SPX Corp (NYSE: SPXC) — Cooling Accelerated; Returns Must Catch Up

Published: 2026-09-11 · Verdict: Accumulate · Entry price: $190 · Price target: $235 · Research confidence: High (88%)

Executive conclusion

Analyst Take

ACCUMULATE below $190; 12–18-month base-case value $235. Investment conviction: medium. Evidence quality: high for reported results, backlog, transaction consideration, and balance-sheet data; medium for management’s demand and capacity estimates; low-to-medium for long-range utilization, acquisition returns, and the durability of current thermal-market economics.

The investment case improved for two independent reasons: the operating numerator rose and the stock-price denominator fell. In the second quarter, SPX reported $679.0 million of revenue, up 22.9%, including 16.9% organic growth. First-half revenue was $1.246 billion. Management raised its 2026 outlook to $2.705–$2.765 billion of revenue, $630–$660 million of adjusted EBITDA, and $8.20–$8.60 of adjusted EPS. On the earnings call, management also raised expected 2026 data-center cooling revenue from approximately $350 million to $430 million and its estimate of annual data-center production capacity at full output from $750 million to approximately $1.1 billion. HVAC backlog reached $918.8 million, 70% above the prior year on a reported basis and approximately 59% higher after excluding acquired backlog. These are material improvements from the evidence available in July. [S1][S2][S6]

At the same time, SPXC closed at $186.95 on September 10, 2026, down 11.7% from the July 17 baseline price of $211.73 and 25.5% below the June intraday high. The price is now only about 4% above its trailing-52-week low. At approximately 50.1 million shares, current equity value is about $9.36 billion. Adding estimated post-Neptronic net debt produces an enterprise value near $10.24 billion, equivalent to 22.3 times the midpoint of management’s adjusted-EPS guidance and 15.9 times the adjusted-EBITDA midpoint. Those remain premium absolute valuations for an equipment manufacturer and serial acquirer, but the earlier record-valuation objection has weakened substantially. [S2][S5][S7]

The central variant is no longer that the market has failed to notice data-center cooling. It has noticed. The differentiated view is narrower: the post-Q2 price reversal may be discounting margin and capital-allocation risks before the evidence shows an order rollover. Current disclosures still show rapidly expanding HVAC revenue and backlog, while AAON, Johnson Controls, and Trane report strong data-center or applied-HVAC demand. The market may therefore be early in treating SPX’s second-quarter margin decline as evidence of demand deterioration. Conversely, investors may still be underestimating how much capital and execution are required to convert a $1.1 billion throughput estimate into acceptable cash returns. [S1][S9][S11][S12]

The counter-case is substantial. HVAC segment margin fell from 25.4% to 22.8% despite 27.6% revenue growth. Management attributed the pressure to factory start-up inefficiencies, tariff and inflation timing, and a difficult prior-year project comparison, but those explanations remain hypotheses until margins recover. First-half cash from continuing operations was $120.2 million against $143.7 million of continuing income; inventories, receivables, and contract assets rose as the company funded growth. First-half acquisitions consumed $439.6 million before Neptronic. Net debt increased from approximately $138 million at year-end 2025 to approximately $447 million at June 27, and Neptronic required another approximately $430 million of cash consideration. Management’s approximately 1.4-times pro-forma leverage estimate remains manageable, but the balance sheet is no longer an unusually underleveraged source of optionality. [S1][S3][S5][S6]

The $235 base case uses analyst—not company—assumptions of approximately $735 million of 2027 EBITDA, $600 million of year-end net debt, 50.5 million diluted shares, and a 17-times enterprise multiple. That arithmetic yields about $235.50 per share. It requires no multiple expansion from the current forward framework, but it does require EBITDA growth, working-capital normalization, and no large additional acquisition. A $130–$150 bear outcome is plausible if 2027 EBITDA remains near $580–$610 million and the multiple falls to 13–14 times. A $280–$305 bull outcome requires EBITDA around $800 million, HVAC margins near 24%–25%, falling debt, and inclusive ROIC approaching the mid-teens.

The near-term decision sequence is clear. Third-quarter orders and backlog must demonstrate that the 2026 revenue step-up is becoming 2027 work rather than exhausting a temporary queue. HVAC margin must recover as identified start-up costs moderate. Cash from operations must catch up with earnings, particularly through inventory and receivable conversion. Neptronic must contribute measurable profit and cash rather than merely adjusted-EPS accretion. The call would weaken materially after two consecutive quarters of declining organic HVAC orders, a material fixed-price cooling charge, full-year adjusted EBITDA below $630 million, persistent sub-23% HVAC margin after start-up costs pass, leverage above 2.5 times without a demonstrably high-return use of funds, or inclusive ROIC below 9%. It would strengthen if 2027 organic growth remains above 8%, HVAC margin exceeds 24%, operating cash conversion exceeds 100%, and inclusive ROIC reaches at least 13%–14%.

Changes since 2026-07-18

The July report’s core operating thesis was confirmed and strengthened. Data-center cooling is a material revenue contributor rather than a promotional association with artificial-intelligence infrastructure. Management’s 2026 estimate rose from approximately $350 million to $430 million, compared with approximately $200 million in 2025 and $150 million in 2024. Estimated full-production capacity rose from $750 million to approximately $1.1 billion. The balance of HVAC also improved from the baseline’s roughly 3% growth estimate to management’s current 5%–6% characterization. [S6]

The earlier valuation conclusion became stale. The stock entered the July report’s $170–$190 accumulation zone even though the adjusted-EPS midpoint rose from approximately $7.95 to $8.40. Forward adjusted P/E therefore declined from roughly 26.6 times to 22.3 times. The previous bear trigger—an operating deterioration accompanying the price decline—had not occurred by September 10. [S2][S7]

The former fortress-balance-sheet description is no longer valid. Year-end 2025 net debt was approximately $137.6 million. At June 27, 2026, debt was $614.7 million and consolidated cash was $168.2 million, implying net debt of $446.5 million; using only cash shown on the continuing balance sheet gives $448.3 million. Neptronic subsequently required approximately $430 million of net cash consideration and was funded with about $340 million of revolver borrowing plus cash. Estimated pro-forma net debt is therefore approximately $877 million, subject to closing adjustments and intervening cash generation. [S1][S5]

The offering arithmetic also requires a precise correction. The base August 2025 deal was 2,659,575 shares at $188, or $500.0 million gross, but the underwriters exercised their 398,936-share option. The actual transaction issued approximately 3.059 million shares, produced $575.0 million gross proceeds, and $551.1 million net proceeds. Thus, the July baseline’s net-proceeds number was correct, while its base share count was incomplete; the audited draft’s attempted reduction of the transaction to only $500 million gross was itself wrong. [S3][S16]

The prior ROIC concern remains open. A consistently calculated trailing accounting ROIC is approximately 12.1%, while a transaction-adjusted estimate after Neptronic is around 11%, depending on acquired EBITDA, tax, and denominator conventions. No evidence of a new SGS-style fixed-price loss was identified, so that prior bear falsifier has not triggered. [S1][S3][S5]

Stock Price Action — Five-Year Event Map

SPXC closed at $186.95 on September 10, 2026. Its trailing-52-week intraday range was approximately $179.62–$251.08, placing the current price about 4.1% above the low and 25.5% below the high. The five-year closing low was $41.90 on April 29, 2022. Despite the recent correction, the stock remains more than four times that trough. Prices below are market facts; causal attributions are analyst interpretations based on contemporaneous filings and announcements. [S7]

The factor model dated September 9 supplies risk context rather than a fundamental explanation. Estimated market exposure was 1.20 and momentum exposure 0.60; liquidity exposure was negative 0.43 and small-size exposure positive 0.30. Residual momentum was slightly negative, residual Sharpe was negative 0.95, and the regression accounted for 48.8% of in-sample return variance over its specified window. No standard errors are supplied, and the sector coefficients are statistical return exposures—not legal classifications, operating segments, or causal evidence. The factor model therefore supports the limited conclusion that SPXC has behaved as an above-market-sensitivity, momentum-exposed security whose recent residual performance weakened. It does not prove crowding, de-crowding, or why the stock fell. [S8]

Verdict: The five-year chart reflects a genuine operating turnaround plus a substantial re-rating. The recent decline is a valuation reset near the bottom of the 52-week range, not yet evidence that the cooling order cycle has broken. The disconfirming fact is that the stock failed to retain its positive Q2 reaction despite higher guidance, indicating that investors are demanding proof of margins, cash, and returns rather than backlog alone.

Business Overview

SPX Technologies is a Delaware corporation with ordinary common stock listed on the New York Stock Exchange. It is not an ADR, MLP, partnership, or K-1 issuer. The company reports two operating segments—HVAC and Detection & Measurement—and generated approximately 80% of 2025 revenue in the United States. It employed roughly 4,700 people and reported no individual customer or commonly controlled customer group accounting for more than 10% of consolidated 2025 revenue. [S3][S7]

HVAC generated $1.518 billion, or 67% of 2025 revenue. Its products fall into four economically distinct groups. Engineered cooling includes Marley cooling towers, Recold fluid coolers and evaporative condensers, and dry or adiabatic heat-rejection products such as OlympusMAX. Engineered air movement and handling includes Cincinnati Fan, TAMCO dampers, Ingénia air handlers, and the Crawford businesses. Heating includes Weil-McLain and Williamson-Thermoflo boilers and Patterson-Kelley commercial systems. Electric heating includes QMark, Berko, Fahrenheat, INDEECO, ASPEQ brands, Thermolec, and now Neptronic’s electric duct heaters, humidification products, controls, valves, and actuators. [S1][S3][S5]

The customer value proposition in engineered cooling is mission-critical but configuration-specific. Data centers, power plants, industrial facilities, and commercial buildings must remove heat reliably. The equipment must satisfy thermal load, water availability, footprint, noise, materials, weather, efficiency, redundancy, and site-integration requirements. A cooling failure can interrupt a customer’s operations, which increases the value of proven performance, engineering support, installation knowledge, replacement parts, and service responsiveness. That supports price and qualification advantages, but it also creates contract risk: customized projects can suffer design revisions, site delays, commodity escalation, warranty claims, and fixed-price cost overruns. SPX’s 2022 experience demonstrates that criticality does not eliminate execution losses. [S3][S14]

Heating economics differ. Boilers and electric heaters serve replacement, renovation, and new-construction demand through distributors, representatives, contractors, and direct channels. Replacement familiarity, dimensions, controls, certifications, and contractor preference can support established brands, but many purchases remain sensitive to price, lead time, code compliance, weather, and distributor availability. This is steadier than a large cooling project but less structurally differentiated.

Detection & Measurement generated $746.9 million, or 33% of 2025 revenue. Radiodetection, Pearpoint, Schonstedt, and Sensors & Software provide underground-locating and ground-penetrating-radar equipment. CUES and ULC Robotics provide pipe inspection, assessment, and rehabilitation technologies. Genfare supplies fare-collection systems. Flash Technology, Sabik, Sealite, and Avlite provide obstruction and navigation lighting. Dielectric provides broadcast equipment, while TCI and KTS/Kranze supply communications and tactical-data-link products. [S3][S13]

These D&M products reduce costly errors or satisfy regulated workflows. Locators help utilities and contractors avoid damaging buried infrastructure. Pipe-inspection systems reduce the need for destructive excavation and support asset-maintenance decisions. Obstruction and marine lighting help customers meet safety requirements. Transit systems collect and account for fares. Communications products support interoperability and program-specific technical needs. The customer often buys a workflow outcome rather than an isolated device, but the strength of that attachment varies: an integrated fare system or qualified defense product is harder to change than a handheld locator purchased through distribution.

The revenue model is predominantly equipment-led and transactional. In 2025, approximately $2.026 billion, or 89.5% of revenue, was recognized at a point in time; approximately $238.9 million, or 10.5%, was recognized over time. Standard equipment contracts commonly run one to three months, while customized projects and certain D&M arrangements extend longer. Remaining performance obligations extending beyond one year were approximately $183.3 million at year-end. These accounting figures do not measure aftermarket revenue: parts and replacement equipment can also be recognized at a point in time. SPX does not disclose a consolidated percentage for recurring service, software, parts, or aftermarket revenue, so the claim that aftermarket is definitively a minority cannot be verified. [S3]

Backlog improves visibility but is not equivalent to recurring revenue or guaranteed demand. HVAC backlog was $918.8 million at June 27, while D&M backlog was $312.4 million. Management stated that certain long-term customer arrangements do not enter backlog until formal purchase orders are received. Those arrangements may improve supplier qualification, forecasting, and allocation, but customers retain timing and volume discretion until orders are placed. Backlog itself can still be affected by cancellation rights, project timing, pricing clauses, and execution requirements that are not fully disclosed. [S1][S6]

Revenue stability is therefore moderate rather than high. Heating replacement, infrastructure maintenance, regulated lighting, installed equipment, and technical service create recurring needs. Cooling, communications, transit, and pipe-inspection projects create order and margin lumpiness. The contrast between 70% reported growth in HVAC backlog and a 14.5% decline in D&M backlog shows that diversification smooths but does not eliminate volatility. [S1]

The business is readily understandable at the operating level: SPX designs, manufactures, integrates, and supports engineered equipment, earns a margin for application expertise and reliable performance, and supplements organic growth through acquisitions. The analytical complexity lies in portfolio heterogeneity and acquisition accounting. HVAC’s cooling scarcity, heating replacement, and air-handling projects do not have identical economics; neither do locators, fare systems, lighting, and defense communications.

Economically valuable assets not fully captured on the balance sheet include internally developed brands, customer qualification, installed-base knowledge, specifications, engineering methods, distributor relationships, field-service capability, regulatory approvals, and the operating routines used to improve acquired businesses. Conversely, purchased franchises are very visible. At June 27, goodwill was $1.234 billion and acquired intangible assets were $1.015 billion, together equal to 95.8% of common equity before Neptronic purchase accounting. The balance sheet therefore both understates internally created franchise value and records how much capital management paid for acquired franchises. [S1]

Verdict: SPX is an understandable engineered-equipment portfolio, not a homogeneous recurring-revenue compounder. HVAC produces two-thirds of revenue and nearly all current acceleration; D&M provides differentiated infrastructure niches but remains project-sensitive. The strongest economic assets are brand, qualification, engineering, installed-base knowledge, and service capability. The disconfirming evidence is the low disclosed over-time revenue share, unquantified aftermarket mix, heterogeneous product economics, and purchased goodwill and intangibles nearly equal to equity.

Industry Dynamics

SPX participates in multiple markets, and no reliable primary-source aggregate market size can be constructed by adding them together. The earlier $4–$5 billion cooling-tower market estimate and claims of a precisely measured three-company global oligopoly are not supported by the current filing record and are omitted. SPX itself says reliable aggregate market-position data are generally unavailable and describes competition from numerous large and small companies. The decision-useful approach is to identify the thermal-system layers, principal scaled participants, and capital-cycle behavior rather than assert unsupported market shares. [S3]

The data-center thermal chain has several profit pools. At the chip and server level, cold plates, immersion systems, fans, and internal loops capture heat. Coolant-distribution units, pumps, and heat exchangers move it through the facility. Air handlers and computer-room units manage air-side loads. Chillers, dry coolers, adiabatic units, evaporative towers, and water systems reject heat outside the computing environment. Controls coordinate the system, while services maintain uptime. SPX’s clearest competitive position is external heat rejection and associated air movement; it does not own every layer.

Marley’s principal scaled heat-rejection competitors include Baltimore Aircoil Company, owned by Amsted Industries, and privately held EVAPCO, alongside regional fabricators and product specialists. The draft incorrectly identified Baltimore Aircoil as an AMETEK business. Johnson Controls and Trane overlap through chillers, air handlers, controls, services, and complete thermal architectures. AAON’s BASX unit competes in data-center air handling and liquid-cooling equipment. Modine participates in data-center cooling systems. Vertiv spans rack-to-grid critical infrastructure. Those companies overlap at the project level but are not interchangeable cooling-tower comparables. [S18][S19]

Liquid cooling does not eliminate thermodynamics, but it can change which supplier captures value. Heat must ultimately leave the computing equipment and facility, yet the final architecture can employ water-cooled chillers and towers, air-cooled chillers, dry coolers, adiabatic systems, or other combinations. Water availability, climate, density, redundancy, efficiency targets, and site constraints determine content. Management’s assertion that liquid cooling preserves an external heat-rejection opportunity is reasonable; the stronger claim that it necessarily preserves cooling-tower content is not. Johnson Controls’ published architectures explicitly include optional towers or dry coolers and a separate zero-water air-cooled design. [S6][S19]

Current demand evidence is unusually strong. SPX’s data-center cooling revenue increased from approximately $150 million in 2024 to $200 million in 2025 and is expected by management to reach approximately $430 million in 2026. HVAC backlog reached $918.8 million. AAON’s June backlog was $1.971 billion, up 98% year over year, with $1.430 billion at BASX. Johnson Controls reported $21.0 billion of backlog, up 32% organically, with data-center and mission-critical demand supporting the increase. Trane reported $12.1 billion of backlog, up 70%, and 50% bookings growth in Americas Commercial HVAC. These figures have different definitions and include non-data-center work, but together they contradict a claim that the broad applied-HVAC order cycle had already collapsed by mid-2026. [S1][S9][S11][S12]

The supply response is equally visible. SPX estimates approximately $1.1 billion of annual data-center capacity at full production. AAON added more than one million square feet of BASX capacity since 2024, described at least $2 billion of revenue capacity, and reported uneven utilization across plants. Trane and Johnson Controls are investing across the thermal stack. BAC and EVAPCO are expanding manufacturing and product offerings. Modine reported rapid data-center sales growth alongside margin pressure and higher capital spending. The evidence supports the conclusion that capital is entering the market; it does not yet prove that supply has exceeded demand. [S6][S9][S10][S12]

This distinction is central to the capital-cycle analysis. During shortage conditions, suppliers can obtain better pricing, favorable mix, high factory utilization, and customer urgency. Capacity announcements then attract labor, machinery, factories, outsourcing, acquisitions, and new products. Returns may remain high if demand grows faster than supply, or fall if lead times normalize and buyers obtain more alternatives. SPX’s current evidence sits between those states: orders are strong, but factory start-up expenses and tariff timing are already suppressing margin. The bear case is not that artificial-intelligence computing disappears. It is that supply catches up before SPX’s new capacity achieves targeted utilization, causing lower price realization and absorption.

Barriers to entry are real but graduated. Large heat-rejection products require thermal and structural engineering, materials and water-treatment knowledge, test capability, fan and drive expertise, field installation, warranties, and application support. Long operating lives make reliability records and parts availability relevant. Heavy or customized products incur freight and site-complexity costs that favor established regional manufacturing. Customer qualification can take time. Patents and proprietary components can improve performance, but the presence of several established suppliers proves that intellectual property is not an impenetrable barrier.

D&M barriers are different. Underground locating equipment can become embedded in training and field procedures. Aviation and navigation lighting must satisfy technical and safety standards. Transit fare systems require procurement, integration, cybersecurity, and continuing support. Communications products require interoperability and program qualification. These factors can reduce switching and protect margins, but public procurement, project funding, and program timing create lumpiness. SPX calls many of these brands market-leading; independent market-share evidence is insufficient to label them monopolies. [S3]

Heating is the most mature part of the portfolio. Demand follows replacement cycles, construction, weather, energy prices, codes, and distributor inventory. Efficiency standards can stimulate upgrades but also require product investment. Brand and contractor familiarity matter, yet customers retain alternatives. Long-run industry profitability should therefore be lower and less scarcity-driven than current data-center heat rejection.

Geographically, SPX is predominantly domestic, with about four-fifths of 2025 revenue generated in the United States. International operations and sales provide incremental opportunity, particularly where industrial, power, infrastructure, and data-center investment rises. Low-cost foreign production is not an existential threat to every product because large customized systems are freight-intensive, locally engineered, and qualification-sensitive. It remains relevant in standardized heaters, electronics, castings, fans, controls, and components. Tariffs can help domestic competitive positioning while simultaneously increasing SPX’s input costs; the net outcome depends on sourcing and price timing. [S1][S3]

Regulation generally supports D&M demand through damage prevention, aviation safety, marine navigation, transit accountability, infrastructure inspection, and defense procurement. Environmental and water constraints can shift thermal architectures rather than simply expand demand for one product. Trade rules, cybersecurity, procurement requirements, emissions standards, and product certification also add cost. No single regulation creates a guaranteed revenue entitlement.

Verdict: The current thermal market combines strong demand, meaningful qualification barriers, and an accelerating supply response. Competition is becoming more intense in the fastest-growing categories, but current primary evidence still shows order growth rather than an industry rollover. The disconfirming evidence for a permanently scarce market is simultaneous capacity expansion, multiple viable heat-rejection architectures, and margin pressure during ramping. The disconfirming evidence for an imminent collapse is record or near-record peer backlog and continued data-center order growth.

Competitive Position

SPX’s competitive position must be evaluated by franchise and by operating outcome. A consolidated segment margin can combine scarcity pricing in cooling, steady replacement demand in boilers, newly acquired air-handling revenue, and factory start-up costs. Calling the whole HVAC segment a wide-moat business would obscure those differences.

Marley is the central cooling franchise. Its economic advantages plausibly arise from a long performance record, engineer and customer familiarity, installed equipment, application expertise, parts and service availability, manufacturing scale, and the ability to support large projects. The customer outcome is reduced performance and execution risk. The financial outcomes expected from a genuine advantage are higher qualification rates, resilient win rates, better price realization, lower selling effort per order, repeat purchases, and profitable aftermarket demand. If those outcomes deteriorate as competitors expand, the current advantage would prove narrower or more cyclical than assumed.

The moat is moderate, not absolute. BAC, EVAPCO, and other providers demonstrate that capable alternatives exist. Large hyperscale and colocation customers possess purchasing power, can standardize designs, and can qualify more than one supplier. Engineering firms and contractors also influence specifications. Long-term arrangements can make SPX a preferred or planned supplier but, according to management, are not purchase orders and do not enter backlog until orders are placed. Customer qualification therefore increases opportunity without guaranteeing volume. [S6][S18]

Products such as OlympusMAX and Everest can create differentiation if proprietary fans, drives, fill, heat exchangers, controls, or factory designs produce more thermal capacity per footprint, lower water consumption, reduced energy use, shorter installation, or lower lifetime cost. Current revenue and backlog prove market relevance; they do not disclose matched bid win rates, warranty cost, price premiums, or customer lifecycle economics. Product differentiation is therefore supported, while moat width remains partly inferred.

The second-quarter margin bridge is the most important current competitive test. HVAC revenue grew 27.6%, including 18.9% organic growth, but segment margin fell from 25.4% to 22.8%. Management’s call attributed approximately 80 basis points each to start-up costs, tariffs, and an unusually favorable prior-year project comparison, with additional inflation pressure. Those figures are approximate and do not form a perfectly isolated bridge. The essential fact is that exceptionally strong demand did not automatically create expanding margin. [S1][S6]

Comfort and electric heating have weaker structural protection. Weil-McLain and other established names benefit from contractor familiarity, replacement compatibility, distributor availability, and code approvals. Those factors can lower customer search and installation risk. They do not remove competition from other boiler and heating suppliers. Switching costs can be moderate in a replacement requiring compatible dimensions or controls and low in a standard new installation.

D&M contains several potentially stronger niches. A utility or contractor standardized on locator equipment incurs training, accessories, calibration, and workflow costs when changing vendors. A transit agency cannot casually replace an integrated fare system. Aviation, marine, and defense products carry certification, safety, and interoperability requirements. The expected financial results are repeat business, stable pricing, lower churn, and attractive margins. However, SPX does not disclose customer retention, installed-base revenue, or product-level ROIC, and D&M backlog declined 14.5% at June 27. These are attractive niche characteristics, not proof of local monopoly economics. [S1][S3]

D&M’s Q2 performance illustrates both quality and lumpiness. Revenue increased 12.9%, segment income increased approximately 43%, and segment margin reached 28.9%. Management said roughly $15 million of high-margin project revenue shifted from the third quarter into the second, accounting for a meaningful portion of the margin increase. Full-year D&M margin guidance of 26.25%–26.75% is therefore a better normalization anchor than the quarterly result. [S1][S2][S6]

Competition across SPX’s businesses occurs through thermal or technical performance, reliability, application engineering, lead time, capacity, service, price, installed-base compatibility, innovation, certification, and project execution. It is not primarily a contest over the cheapest labor input. Price becomes more important when specifications standardize or supply becomes abundant; engineering and reliability matter more when failure costs are high.

Peer comparisons should remain layered. AAON is useful for observing data-center air-side and liquid-cooling demand, but not as a direct proxy for cooling-tower economics. Johnson Controls and Trane illustrate large-system design, chillers, controls, and service, but have broader installed service networks. IDEX, ITT, and AMETEK are useful capital-allocation and niche-industrial benchmarks, not direct product substitutes. BAC and EVAPCO are more direct heat-rejection competitors but disclose less financial information.

SPX’s current competitive evidence is stronger on relevance than on excess returns. Backlog growth, customer expansion, and revenue revisions show that the products are being selected. HVAC’s falling quarterly margin, a roughly 11%–12% inclusive ROIC, powerful customers, multiple qualified competitors, and an expanding supply base prevent a stronger conclusion. [S1][S3][S6]

Verdict: SPX owns above-average competitive assets, particularly Marley and several regulated or workflow-embedded D&M brands. Their mechanisms are qualification, performance history, engineering, installed-base knowledge, service, and scale—not monopoly patents. The strongest confirming evidence is order and backlog conversion. The strongest disconfirming evidence is margin compression during peak growth, customer discretion under long-term arrangements, and inclusive returns that remain below elite niche-industrial levels.

Growth History and Forward Opportunities

Revenue increased from $1.220 billion in 2021 to $2.265 billion in 2025, a compound annual rate of approximately 16.8%. Operating income followed a less linear path: $104.7 million in 2021, $64.8 million in 2022, $221.9 million in 2023, $308.3 million in 2024, and $350.4 million in 2025. The 2022 trough included project losses and impairments; the subsequent recovery reflects both operating repair and a changed acquisition portfolio. [S3][S7][S14]

Growth has been approximately half organic and half acquired in recent years. In 2025, revenue increased 14.2%, consisting of 6.3% organic growth, 7.8% acquisitions, and 0.1% currency. In Q2 2026, growth accelerated to 22.9%, consisting of 16.9% organic growth, 5.8% acquisitions, and 0.2% currency. Current organic quality is therefore materially better than the recent multi-year average, although much of the acceleration comes from data-center cooling. [S1][S3]

Management’s 2026 outlook is $2.705–$2.765 billion of revenue, $630–$660 million of adjusted EBITDA, and $8.20–$8.60 of adjusted EPS. Midpoints imply approximately 20.7% revenue growth and 27% adjusted-EBITDA growth relative to 2025 adjusted EBITDA. The guidance is non-GAAP and SPX says a forward GAAP reconciliation is not practicable because some exclusions cannot be reasonably predicted. [S2]

Data-center cooling is the primary growth opportunity. Management’s disclosed revenue sequence is approximately $150 million in 2024, $200 million in 2025, an original 2026 plan around $300 million, a first-quarter increase to $350 million, and a second-quarter increase to approximately $430 million. The repeated upward revisions are evidence of near-term demand and execution. They are management estimates, not completed revenue or independently guaranteed customer commitments. [S6]

Management expects approximately $1.1 billion of annual data-center revenue capacity at full production, with major improvements targeted through the second half of 2028. The network includes Olathe, Springfield, TAMCO capacity in Tennessee, Madison, Alabama assembly, and new product lines. Some capacity comes from lean flow and throughput improvements rather than greenfield capital alone. That makes the option more capital-efficient, but it does not remove utilization risk.

The gap between $430 million of expected 2026 revenue and $1.1 billion of stated capacity is the central long-range question. Capacity is an engineering estimate of potential throughput, not revenue guidance. Filling it requires additional customer projects, purchase orders, manufacturing labor, supply-chain performance, working capital, and continued competitiveness. Underutilization would pressure absorption and returns even if data-center construction continued growing at a slower rate.

The outlook outside data centers improved modestly. On the Q2 call, management agreed that the balance of HVAC was tracking at approximately 5%–6%, compared with the prior report’s roughly 3% estimate. Heating, air handling, and acquired products contribute to that result, making a clean organic normalization difficult. Investors should test the rate after acquisition effects, price, and weather normalize. [S6]

D&M’s near-term outlook is positive but not comparable with cooling. Q2 organic growth was 12.8%, yet backlog declined to $312.4 million and a high-margin project shifted approximately $15 million into Q2. Full-year segment guidance of $750–$770 million is close to 2025 revenue of $746.9 million, implying that the strong quarter should not be annualized. [S1][S2][S6]

Neptronic provides a strategically coherent adjacent opportunity. It adds humidification, controls, actuators, valves, and electric duct heating to SPX’s thermal portfolio. The acquired company had approximately $75 million of annual revenue and cost about $430 million, or 5.7 times sales. SPX described the EBITDA multiple as modestly above 12 times. Strategic fit can improve cross-selling and engineering breadth; it does not establish that the price earns an adequate return. [S5]

Additional opportunities include replacement of aging utility and municipal infrastructure, regulated locating and lighting demand, pipe inspection and rehabilitation, transit modernization, communications programs, and further bolt-on acquisitions. Future deals should be viewed as uses of capital rather than free growth options, particularly after the leverage reset.

Near-term catalysts include third-quarter order disclosure, further backlog conversion, sequential HVAC margin recovery, operating cash-flow acceleration, initial Neptronic contribution, and evidence that Madison and other facilities are ramping on schedule. Negative catalysts include order deferrals, working-capital persistence, tariff escalation, integration charges, or a revised capacity timetable.

Verdict: The near-term product outlook is strong and better evidenced than in July. Data-center revenue revisions, HVAC backlog, and 16.9% consolidated organic growth establish real momentum. The disconfirming evidence is the large utilization gap to planned capacity, declining D&M backlog, timing-driven D&M margins, and the continued importance of acquisitions. Growth is credible; its post-2027 breadth and capital efficiency remain unproven.

Financial Quality

Multi-year operating record

*Net income in 2021–2023 contains material discontinued-operation effects and is not a clean measure of continuing performance. The table reconciles Company Financials history to SPX’s filings; filing operating income of $350.4 million, rather than an aggregator-derived $352.2 million, is authoritative for 2025. [S1][S3][S7][S14]

Current earnings are above the company’s recent mid-cycle level. Reported operating margin recovered from 4.4% in 2022 to approximately 15.5% in 2024–2025 and 16.3% in the first half of 2026. Part of the improvement is structural: portfolio divestitures, acquired higher-margin businesses, operating initiatives, and organic scale. Part reflects recovery from an impaired project year and exceptionally strong thermal demand. It is unsupported to assert that normalized HVAC margin must return to the high teens; that belongs in a bear scenario, not the factual record.

GAAP and adjusted results differ materially. In 2025, GAAP diluted continuing EPS was $5.06 and adjusted EPS was $6.76. In Q2 2026, GAAP diluted continuing EPS was $1.56 and adjusted EPS was $2.02. Q2 adjustments included acquired-intangible amortization, acquisition-related charges, special charges, and other items. Intangible amortization is noncash in the current period, but the associated assets were purchased with prior cash or stock. Acquisition and integration costs also recur economically when acquisitions are a recurring strategy. Adjusted results are useful for comparing operating periods; they should not be treated as identical to owner earnings. [S1][S2][S3]

Stock-based compensation is not the dominant adjustment. Long-term incentive compensation expense was $16.7 million in 2025, less than 1% of revenue. Dilution nevertheless matters because the 2025 offering increased the share base: diluted weighted-average shares rose from 46.5 million in 2021 to 48.5 million in 2025 and 50.6 million in the first half of 2026. Per-share returns must therefore exceed both acquisition spending and net issuance. [S1][S3]

Cash quality was strong in 2025 and less complete in early 2026. Cash from continuing operations was $335.6 million in 2025 against $245.5 million of continuing income. Capital expenditure was $92.1 million, producing $243.5 million of simple free cash flow. Comparable simple free cash flow was approximately $219.9 million in 2023 and $275.1 million in 2024. The 2025 decline despite earnings growth reflected the step-up in capacity spending. [S3]

In the first half of 2026, cash from continuing operations was $120.2 million against $143.7 million of continuing income, and capital expenditure was $39.6 million, producing $80.6 million of simple free cash flow. Inventory increased from $302.2 million at year-end to $374.0 million; receivables and contract assets also increased. Acquisitions contributed to balance-sheet changes, so not every dollar is an organic cash use. The cash-flow statement nevertheless establishes that current earnings have not yet fully converted. [S1]

The proxy’s $469.0 million incentive measure should not be confused with ordinary free cash flow. Starting with $335.6 million of operating cash and $92.1 million of capex gives $243.5 million. Adding acquisition and integration cash costs produces $294.3 million of company-presented adjusted free cash flow. The compensation calculation then excludes $57.3 million of tax payments, $41.7 million of net interest, and $75.7 million of other items to reach $469.0 million. That metric exceeded the maximum performance threshold and produced a 200% payout on the 25%-weighted cash component. It measures selected operating execution, not cash available to equity holders after normal claims. [S4]

Capital intensity is rising. Net property, plant, and equipment increased from $140.5 million in 2023 to $218.0 million in 2024, $290.8 million in 2025, and $324.8 million at June 2026. Capital expenditure increased from $23.9 million in 2023 to $92.1 million in 2025; $62.0 million of 2025 capex related to HVAC capacity. First-half 2026 capex was $39.6 million, including $27.4 million for capacity expansion. At year-end, SPX expected 2026 capex of $135–$165 million. The business remains less physically capital-intensive than basic manufacturing, but the cooling strategy is increasing both fixed assets and working capital. [S1][S3]

ROIC and return quality

A transparent accounting estimate uses trailing GAAP operating income of approximately $399.9 million: $350.4 million in 2025, less $153.2 million in H1 2025, plus $202.7 million in H1 2026. Applying a normalized 22% cash tax rate produces NOPAT near $312 million. Average invested capital—common equity plus debt less consolidated cash—was approximately $2.584 billion between year-end and June. The resulting accounting ROIC is approximately 12.1%. This is an analyst estimate, not a company-reported measure, and Company Financials’ standardized ratio is modestly lower because of methodology differences. [S1][S3][S7]

Adding Neptronic’s approximately $430 million consideration while inferring NOPAT from a transaction multiple modestly above 12 times reduces estimated pro-forma ROIC to roughly 11%–11.3%. The calculation is sensitive to acquired cash, purchase accounting, taxes, integration costs, synergies, and whether beginning, ending, or average capital is used. The decision-useful conclusion is a range: SPX currently appears to earn above an estimated 9%–10% cost of capital, but not by enough to presume that every acquisition creates value.

Excluding goodwill and acquired intangibles from invested capital would produce a much higher return but would ignore the cash paid to acquire current earnings. For a serial acquirer, inclusive capital is the more conservative and economically relevant denominator. At June 27, goodwill and intangible assets totaled $2.250 billion, equal to 95.8% of common equity. Net tangible equity was only about $99 million before Neptronic purchase accounting. This is not an immediate solvency problem; it increases sensitivity to acquisition performance and impairment.

The balance sheet remains liquid but changed materially. June debt was $614.7 million. Cash shown on the continuing balance sheet was $166.4 million, while consolidated cash including $1.8 million in held-for-sale assets was $168.2 million. Net debt was therefore $448.3 million or $446.5 million depending on convention. After Neptronic, estimated net debt is approximately $877 million. Management described pro-forma bank leverage as approximately 1.4 times, below its stated 1.5–2.5-times long-term target range. The $2.025 billion senior facility includes a $1.5 billion revolver and a $500 million term loan maturing in September 2030. [S1][S5][S6][S17]

Off-balance-sheet and economic obligations appear manageable relative to the enterprise but are not zero. Year-end operating-lease liabilities were approximately $75.8 million, with $87.9 million of undiscounted future payments. SPX also disclosed standby letters of credit, surety bonds, ordinary-course warranties, legacy environmental and product matters, and contingent or integration obligations associated with acquisitions. No disclosed individual item appears capable by itself of overwhelming current liquidity. Large fixed-price project commitments remain an economic risk even when they are not balance-sheet debt. [S3]

Revenue recognition, purchase accounting, goodwill testing, estimates of project cost, warranty reserves, and non-GAAP adjustments are the load-bearing accounting judgments. No material accounting-policy change was identified in the latest 10-Q. The principal concern is not an identified GAAP departure; it is that adjusted presentation and acquisition accounting can make economic return quality look stronger than inclusive cash returns.

Verdict: Financial quality is above average but not elite. Growth is reported and organic, margins recovered materially, 2025 cash generation was strong, and leverage remains manageable. The disconfirming evidence is H1 cash conversion below earnings, rising capital intensity, large acquired intangibles, material adjusted-versus-GAAP differences, and inclusive ROIC near 11%–12% rather than the mid-teens.

Capital Allocation

SPX’s allocation model is to invest in organic capacity, acquire adjacent engineered-equipment businesses, maintain access to credit, and retain rather than distribute most cash. The strategy has produced faster growth and higher margins, but its value depends on transaction price and realized cash returns.

Major recent transactions include Cincinnati Fan for approximately $145 million, TAMCO for $126 million, ASPEQ for $422 million, Ingénia for $292 million, KTS/Kranze for approximately $340 million including amounts linked to future service, Sigma & Omega for $143 million, Thermolec for $140.2 million, Crawford’s retained businesses for $299.4 million net of acquired cash, and Neptronic for approximately $430 million. First-half 2026 acquisition cash outflow was $439.6 million before Neptronic. [S1][S3][S5][S13]

Management previously described its acquisition history as approximately 10.5–11 times EBITDA before synergies and approximately 9 times after expected synergies. Neptronic was explicitly priced modestly above 12 times EBITDA and at approximately 5.7 times revenue. The higher price may be justified by margins, growth, product complementarity, and cross-selling, but none of those establish value creation until cash returns exceed the cost of capital. [S5][S6]

The acquisition record is strategically coherent but economically under-disclosed. Acquisitions expanded SPX into electric heating, air handling, communications, pipe robotics, and adjacent thermal controls. Segment margins and consolidated earnings improved. However, SPX does not report acquired-business ROIC, cash payback, or complete standalone EBITDA for its major transactions. Consolidated inclusive ROIC around 11%–12% suggests value creation is possible but leaves limited room for overpayment or integration failure.

Organic reinvestment is increasingly important. HVAC capacity accounted for $62.0 million of 2025 capital expenditure and $27.4 million in H1 2026. Lean improvements that release throughput without proportionate capital would be highly valuable; underutilized factories would have the opposite effect. Capital allocation should therefore be judged on capacity utilization and cash ROIC, not merely announced throughput.

The August 2025 equity offering was a rational use of a highly valued currency, but its full size must be stated correctly. The company initially priced 2.660 million shares at $188 and the underwriters exercised an option for another 398,936 shares. Actual issuance was approximately 3.059 million shares, gross proceeds were $575.0 million, and net proceeds were $551.1 million. Issuing equity at a premium valuation reduced balance-sheet risk and funded acquisitions, but shareholders still require acquired returns to exceed the dilution cost. [S3][S16]

The board renewed authority in May 2026 to repurchase up to $100 million per fiscal year, but SPX made no repurchases in the first half of 2026 and none in 2023–2025. The last material open-market repurchase period was 2022, when the company acquired approximately 706,827 shares for $33.7 million, or about $47.68 per share. An unused authorization is not shareholder return; the relevant measure is net share-count reduction after employee issuance and offerings. [S1][S3]

SPX pays no dividend. Given its capacity program, acquisition strategy, and higher post-Neptronic leverage, retaining cash for investment and debt reduction is financially coherent. Dividend coverage is therefore not applicable, but the absence of distributions raises the burden on management to demonstrate attractive reinvestment returns. [S3]

Insider transactions require classification. The evidence reviewed through September 10 identified grants, gifts, and estate-related transfers, not a recent open-market purchase. The latest 10-Q separately said no director or officer adopted, modified, or terminated a Rule 10b5-1 or non-10b5-1 trading arrangement during the quarter. Neither fact proves insiders are bearish; the absence of an open-market buy is weak evidence compared with a discretionary purchase or sale. [S1][S15]

Compensation creates mixed incentives. The 2025 annual incentive weighted adjusted EBITDA at 50%, the compensation-specific adjusted cash-flow measure at 25%, and adjusted revenue at 25%; aggregate corporate performance paid 163.9% of target. Long-term awards include relative-TSR performance shares, options, and time-based restricted shares. CEO Gene Lowe’s reported 2025 compensation was approximately $9.34 million, and ownership guidelines require meaningful equity holdings. [S4]

The structural weakness is the absence of an explicit ROIC or economic-profit metric. Revenue, adjusted EBITDA, and a heavily adjusted cash measure can all increase after acquisitions even when purchase consideration earns only the cost of capital. Relative TSR and executive ownership help align management with per-share outcomes, but neither directly prevents low-return acquisition growth.

Verdict: Capital allocation has been strategically coherent and often opportunistic, particularly the 2025 equity issuance and adjacent thermal acquisitions. The disconfirming evidence is Neptronic’s above-range multiple, higher leverage, dilution, an unused repurchase authorization, limited asset-level return disclosure, and incentives without a direct return-on-capital governor. The next allocation priority should be cash conversion and proof of returns from existing investments rather than transaction volume.

Changes and Headwinds — Last Two Years

The business environment changed materially in four ways: data-center demand accelerated; SPX committed to a larger capacity network; acquisitions broadened the HVAC and communications portfolios; and tariffs, working capital, and factory execution became more visible. [S1][S5][S6]

External and internal causes interact. The external driver is rapid investment in high-density computing infrastructure. Internal actions include developing dry and adiabatic products, expanding factories, improving production flow, qualifying with customers, pricing against tariffs, and buying adjacent air-handling and control businesses. Organic orders demonstrate real external demand; market-share retention, throughput, price-cost performance, and cash conversion will determine whether internal execution creates excess returns.

The capacity plan expanded from approximately $750 million to $1.1 billion of annual data-center revenue capability at full production. Facilities and initiatives include Olathe, Springfield, TAMCO capacity in Tennessee, Madison, Alabama, and the Everest platform. The increase includes productivity and flow improvements, not merely new square footage. That is positive for potential capital efficiency but places greater weight on operational execution. [S6]

The acquisition program changed the product architecture and balance sheet. KTS expanded communications; Sigma & Omega added HVAC products; Thermolec added electric heating; Crawford added air handlers and coils; Neptronic added humidification, controls, valves, actuators, and electric duct heating. These businesses can produce a broader engineered thermal offering, but integration, customer overlap, purchase accounting, and management attention are now larger risks. [S1][S3][S5]

Tariffs and input inflation became measurable headwinds. Management attributed part of Q2 HVAC margin pressure to tariff costs that preceded full price recovery, along with broader inflation. Domestic manufacturing can improve responsiveness and reduce some exposure, but it does not eliminate imported components or commodity costs. Price-cost timing should be monitored through gross margin rather than assumed temporary.

Management change is also underway. John Swann is expected to retire from leadership of D&M, with Eric Kaled identified as successor. Leadership transitions are not evidence of operational trouble, but D&M’s project mix, acquisitions, and high current margin make execution continuity relevant. [S6]

No material accounting-policy change was identified in the current 10-Q. Purchase accounting and intangible amortization became more important because of transaction volume. The filing record reviewed did not reveal a new litigation, environmental, or regulatory matter large enough to alter the investment case. Ordinary-course legacy obligations remain.

The internal-versus-external scorecard for the next year is straightforward. Sustained order growth across several customers would support a broad external opportunity. Higher throughput and recovering margins would demonstrate internal execution. Revenue growth accompanied by persistent inventory, start-up costs, and weak cash conversion would indicate that SPX captured volume without yet proving economic returns.

Verdict: The past two years improved SPX’s opportunity set and product breadth while concentrating execution risk in cooling capacity and acquisition integration. Demand is externally favorable; SPX’s actions have converted it into revenue and backlog. The disconfirming evidence is margin compression, cash absorption, and higher leverage. The next proof point is profitable, cash-generative throughput—not another capacity or acquisition announcement.

Risk Analysis

Several risks interact nonlinearly. A demand slowdown alone would be manageable at current leverage. A slowdown combined with underutilized factories, working-capital persistence, an acquisition impairment, and multiple compression could produce much larger equity losses. Conversely, backlog and liquidity reduce immediate solvency risk but do not protect the multiple if returns disappoint.

A catastrophic investment loss would require more than an ordinary cooling slowdown. A plausible path would combine large fixed-price or product failures, sustained facility underutilization, several overpriced acquisitions, impairment, leverage moving above covenant comfort, restricted credit access, and a severe valuation contraction. No current filing indicates that combination, but the 2022 project loss proves that execution risk is not theoretical. [S1][S14][S17]

A literal total loss is very unlikely. SPX owns profitable operating franchises, has positive adjusted and GAAP operating earnings, diversified customers, and a credit facility extending to 2030. A more realistic catastrophic equity outcome is a decline exceeding 50% if EBITDA falls toward $450–$500 million, net debt remains elevated, and the market applies a 10–12-times multiple. Equity would retain value in that scenario, but the residual claim would be sharply compressed.

Verdict: The dominant risk is the interaction of capacity, acquisitions, cash conversion, and valuation—not near-term insolvency. Backlog lowers immediate revenue risk but cannot remove cancellation, execution, margin, or working-capital risk. The bear case should be tested through orders, cash, and inclusive returns rather than a generic prediction that AI investment ends.

Valuation Discussion

At $186.95 and approximately 50.086 million shares, SPXC’s equity value is about $9.364 billion. June debt of $614.7 million less consolidated cash of $168.2 million produced $446.5 million of pre-Neptronic net debt. Adding approximately $430 million of net cash consideration gives estimated pro-forma net debt of $876.5 million and enterprise value near $10.24 billion. The estimate is subject to acquired cash, closing adjustments, and cash generated after June. [S1][S5][S7]

Current reference multiples are:

• 22.3 times the $8.40 midpoint of adjusted-EPS guidance;
• approximately 33 times estimated trailing GAAP EPS;
• 15.9 times the $645 million midpoint of adjusted-EBITDA guidance;
• 3.7 times expected 2026 revenue on enterprise value; and
• approximately 19 times trailing reported EBITDA before a full annualized Neptronic contribution.

The 15.9-times figure has a timing limitation: enterprise value includes the full Neptronic purchase price while 2026 guidance includes only a post-closing contribution. A fully annualized pro-forma denominator would modestly reduce the multiple, but acquired EBITDA is not disclosed precisely enough to present a false point estimate. [S2][S5]

Adjusted P/E flatters economic earnings by excluding acquired-intangible amortization and recurring acquisition expenses. GAAP P/E treats the amortization as a current expense even though the acquisition cash was spent earlier. Neither is sufficient alone. EV/EBITDA, free cash flow, and inclusive ROIC provide complementary views.

Company Financials’ consistent current-market-cap dataset shows trailing EV/EBITDA of approximately 19 times for IDEX, 23 times for ITT and AMETEK, 26 times for Trane and Johnson Controls, and above 30 times for AAON under its reported-EBITDA convention. These are not like-for-like forward multiples: fiscal calendars, acquisition timing, service mix, and adjustments differ. They establish that investors continue to pay high multiples for differentiated HVAC and niche-industrial franchises, not that SPX deserves any specific exit multiple. [S7]

SPX has less recurring service exposure than Trane or Johnson Controls and less demonstrated return quality than the best niche compounders. It has faster current organic growth than many diversified industrials and a stronger near-term backlog signal. A valuation between conventional diversified industrials and premium thermal-growth peers is therefore economically coherent.

Own-history context also changed. The July report’s record price-to-sales percentile cannot be carried forward after an 11.7% price decline and higher revenue guidance. Current equity value is approximately 3.4 times expected 2026 revenue and enterprise value approximately 3.7 times. Historical percentile claims require a fully recomputed series using consistent enterprise-value and revenue definitions; the stale 93.9th-percentile conclusion is not retained. [S2][S7]

Scenario framework

The base case assumes no major additional acquisition and meaningful working-capital conversion. At 17 times $735 million of EBITDA, less $600 million of debt and divided by 50.5 million shares, equity value is approximately $235.50. At 16.5 times, it is approximately $228.30. The outcome is therefore driven primarily by EBITDA and deleveraging rather than a heroic multiple increase.

The bear case does not require data-center construction to collapse. It requires lead times, price, mix, or utilization to normalize faster than SPX’s fixed-cost base. The bull case requires more than revenue: it requires margin recovery, cash conversion, and inclusive ROIC moving toward 14%–15%.

A reverse-valuation view reaches the same conclusion. At roughly 16 times the current EBITDA guide, the market prices several years of above-market growth, successful integration, and low-20s adjusted margins. It no longer prices uninterrupted 20% organic growth, but it still offers little protection if EBITDA remains near $600 million. The market correctly recognizes the quality and relevance of Marley, the backlog, and the capacity option. It may be too pessimistic if Q2 start-up pressure is temporary. It may remain too optimistic about how easily $1.1 billion of capacity can be filled at current economics.

Verdict: Valuation is no longer the thesis-breaking fact it was in July, but the stock is not statistically cheap. Acceptable returns require 2027 EBITDA above roughly $700 million and falling leverage. The primary sensitivity is cash-generative utilization; the selected terminal multiple is secondary.

Variant Perception

The central investor questions are concrete: What portion of the $430 million data-center estimate is represented by purchase orders? How much 2027 work is already in backlog? What cancellation, escalation, deposit, and fixed-price terms apply? Can HVAC margin return above 24%? How much capex is still required for $1.1 billion of capacity? What cash return will Neptronic earn? Will working capital reverse? Why is ROIC absent from compensation? [S1][S4][S5][S6]

The apparent consensus is that SPX is a high-quality industrial compounder with a scarce data-center thermal franchise, an effective acquisition engine, and a multi-year capacity runway. The strongest bull case is that external heat rejection remains required under both air and liquid architectures; customer and product breadth is expanding; repeated revenue revisions indicate share and demand momentum; lean improvements add throughput efficiently; non-data-center HVAC grows 5%–6%; and cash rapidly reduces post-Neptronic leverage.

The strongest bear case is that a temporary capacity shortage is being treated as a permanent moat. Customers can qualify multiple suppliers, alternative thermal architectures can reduce tower content, competitors are expanding, and SPX’s own HVAC margin fell during peak growth. Acquisitions obscure organic economics, adjusted EPS omits recurring economic costs, capacity and working capital absorb cash, and inclusive ROIC remains only modestly above the estimated cost of capital.

Five load-bearing assumptions determine the result:

1. Data-center orders remain durable through the 2028 ramp. Supporting evidence would be rising 2027 purchase orders and backlog after the 2026 revenue step-up. Disconfirmation would be two quarters of negative organic HVAC orders or customer deferrals. [S1][S6]
2. New capacity earns attractive incremental returns. Supporting evidence would be HVAC margin above 24% while utilization rises. Disconfirmation would be sub-23% margin after identified start-up costs pass or persistent inventory accumulation. [S1][S6]
3. Acquisitions earn more than the cost of capital. Supporting evidence would be pro-forma ROIC approaching 14% as leverage falls. Disconfirmation would be impairment, repeated restructuring, or ROIC below 9%. [S3][S5]
4. Cash conversion follows earnings. Supporting evidence would be full-year operating cash at least equal to continuing income and slower working-capital growth. Disconfirmation would be continued contract-asset and inventory absorption. [S1]
5. The valuation can be earned without renewed momentum. Supporting evidence would be earnings-driven returns at a stable or lower multiple. Disconfirmation would be negative revisions combined with high market sensitivity. [S7][S8]

The factor model adds positioning context, not fundamental evidence. Market exposure of 1.20 and momentum exposure of 0.60 indicate historical sensitivity to broad-market and momentum returns under the model. Negative residual momentum and residual Sharpe indicate recent company-specific weakness under its definitions. With R-squared of 48.8%, more than half of in-sample return variance was not accounted for. The model cannot establish investor crowding or the cause of the selloff. [S8]

Three transferable research hypotheses survive testing against SPX. First, production-capacity announcements update revenue optionality before they update ROIC confidence because orders, utilization, margin, and working capital remain necessary. Second, long-term supplier arrangements update qualification confidence before demand-guarantee confidence when only purchase orders enter backlog. Third, adjusted-EPS accretion from acquisitions does not establish value creation without including transaction consideration and recurring integration requirements in return calculations. Unrelated retrieved rules from biotechnology, restaurants, and other business models were rejected as out of scope rather than forced into the analysis. [S1][S3][S4][S5][S6][S9]

Verdict: The differentiated view is not that investors missed cooling demand. It is that the current price offers exposure to an operating outlook that improved while correctly charging more for leverage, margin, and conversion risk. The strongest challenge to that view is that order evidence has not yet matured into higher HVAC margin, post-acquisition cash returns, or mid-teens ROIC.

Fact vs. Interpretation

Verdict: The highest-confidence facts concern reported revenue, backlog, margin, cash, debt, and transaction consideration. The most consequential uncertainties concern the enforceability of demand, utilization, normalized margin, acquisition returns, and terminal valuation.

Open Questions

1. What portion of the approximately $430 million 2026 data-center estimate is represented by firm purchase orders, and what portion relies on customer forecasts under long-term arrangements? [S6]
2. What cancellation rights, deposits, escalation clauses, and fixed-price exposures apply to HVAC backlog? [S1][S3]
3. How much 2027 data-center revenue is already in backlog, and how concentrated is it among hyperscalers, colocation operators, contractors, and engineering firms? [S1][S6]
4. What standalone revenue, EBITDA, free cash flow, invested capital, and synergy schedule will Neptronic contribute after purchase accounting? [S5]
5. How much additional capex is required to reach $1.1 billion of capacity, and what utilization clears the cost of capital? [S3][S6]
6. What percentage of consolidated revenue and gross profit comes from parts, service, software, maintenance, and replacement activity? [S3]
7. Will second-half collections and shipments reverse the first-half working-capital build? [S1]
8. What is normalized D&M margin after removing project timing, and when will backlog stabilize? [S1][S6]
9. Can non-data-center HVAC sustain 5%–6% organic growth after acquisition, price, and weather effects normalize? [S6]
10. Will the compensation committee add ROIC or economic profit to incentive metrics? [S4]

Verdict: These are not peripheral disclosure gaps. Answers to the order-quality, utilization, cash, and acquisition-return questions determine whether SPX merits a premium compounder valuation or a more cyclical equipment multiple.

What Must Be True

Bull tests

• Data-center cooling revenue must reach approximately $430 million in 2026, and 2027 purchase orders must continue growing without materially weaker cancellation or pricing terms. The current evidence supports the 2026 estimate but does not disclose the firm-order percentage. [S1][S6]
• HVAC segment margin must recover above 24% as factory start-up costs and tariff timing moderate. The measurable starting point is 22.8% in Q2 versus full-year guidance of 24.5%–25.0%. [S1][S2]
• Full-year cash from operations must at least match continuing income, with inventory, receivables, and contract assets growing more slowly than revenue. H1 cash conversion did not meet that test. [S1]
• Net leverage must decline toward or below 1.0 times by the end of 2027 without another large equity issuance or acquisition. Management’s current pro-forma estimate is approximately 1.4 times. [S5][S6]
• Inclusive ROIC must exceed 13% and trend toward 14%–15% as Neptronic and new capacity mature. Current accounting and transaction-adjusted estimates remain near 11%–12%. [S1][S3][S5]
• Non-data-center HVAC must sustain at least 5% organic growth and D&M backlog must stabilize. Current management commentary supports the HVAC premise, while the latest D&M backlog contradicts the stabilization premise. [S1][S6]

The bull thesis is falsified by two consecutive quarters of negative organic HVAC orders, a material fixed-price cooling charge, persistent sub-23% HVAC margin after identified start-up costs pass, material capacity underutilization after the 2028 network is available, or inclusive ROIC remaining near 10%–11% despite full production. [S1][S6][S14]

Bear tests

• Competitive capacity must cause lead times, price, or utilization to normalize before SPX fills its network. Current peer investment supports the supply premise, but strong backlog means the rollover has not yet occurred. [S9][S10][S11][S12]
• HVAC margin must remain below 23% after start-up expenses normalize, indicating that current pressure is structural rather than transitional. [S1][S6]
• Working capital must continue consuming cash and prevent meaningful debt reduction. A second-half reversal would contradict this premise. [S1]
• Neptronic and other acquisitions must increase adjusted EPS without lifting inclusive ROIC. Asset-level cash disclosure or consolidated ROIC above 14% would falsify this claim. [S3][S5]
• The market must revalue SPX toward 13–14 times EBITDA as organic growth slows. Stable valuation alongside EBITDA above $700 million would undermine this premise. [S2][S7]

The bear thesis is falsified if 2027 organic growth remains above 8%, HVAC margin exceeds 24%, operating cash conversion exceeds 100%, net leverage falls below 1.0 times, and inclusive ROIC reaches at least 14%. Those outcomes would show that capacity and acquisitions generated durable economic returns rather than only revenue and adjusted earnings. [S1][S2][S3][S5]

Public source appendix

• S1: SPX Technologies Form 10-Q for the quarter ended June 27, 2026 — primary filing; published 2026-07-30; Financial statements; Notes 3, 7, 10 and 13; cash flows; segment results; backlog; subsequent Neptronic event; repurchases and trading arrangements
• S2: SPX Technologies second-quarter 2026 earnings release — primary company release; published 2026-07-30; Q2 results; organic-growth bridge; updated revenue, adjusted EBITDA and adjusted EPS guidance; segment outlook; non-GAAP reconciliations
• S3: SPX Technologies 2025 Form 10-K — primary filing; published 2026-02-25; Items 1, 7 and 8; revenue recognition; segment results; acquisitions; offering outcome; cash flow; capex; debt; leases; goodwill and intangibles
• S4: SPX Technologies 2026 proxy statement — primary filing; published 2026-03-31; Executive compensation; incentive metrics; adjusted-cash-flow reconciliation; long-term awards; ownership guidelines and beneficial ownership
• S5: SPX Technologies announcement of completed Neptronic acquisition — primary company release; published 2026-07-23; Closing date; CA$605 million/approximately US$430 million consideration; revenue; acquisition multiple; facility, employees and products
• S6: Company Financials — SPXC Q2 and Q1 2026 earnings-call transcripts — third-party transcript service reconciled to filings; published 2026-07-30; Q2 and Q1 prepared remarks and Q&A: data-center revenue, capacity, long-term arrangements, margins, non-data-center HVAC growth, leverage, acquisition multiple and D&M timing
• S7: Company Financials — SPXC profile, financial statements, ratios, enterprise value and market data — third-party financial dataset reconciled to filings; publication date unavailable; Exchange-qualified NYSE:SPXC profile; annual and quarterly statements; standardized ratios; prices through September 10, 2026; peer valuation calculations reconciled to filings
• S8: The factor model — SPXC snapshot dated September 9, 2026 — quantitative diagnostic; published 2026-09-09; Statistical factor exposures, residual signals and regression diagnostics
• S9: AAON Form 10-Q for the quarter ended June 30, 2026 — peer primary filing; published 2026-08-07; Revenue; gross margin; total and BASX backlog; manufacturing capacity; working capital and ramp costs
• S10: AAON August 2026 investor presentation — peer primary company presentation; published 2026-08-18; Data-center product portfolio; manufacturing capacity; plant utilization; backlog and company-presented ROIC
• S11: Johnson Controls third-quarter 2026 results — peer primary company release; published 2026-07-29; Organic orders; $21 billion backlog; Americas backlog; data-center and mission-critical demand; margins and guidance
• S12: Trane Technologies second-quarter 2026 results — peer primary company release; published 2026-07-30; Organic bookings; commercial-HVAC bookings; $12.1 billion backlog; revenue, margins and guidance
• S13: SPX Technologies 2024 Form 10-K — primary filing; published 2025-02-26; Ingénia and KTS transactions; segment results; backlog; goodwill and acquired intangibles
• S14: SPX Technologies 2022 Form 10-K — primary filing; published 2023-02-24; 2022 operating trough; large cooling-project losses; impairments; continuing operations and historical statements
• S15: SPX Technologies Form 4 — Brian Deck restricted-stock award — primary insider filing; published 2026-07-27; Transaction code and award details distinguishing a grant from an open-market purchase
• S16: SPX Technologies Form 10-Q for the quarter ended September 27, 2025 — primary filing; published 2025-10-30; Actual August 2025 offering outcome: approximately 3.059 million shares, $575.0 million gross and $551.1 million net proceeds
• S17: SPX Technologies September 2025 credit-facility amendment — primary filing; published 2025-09-11; $2.025 billion aggregate facilities; $1.5 billion revolver; $500 million term loan; September 2030 maturity
• S18: Baltimore Aircoil Company history — competitor primary company history; publication date unavailable; Corporate history and Amsted Industries ownership since 1985
• S19: Johnson Controls data-center thermal-management architecture — peer primary product documentation; publication date unavailable; Heat capture, transfer and rejection architecture; air- and water-cooled chillers; towers and dry-cooling alternatives