why the guide raise was not purely a data-center story.
• Neptronic (C$605M / ~US$430M cash, announced 2026-07-23, completed in Q3) contributes roughly $75M of 2026 revenue over about five months and $0.05–$0.06 of 2026 EPS, at low-40s segment income and mid-40s EBITDA margins — a higher-margin components and controls extension inside the same building.
• Targets kept moving after the print: Truist to $301 (2026-08-03), B. Riley to $285 and Wells Fargo to $250 (both 2026-07-31), Oppenheimer $280 (2026-07-20). The 19-analyst average sits near $273.
• Management framed 2027 data-center visibility from hyperscaler customers as something they "feel very good about," with 2026 capex guidance of $135–$165M weighted to the back half funding the expansion.

Bear Case

• The price response is the problem. A beat, a raise and a capacity-target increase on 2026-07-30 left the 2026-08-14 close at $217.32, still 11.8% under the 52-week high of $246.41 with RSI at the midline. When better news buys less price, the marginal bid is thinner than the headline flow suggests.
• HVAC segment margin fell 260bps in Q2 on capacity start-up costs and tariffs. The revenue ramp and the margin ramp are not arriving together, and the start-up drag scales with the $135–$165M capex program.
• The sell-side is fully caught up: consensus Strong Buy, buy-side consensus percentage cited at 85, average target near $273 with the high at $301. Upgrades from here require estimate revisions, not rating changes.
• Valuation still requires the ramp. The raised FY26 adjusted EPS midpoint of $8.40 is measured against a 2026-08-14 close of $217.32, and adjusted EPS excludes acquisition and amortization items — Q2 GAAP diluted EPS was $1.55 against $2.02 adjusted, per the Q2 10-Q summary.
• Backlog at $919M is an order book, not recognised revenue; the 2028 timeline on the $1.1B capacity target puts most of the payoff outside any 2026 estimate.
• No company-dated catalyst between now and late October. A name with a raised guide and a stalled price has roughly ten weeks with nothing of its own to trade on.

Setup & Price Structure

• Reference close 2026-08-14: $217.32. 52-week high $246.41 (-11.8%). RSI(14) 49.0. Three-month return +8.1%.
• The $205 origin of the May–June advance was lost ahead of the print — the 2026-07-30 regular session closed at $199.30 — and the earnings reaction reclaimed it. Post-print price has held above that shelf without extending toward the high.
• The post-earnings range therefore sits between the reclaimed $205 area beneath and the $246.41 high above, with price nearer the floor than the ceiling after two weeks of drift.
• Crowding observables: retail-facing backward-looking coverage clustering (Benzinga, 2026-08-13, citing a 27.99% average annual return over five years and 16.13% annualized outperformance); three target raises in the fifteen sessions after the print; no earnings date inside the next 30 days, so the usual pre-print positioning pressure is absent. The recent-filings record for this window shows no insider transactions.
• RSI at 49.0 two weeks after a beat-and-raise is a neutral momentum reading on maximal fundamental news — the observable that dates the maturing label.

Catalyst Calendar (next 30 days)

• 2026-08-16 → 2026-09-15 — no company-dated event scheduled. Q3 2026 results are not on the calendar; the comparable print landed in late October.
• ~2026-08-26 (est.) — NVIDIA fiscal Q2 results. Second-order: the data-center capex tone that reprices the cooling complex.
• 2026-09-30 — Q3 close, the first quarter carrying Neptronic (acquired 2026-07-23) for most of the period.
• ~2026-10-21 (est.) — Vertiv Q3 results, the closest listed cooling comparable.
• ~2026-10-29 (est.) — SPXC Q3 2026 results: first test of the $919M backlog, the $430M 2026 data-center revenue path and the HVAC margin after the Q2 260bps decline.

What Would Change Our Mind

The structure that matters now is the earnings gap itself. The 2026-07-30 release moved price from a broken shelf to a reclaimed one; giving that back would mean the market has fully unwound a raise that added $0.45 to the EPS midpoint and $350M to the capacity target. A weekly close below $199 marks that unwind — the level is the pre-guidance regular-session close on the day of the print.

Secondary conditions, each observable rather than interpretive: a Q3 print (~2026-10-29, est.) that shows HVAC backlog under $919M without a matching revenue step-up, or that trims the $430M 2026 data-center revenue figure; a second consecutive quarter of HVAC segment margin decline, which would recast the capacity build as a margin problem rather than a timing one; a hyperscaler guiding 2027 capex flat or lower, which removes the demand assumption behind the $1.1B capacity target. On the other side, a weekly close back above $246.41 on volume would date a move from maturing back toward accelerating and would say the raise finally got paid.

Correlation Notes

• Trades as a second-order derivative of hyperscaler capex alongside Vertiv and Eaton; group re-ratings can move it independently of its own backlog, which cuts both ways given the $919M order book already disclosed.
• HVAC comparables (Modine, Carrier, Trane) share the tariff and input-cost exposure that took 260bps out of the Q2 HVAC margin — a steel, aluminium or copper move reads through to the same line.
• Neptronic adds a Canadian cost base in Montreal, so USD/CAD becomes a small, new input to reported margin from Q3 onward.
• The name's beta to industrials-wide risk episodes has been demonstrated: the July drawdown came from multiple compression and a macro flush, with no company datapoint attached.