Nextpower Inc. (NASDAQ: NXT) — The Tracker Leader Whose Margin Is Half Washington’s, Priced for the Bull Case

Formerly Nextracker Inc. (renamed November 2025; ticker unchanged). Utility-scale solar equipment. Report date: 2026-07-02. Fiscal year ends March 31; FY26 = year ended March 31, 2026.

⚡ Claude’s Take

This block is the author’s own subjective opinion. It is general information, not investment advice. The analysis that follows takes no position and carries no price target — this block is the one exception.

Verdict: HOLD / AVOID-at-this-price / accumulate-on-weakness / not-a-short. Medium conviction. Fair-value zone ≈ $90–115 — roughly a mid-teens-to-~20x multiple on ~$4.40 of FY27 adjusted EPS, haircut for the fact that half the earnings base is a fading government credit. I’d accumulate a genuinely great franchise below ~$90 (into the low-$90s/high-$80s), and I would not chase it above ~$130. At today’s $112.84 you are paying a full growth-equity multiple for a business whose guided year delivers flat-to-declining per-share earnings.

Here’s the tension the tape is missing. Nextpower is a real, high-quality business — the undisputed #1 in solar trackers for ten straight years, asset-light (it engineers; a 100±factory contract-manufacturing network builds the steel), net cash ~$1.1B, and it grew 18–20% and stayed solidly profitable straight through the 2024 downturn that pushed its #2 (Array) into a $240M loss. That is the signature of a durable, if narrow, moat. But the reported 32.6% gross margin and ~30% ROIC that justify the multiple are roughly half a subsidy. Strip the §45X manufacturing credit ($379.9M in FY26, up from $224.9M) out of cost of sales and gross margin is ~22% — and flat in dollars on +20% revenue — while ex-credit operating income actually shrank ~23% last year. The market is paying ~25–33x forward earnings and ~19–20x forward EBITDA for a steel-and-motors hardware maker whose underlying, unsubsidized economics are eroding under price competition, and whose §45X tailwind steps down 25%/year from 2030. On top of that sits a policy-timing cliff: OBBBA accelerated the phase-out of the §48E/45Y project credits NXT’s customers depend on, so the record $5.25B backlog is partly demand pulled forward into FY26–27, with an air-pocket risk in FY28.

Framing: this is a high-beta clean-energy momentum name — statistically it is the solar factor (industry beta +2.2 to +2.5, market beta 1.36, strongly negative low-volatility and interest-rate loadings) — that doubled in a year, sits ~28% off its May-2026 all-time high, and trades at the richest price-to-sales of its short public life. Not a falling knife; not abandoned value; a crowded, leveraged bet on two exogenous variables (solar-policy durability and rates) dressed up as a pure quality-compounder story. The valuation work says the base case roughly re-earns today’s price over three years while the bear case halves it — a poor risk/reward at spot.

Conviction: medium. The single fact that flips me bullish: durable positive book-to-bill and re-accelerating organic revenue with adjusted gross margin stabilizing at 22%+ through the ITC phase-out — i.e., proof the business grows without borrowing FY28’s demand and holds margin without the subsidy expanding. The single fact that flips me bearish (to an active avoid): two consecutive quarters of backlog decline or negative bookings, or ex-45X gross margin breaking below ~20%, signaling the core hardware is commoditizing faster than the platform can offset. Tag: the tracker king whose margin is half Washington’s, priced for the bull case.

📈 Stock Price Action — ~3.4-Year Event Map (since the Feb-2023 IPO)

Nextpower has only traded since its February 2023 IPO, so this covers the full public life, not a five-year window. Price moves are FACTS; the attributed causes are INTERPRETATION. No price target, no recommendation, no support/resistance implied.

Arc. Nextracker (as it then was) IPO’d at $24 (first trade $30.31) in February 2023, spent two years going roughly nowhere, then roughly doubled in the twelve months into a closing all-time high of ~$156 (intraday $163.13) on May 29, 2026. It now trades at $112.84 — about 28% below the closing peak but still +114% off its 52-week low ($52.61) — inside a 52-week intraday range of $52.61–$163.13. In plain terms: an IPO that idled for two years, then a violent clean-energy melt-up, now taking its first real breather.

Cycle narrative.

1. IPO & drift (2023). Priced at $24, opened at $30.31, and spent 2023 range-bound in the low-$30s–low-$40s as the Flex-controlled float found hands.
2. First IRA melt-up (late-2023→Feb-2024). Utility-solar demand optimism and solid prints carried it to ~$56 — the crest of the first enthusiasm wave.
3. 2024 solar bear (Apr–Sep 2024). A grind to ~$37 tracked the whole solar complex lower as long rates stayed high and the November election stoked IRA-repeal fears; NXT’s negative rate/low-vol loadings made it a high-beta casualty.
4. Rebound (Jan 2025). An oversold bounce as §45X-preservation odds firmed.
5. §45X clarity (May 2025). The Q4/FY25 beat plus OBBBA preserving the manufacturing credits reset the demand narrative.
6. The doubling (Aug 2025→Jan 2026). Three straight FY26 beats layered onto the data-center/power-demand theme drove the stock from ~$54 to ~$117.
7. The blow-off top (May 2026). The Q4/FY26 print, a raised FY27 guide, the Prevalon storage acquisition, and a sell-side PT flurry to $142–182 marked the ATH.
8. First real pullback (Jun 2026→now). A tech-and-solar selloff and post-double profit-taking pulled it back to $112.84 even as the Zimmermann deal and new products landed.

1. Executive Summary

Nextpower Inc. (NASDAQ: NXT), until November 2025 named Nextracker, is the world’s largest maker of utility-scale solar trackers — the steel structures, motors, controllers, and control software that tilt photovoltaic panels to follow the sun, lifting a project’s energy yield by up to ~25% versus fixed-tilt and improving its levelized cost of energy. It has shipped 160+ GW cumulatively across six continents and has held the #1 global position by GW for ten consecutive years. The business is unusually capital-light: NXT designs and engineers, while a network of 100+ contract-manufacturing facilities in 19 countries — built “with close to no capital investment” — fabricates and drop-ships components near project sites. That model produces a clean, net-cash (~$1.1B) balance sheet, ~$514M of free cash flow, and headline returns (ROIC 28.9% in FY26) that screen as elite. Revenue compounded from ~$1.46B (FY22) to $3.56B (FY26), up 20% last year, with a record backlog above $5.25B.

The moat is real but narrow, and the earnings quality is materially lower than the income statement advertises. The competitive advantage is a composite of bankability/track-record intangibles (lenders and independent engineers will not finance an unproven 30-year structure holding a nine-figure asset), EPC spec-in switching costs, and steel-fab procurement scale. The proof is the contrast with Array Technologies, the #2, which lost $240M in 2024 and has posted net losses in four of the last five years while NXT grew and earned 32%+ gross margins — the textbook signature of a durable advantage surviving a downturn. But the central finding of this report is that the reported gross-margin expansion is a government subsidy, not operating improvement. The IRA §45X advanced-manufacturing credit — captured via supplier vendor rebates and direct assignment on eligible torque tubes and fasteners, and booked as a reduction of cost of sales — contributed $379.9M in FY26 (up from $224.9M in FY25 and $121.4M in FY24). Strip it out and FY26 gross profit was flat ($780.2M vs $783.9M) on +20% revenue, gross margin was ~22%, not the reported 32.6%, and ex-credit operating income actually declined ~23%. The credit equals 54.5% of FY26 GAAP operating income; ex-credit ROE is ~14% and ROIC mid-teens — respectable, not exceptional.

The valuation prices the quality and a growth re-acceleration the guided year does not show. At $112.84 (≈$17.7B market cap, ≈$16.6B EV), NXT trades at ~30x trailing and ~25–33x forward earnings and ~19–20x forward adjusted EBITDA — a growth-equity multiple, and the richest price-to-sales of its short public life (92.9th own-history percentile). Yet management’s raised FY27 guide embeds roughly flat adjusted EPS and lower GAAP EPS; Q4 FY26 adjusted EBITDA margin already fell to 22.9% from 26.2% a year earlier. The market is therefore paying for a re-acceleration beyond FY27 — international penetration, storage (Prevalon), power conversion — against a policy-timing cliff: OBBBA (2025) preserved §45X but accelerated the phase-out of the §48E/45Y project credits that drive NXT’s customers’ returns, so much of the record backlog is demand pulled forward, with an FY28 US air-pocket risk. Scenario analysis is downside-skewed: a base case that roughly re-earns today’s price over three years, a bull case ~+45–65%, and a policy-driven bear case ~−50% requiring no operational catastrophe. The company is simultaneously repositioning — the Nextracker→Nextpower rebrand marks a pivot into foundations, eBOS, robotics, storage, and inverters via seven-plus acquisitions in 24 months — which expands the TAM story but dilutes the focused, high-return tracker moat and risks empire-building; management incentives reward scale (revenue, EBITDA), not returns on capital, and insiders are net sellers with zero open-market purchases. This is a genuinely good business whose price already assumes the bull path with little margin of safety.

2. Business Overview

What NXT sells. Nextpower is the world’s largest maker of single-axis solar trackers — the steel structure, motors, controllers, and software that mount utility-scale photovoltaic (PV) panels on a rotating horizontal torque tube and tilt them east-to-west to follow the sun across the day. The physics is the entire value proposition: by keeping panels normal to the sun, a tracker lifts a project’s annual energy yield by up to ~25% versus fixed-tilt (stationary) mounting (FACT — FY26 10-K Item 1, citing Lazard and a Joule/Cell Press techno-economic study). Because the incremental energy revenue over a 30-year plant life typically exceeds the incremental hardware cost, trackers improve the risk-adjusted levelized cost of energy (LCOE) and have become standard equipment: the 10-K states that “the majority of utility-scale projects installed today in mature markets” use trackers. NXT has shipped more than 160 GW cumulatively to six continents and claims the #1 position by GW shipped for ten consecutive years (FACT).

The flagship product is NX Horizon, a self-powered, independent-row tracker — each row carries its own small PV panel powering its own motor/controller, so rows commission before grid power is available and can be angled independently, row by row. Around this core, NXT has layered an expanding product family (FACT, Item 1):

• NX Horizon-XTR / XTR-1.5 — terrain-following variants that conform to sloped or uneven sites, cutting cut-and-fill earthworks and expanding the buildable-site universe.
• NX Horizon with Hail Pro / Hail Pro-75 — automated hail-stow (up to 75° tilt) using weather data; a genuine differentiator as insurers price hail risk into utility solar. Wind and hurricane stow live in the software.
• TrueCapture and NX Navigator — separately licensed software. TrueCapture is a yield-optimization control system that adjusts each row for topography, irradiance and diffuse-light conditions (claimed 1–2% energy-loss recovery); NX Navigator is the monitoring/operability and weather-mitigation console. This is the closest thing NXT has to a recurring, high-margin software attach.
• Adjacencies added by acquisition (FY25–26): NX Foundation Solutions (Ojjo / Solar Pile International — earth-truss and driven-pile foundations for difficult soils), Steel Frames (Origami Solar — roll-formed steel PV frames vs. extruded aluminum), eBOS (Bentek — electrical balance-of-system harnesses/combiners), AI & Robotics (OnSight — autonomous inspection), soiling sensors (Fracsun), and battery energy storage (Prevalon, announced May-2026). The June-2026 launches of the NX Gemini 2P (two-panel-in-portrait tracker) and NX Anchor integrated foundation extend the range into Europe.

How it makes money — and the recurring-revenue question. Revenue is overwhelmingly project-based hardware sales, recognized as product ships to named project sites — 88% recognized over time ($3,117.4M FY26), with point-in-time revenue jumping to $442.0M (from $77.0M) as acquired product lines enter the mix. Solar-tracker system sales were ~88% of FY26 revenue; non-tracker platform sales ~12%, up from ~8% in FY25 — and management flags that non-tracker (software, eBOS, foundations, robotics) is growing faster than the tracker base (FACT). The software (TrueCapture / NX Navigator) is licensed separately and can be deployed to the installed fleet, which is the seed of a recurring, attach-driven model — but NXT does not disclose software as a separate revenue line, so its scale is unquantified (OPEN QUESTION). Bottom line: this is a lumpy, project-cycle equipment business, not a subscription business. Revenue is “recurring” only in the sense that a large, growing base of EPC/developer relationships places repeat orders; individual quarters swing with project timing, interconnection, and financing.

Capex-light manufacturing — the defining structural feature. NXT designs and engineers; it does not own most of its factories. As of March 2026 it ran ~1,500 MW/week of capacity (~80 GW/year) through 100+ contract-manufacturing facilities in 19 countries across five continents — built “with close to no capital investment” (FACT, Item 1). In the US it has secured raw-steel-coil supply from domestic mills feeding 30+ US fabricators (>40 GW of primary-component capacity) deliberately co-located near mills and project sites to minimize freight and qualify for domestic content. It owns manufacturing only “on a very limited basis” — controllers in Brazil, eBOS in California, and a Saudi tracker-component JV (Nextpower Arabia, with the Abunayyan group). This asset-light network is simultaneously the cost/flexibility moat and the source of NXT’s headline returns: almost no invested capital sits in fixed assets (capex was ~1.4% of revenue in FY26). It also drop-ships direct to site, holding only ~300,000 sq ft of contingency warehousing globally.

Customers and geography. Customers are EPCs, developers, IPPs and plant owners — 275+ active customers across 40+ countries (FACT). Concentration had improved to no >10% customer in FY25, but in FY26 one customer (“Customer G”) reached ~12% of revenue — the first >10% customer in three years, worth monitoring though not yet acute (FACT). The Volume Commitment Agreement (VCA) program — multi-project, multi-year framework contracts since FY23 — underpins the >$5.25 billion backlog (FACT). Geographically, however, the story is more US-concentrated than the “global platform” branding implies: FY26 was 77% US / 23% Rest-of-World, and RoW revenue actually fell 11% to $828.7M on Latin American weakness while US revenue grew 34% to $2,730.7M (see Financial Quality and Growth).

Verdict. A high-quality, capex-light, market-leading equipment franchise with a genuine energy-yield value proposition and a widening product platform — but a project-cyclical, hardware-dominated (88%) revenue model whose “recurring” and “software” characteristics remain more narrative than disclosed financial fact, and whose growth in FY26 was entirely US-driven.

3. Industry Dynamics

Market size and the demand super-cycle. NXT sits downstream of utility-scale solar deployment, one of the fastest-growing segments of global electricity supply. US utility-scale solar installations exceeded ~30 GW in 2025 (total US installed base ~266 GW), and global annual PV demand is on the order of ~500–600 GW (per earlier public First Solar analysis and U.S. utility-solar market data, 2025–2026). NXT’s own 10-K leans hard on the “electricity super-cycle” framing — AI/data-center load growth, electrification of transport and heat, and solar’s status as the cheapest incremental generation in many regions (Lazard: utility-solar generation cost fell ~84% 2009–2025) (FACT). The demand backdrop is genuinely strong; the debate is almost entirely on the policy side of demand, not the physics.

Tracker penetration — the structural growth vector. Trackers are now the default in mature markets (US, Australia, India, Latin America, parts of Europe) — the majority of US ground-mount is tracked — but penetration is materially lower in developing markets (Middle East, Africa, parts of Asia/Europe) where fixed-tilt still competes on upfront cost (FACT/INTERPRETATION). The bull’s structural argument is that international tracker attach converges toward mature-market levels, expanding NXT’s addressable GW faster than utility solar itself grows. That thesis is real over a decade — but note it is an attach-rate story that NXT’s own FY26 RoW numbers (down 11%) did not deliver.

Competitive structure — a consolidated oligopoly, not a commodity scrum. This is the single most important contrast with the module industry next door. The 10-K names competitors as Array Technologies (ARRY, the clear #2), Arctech Solar, GameChange Solar, PV Hardware (PVH), Shoals, and Trina (FACT). NXT holds an estimated ~30%+ of the global tracker market and higher in the US, with ARRY the only other at-scale Western player. Unlike PV modules — where First Solar fights ~105 GW/year of new (largely Chinese) capacity against collapsing global ASPs in a textbook Marathon capital-cycle bust — the tracker market is a two-strong oligopoly at the top with a fragmented tail. The 10-K is explicit about the barrier: “customers’ reluctance to purchase products from new entrants with a limited history has resulted in a bifurcation of providers based on their track record.” Trackers hold PV modules 30+ years through wind and hail; a tracker failure is a plant-level catastrophe, so lenders and EPCs will not spec an unproven structure. That bankability gate is a real barrier to entry that the module industry lacks.

Profit pools and pricing. A tracker is partly commoditized steel (torque tubes, piles, fasteners) plus value-add (motors, controllers, software, engineering, bankability). Steel is largely pass-through — NXT indexes contracts and absorbs/passes commodity moves — so the durable margin sits in the engineering/software/procurement layer. NXT earns ~32–34% reported gross margin, but, as the Financial Quality section shows, roughly a third of that is the §45X credit; the underlying design-and-supply-chain margin is ~22%, still roughly double ARRY’s volatile through-cycle average — which tells you the value-add layer and scale procurement are real, but not that they confer monopoly pricing. The risk: as steel is pass-through and the structure partly commoditized, the incremental economics are thinner than the average — FY26 incremental operating margin was only ~9.7%, a tell that at the margin NXT is taking price/mix pressure (international, lower-priced projects, dilutive M&A).

Regulation — the dominant swing variable, but different from a module maker’s. The US policy stack matters enormously, but NXT’s exposure differs from First Solar’s:

• §45X manufacturing credit — NXT’s torque tubes ($0.87/kg) and structural fasteners ($2.28/kg) qualify through 2029, stepping down 25%/year in 2030–32 and ending after 2032 (FACT). Historically the credit accrued to NXT’s suppliers and was captured indirectly; beginning CY2025 NXT also claims some 45X directly by assignment. This is a margin tailwind but a smaller and more indirect one than a module maker’s 45X.
• Domestic-content ITC adder — projects using US-made iron/steel/components get a bonus credit (30%→40% ITC). NXT built a US supply chain specifically to sell a domestic-content-compliant tracker — a demand pull toward NXT’s US output.
• OBBBA (2025) — the demand-timing cliff. This is the key negative. OBBBA preserved §45X (NXT’s margin) but accelerated termination of the §48E/45Y project (ITC/PTC) credits that NXT’s customers rely on: projects must begin construction by ~July 4, 2026 to use a four-year safe harbor, else be placed in service by December 31, 2027; the 5% cost safe harbor was eliminated in September 2025 (Notice 2025-42) (FACT). Plus FEOC restrictions excluding China-linked entities from the credits. Interpretation: this pulls construction starts forward into 2026, then creates real air-pocket risk for US utility-solar deployment in 2027–2028 — the single biggest cyclical threat to NXT’s US-heavy revenue.
• Trade-policy tailwind: AD/CVD tariffs, UFLPA, and (June 2026) reported moves to restrict foreign inverters push developers toward domestic-content, FEOC-clean equipment — advantaging NXT’s US network over Chinese-linked rivals (Arctech, Trina).
• Interconnection queues and high rates compress project IRRs and delay timing — a demand headwind that hit the whole space in 2024 (and halved ARRY) but which NXT rode through on backlog.

Marathon capital-cycle read. Is capital flooding into tracker manufacturing? No — the opposite. The #2 (ARRY) has been loss-making and shrinking, Chinese entrants are FEOC-blocked from the lucrative US pool, and NXT itself adds contracted capacity with near-zero capital. This is a consolidating, not a flooding, supply side — a favorable Marathon signal — in sharp contrast to the module bust.

Verdict. Structurally attractive by the standards of clean-energy equipment — a consolidated tracker oligopoly with real bankability barriers, a strong secular demand tailwind, and a supply side rationalizing rather than flooding — but heavily levered to US policy timing (the OBBBA project-credit cliff) and cyclical project financing. Better structure than the commoditized module industry; worse cyclicality and policy-dependence than a boring industrial.

4. Competitive Position

Name the moat. NXT’s advantage is narrow but real, and it is a composite of three of Greenwald’s genuine advantage types rather than one dominant mechanism:

1. Intangibles — bankability / track record (the primary moat). The 10-K itself identifies the barrier: buyers “reluctan[t] to purchase products from new entrants with a limited history,” producing a “bifurcation of providers based on their track record.” A utility-scale tracker must survive 30 years of wind/hail/snow holding a nine-figure asset; a structural failure is an uninsurable-scale loss. Lenders, independent engineers, and insurers therefore gate financing on proven trackers. NXT’s 160+ GW installed base across six continents and ten-year #1 status is the credential — and it is self-reinforcing: every incremental GW deepens the field-performance dataset that independent engineers bless. This is a genuine intangible barrier, backed by roughly 329 US + 498 non-US patents on architecture, controls, and foundations. Pressure-test: it does not stop a proven #2 (ARRY) or a credentialed regional player (PVH in Iberia, Arctech in MENA/Asia) — it is a barrier against new entrants and the commodity tail, not against the established duopoly.

2. Switching costs / spec-in (secondary). Trackers are designed into a project early; EPCs standardize installation crews, software (NX Navigator), and O&M procedures around a vendor; the installed base runs proprietary controllers and TrueCapture firmware. This creates developer-level captivity and repeat-order stickiness (the VCA program formalizes it into multi-year commitments). Pressure-test: switching costs are moderate, not iron — a developer can and does multi-source across projects; there is no per-plant lock-in comparable to enterprise software.

3. Scale / cost advantage in a fragmented steel-fab supply chain (secondary). NXT’s 100±fabricator, 19-country network and centralized steel procurement give it landed-cost and lead-time advantages a sub-scale rival cannot replicate, plus the ability to hit domestic-content thresholds market by market. The financial proof is the gross-margin gap versus ARRY.

Software (TrueCapture) is a feature, not yet a moat. It contributes yield and stickiness, but it is not separately monetized at disclosed scale, and competitors offer analogous control systems. Treat it as reinforcing the intangibles/switching-cost moat, not as a standalone network-effect or data moat — the 10-K gives no evidence the installed-base data compounds into an unassailable advantage. If a “moat” claim can’t be tied to a financial outcome that erodes without it, it isn’t a moat — and TrueCapture’s financial signature is invisible in the disclosures.

The decisive evidence — Greenwald’s share-stability and ROIC tests, run against ARRY. The cleanest proof of a moat is stable share plus persistent excess returns through a downturn. The NXT-vs-ARRY contrast is stark (ROIC.ai; ARRY on calendar years, NXT on March fiscal years):

The read: during the 2024 tracker downturn (rates/interconnection/project delays) that cut ARRY’s revenue by ~42% and drove it to a −$240M loss, NXT grew 18% and earned a 32.5% gross margin and $509M of net income. ARRY’s gross margin has swung violently (8% → 13% → 26% → 33% → 23% across 2021–25) and it has posted net losses in four of the last five years (with expensive preferred stock draining another ~$52–60M/year in dividends). NXT’s margins are stable in the low-30s (albeit subsidy-supported) and it has been solidly, growingly profitable every year. NXT gained share and profitability precisely when the industry punished the weaker player — the textbook signature of a durable competitive advantage. ARRY is not a peer NXT is racing; it is a distressed #2 NXT is out-executing.

The erosion tells (be direct). The advantage is narrow and there are cracks worth naming:

• Returns are very high but falling: ROIC 54.3% (FY24) → 36.9% (FY25) → 28.9% (FY26) — still multiples of WACC, but decaying as the M&A-built capital base grows and acquired non-tracker businesses (lower-margin foundations/eBOS) dilute the mix. And, critically, most of that ROIC is the §45X credit (see Financial Quality): ex-subsidy ROIC is mid-teens.
• Incremental operating margin was only ~9.7% in FY26 — far below the ~20% average op margin, evidence that the next dollar of revenue (international, competitively-priced projects, dilutive tuck-ins) carries much thinner economics than the installed franchise. Operating margin has stepped down 23.5% → 21.6% → 19.6% over three years.
• Trackers are partly commoditized steel, so the moat protects the leader’s premium and the domestic-content/bankability pool, but it does not confer monopoly pricing — hence the pass-through economics and the RoW price weakness.

Verdict. A real but narrow moat — bankability/track-record intangibles, reinforced by EPC spec-in switching costs and steel-fab scale — that decisively separates NXT from the commodity tail and from a floundering #2, and passes the Greenwald share-stability and excess-return tests through a genuine downturn. It is not a wide, pricing-power moat: the underlying product is partly commoditized, returns are decaying off an extraordinary base (and are substantially subsidy-driven), and incremental economics are thinning. Durable advantage — yes; unassailable — no.

5. Growth History and Forward Opportunities

Historical growth — high magnitude, decelerating volume, increasingly acquired. NXT compounded revenue from roughly $1.46B (FY22) to $3.56B (FY26) — a ~25% CAGR — with the recent path FY24 $2,499.8M → FY25 $2,959.2M (+18.4%) → FY26 $3,559.4M (+20.3%). Decomposing FY26: GW delivered rose from 33.6 to 38.0 (+13%) while revenue rose +20%, so ~13 points of volume plus ~7 points of price/mix (FACT, MD&A). Critically, volume growth is decelerating — GW delivered grew +29% in FY25 but only +13% in FY26 — and the revenue beat over volume came from ASP/mix, not units. The company is shipping more dollars per GW as software/eBOS/foundations attach rises (non-tracker 8%→12% of revenue), a positive mix story, but the underlying unit engine is slowing.

Organic vs. acquired — the mix is tilting toward M&A. Historically organic (share gains + market growth), NXT’s FY25–26 growth is now materially supplemented by a rapid tuck-in cadence (FACT, 10-K notes + 8-Ks):

• Ojjo (Jun-2024) and Solar Pile International — foundations (earth-truss / driven-pile).
• Bentek (May-2025) — eBOS / electrical infrastructure, US fabrication footprint.
• OnSight (mid-2025) — autonomous robotic inspection; launched an AI & Robotics initiative and named a Chief AI & Robotics Officer.
• Origami Solar (Sep-2025) — roll-formed steel PV frames.
• Fracsun (Nov-2025) — soiling-measurement sensors.
• Zigor Corporation & Apex Power (subsequent event, ~May-2026) — power electronics.
• Prevalon Energy (announced May-2026) — battery energy storage (BESS), for up to ~$365M ($150M cash + $50M stock + up to $165M contingent). (This is a 100% acquisition, not a JV.)
• Zimmermann PV-Steel Group (Jun-2026) — European steel/foundations, up to €330M, adding product lines and ~15 countries.

Goodwill rose from $371M to $489M in FY26 (FACT). Seven-plus acquisitions in ~24 months is a meaningful shift toward inorganic growth (see Capital Allocation for the discipline question).

Forward opportunities — genuine, but front-loaded with US-policy and international-execution risk.

• Backlog visibility: >$5.25 billion at FY26 year-end (VCA + POs) provides real near-term cover — but management explicitly warns backlog is subject to cancellation, delay, and scope reduction, so it is not a guaranteed revenue schedule.
• International penetration (the bull’s vector): LatAm, MEA (Nextpower Arabia / Saudi JV), Europe (Zimmermann), India (Hyderabad R&D + sales hub), and Australia. The problem: FY26 RoW revenue fell 11% to $828.7M on LatAm weakness while US grew 34% — so the international-attach thesis is, so far, aspiration contradicted by the actuals. This is the growth story’s biggest credibility gap. (FACT / INTERPRETATION)
• Platform/adjacency expansion: non-tracker (software, eBOS, foundations, steel frames, robotics, and now storage via Prevalon) is growing faster than trackers and lifts revenue-per-GW and stickiness — the highest-quality element of the forward story, though it also dilutes gross margin and returns.
• Storage/BESS (Prevalon): opens a large adjacent TAM (solar-plus-storage is the cheapest incremental generation in many US markets) and leverages the same developer/EPC customer base — but it is a new, competitive market (Tesla, Fluence, Sungrow) where NXT has no bankability incumbency.
• FY27 guidance (raised, May-2026): revenue $3.8–4.1B (from $3.6–3.8B), adjusted EBITDA $825–900M (from $800–900M), adjusted diluted EPS $4.21–4.59, GAAP EPS $3.19–3.56 — but see Valuation: the raise is revenue-led while per-share earnings are roughly flat.

The overhang on growth quality — the OBBBA cliff. With 77% of revenue in the US and OBBBA accelerating the §48E/45Y project-credit deadlines (begin-construction by ~July 4, 2026), a pull-forward of US construction starts into FY26–27 followed by an air-pocket in FY28+ is a live risk. Much of the recent US strength may be developers racing to safe-harbor projects — i.e., borrowing demand from the future.

Verdict. High-magnitude but mixed-quality growth. The organic core is real (durable share gains, a strong secular US utility-solar/AI-load tailwind, a >$5.25B backlog, and a rising, sticky non-tracker attach). But three qualifiers pull the grade down: (1) FY26 growth was entirely US-driven while international revenue declined, undercutting the headline international-penetration thesis; (2) growth is increasingly acquired (7+ deals, goodwill +$118M in one year); and (3) unit growth is decelerating and sits atop a US policy pull-forward that risks an out-year air-pocket. This is a leader growing well for now, on a demand base more policy-timed and US-concentrated than the “global platform” branding suggests.

6. Financial Quality

Revenue growth and composition. NXT compounded revenue from ~$1.46B (FY22) to $3,559.4M in FY26 — FY24 $2,499.8M, FY25 $2,959.2M (+18.4%), FY26 (+20.3%). Growth is real and volume-led. Revenue is 88% recognized over time ($3,117.4M) with a jump in point-in-time revenue to $442.0M FY26 (from $77.0M) as acquired product lines enter the mix. Geographic mix shifted more US-concentrated: US revenue grew +34% to $2,730.7M (77%, up from 69%) while Rest-of-World fell −11% to $828.7M (23%). One customer (“Customer G”) reached ~12% of revenue in FY26 — the first >10% customer in three years; concentration is emerging but not yet acute.

The gross-margin step-up is a §45X subsidy, not operating improvement — the central QoE finding. Reported gross margin rose from ~10% (FY22) to 15% (FY23) to 32–34% (FY24–26). The consensus narrative attributes this to US content, logistics normalization and pricing. The filings say otherwise. NXT is not a direct §45X claimant on cells or modules — it does not make them. Instead it contracts with US steel/component suppliers (torque tubes, structural fasteners) who either pay NXT a “vendor rebate” for a share of the Advanced-Manufacturing Production Credit, or assign the credit to NXT under IRC §6418. NXT books this as a reduction of inventory and then a reduction of cost of sales (10-K MD&A + Note 13):

The implication is stark: ex-45X gross profit was essentially flat in FY26 ($783.9M → $780.2M, −0.5%) on +20% revenue. The entire +$151.3M (+15%) reported gross-profit increase is the +$155M growth in the §45X credit. On management’s own adjusted numbers the picture is identical — ex-45X adjusted gross profit went ~$798.6M → ~$803.6M (+0.6%), and ex-45X adjusted gross margin fell from ~27.0% to ~22.6%. The reported 32–34% is a policy-credit margin; the underlying design-and-supply-chain margin is ~22% and declining under price competition (Array, GameChange, PVH) — even as the higher-margin US mix rose to 77%, which should have helped.

Operating leverage is negative once the subsidy is stripped. §45X equals 54.5% of FY26 GAAP operating income ($379.9M of $697.3M) and ~53% of pretax income. Ex-45X operating income declined ~23% YoY: FY25 $414.2M (14.0% margin) → FY26 $317.4M (8.9% margin). This is why reported operating margin compressed (23.5% → 21.6% → 19.6%) and FY26 incremental operating margin was only 9.7%: operating expense grew +$93M (+25% — R&D nearly tripled in two years, $42.4M → $120.9M; SG&A up 86% to $341.9M) against flat ex-credit gross profit. The company is spending into a business whose underlying unit economics are eroding, with the subsidy masking it in the headline.

Adjusted vs GAAP. Even on management’s preferred metrics, FY26 margins compressed: adjusted gross margin 34.6% → 33.3%, adjusted net margin 21.3% → 19.3%. Adjusted figures add back SBC ($120.3M — 3.4% of revenue, ~17.5% of adjusted operating income; material and dilutive) and intangible amortization. Adjusted EBITDA was ~$853.7M (vs unadjusted EBITDA $727.9M); adjusted diluted EPS ~$4.40 vs GAAP diluted ~$3.83. Share count crept 140.8M → 149.4M basic (FY24 → FY26), ~156M diluted — steady low-single-digit dilution from equity comp.

Up-C / TRA and the equity bridge. Historical negative book equity (−$3.0B FY22; −$3,076M equity-before-minority FY23) was an Up-C artifact, not distress — Flex’s economic stake sat in redeemable non-controlling interest ($3,560M FY23). The FY23→FY24 share jump (45.9M → 140.8M) is the Up-C collapse: LLC units folded into Class A as Flex fully distributed and exited (Jan-2-2024), and TPG (TPG Rise Climate, a ~$500M pre-IPO investor at a $3.0B valuation) exchanged its remaining units Feb-5-2025. Equity flipped positive ($992M → $1,628M → $2,334M FY26) as cumulative earnings pared the retained-earnings deficit (−$1,971.5M) against $4,305.7M of APIC. A residual Tax Receivable Agreement obligation of $393.2M (FY26; $419.4M FY25) pays Flex’s affiliate plus TPG 85% of realized tax benefits from the basis step-up; FY26 cash TRA payment was $27.4M — a real but modest, multi-decade cash drag. Tangible book was positive throughout.

Balance sheet, cash flow and returns. Cash $1,095M (FY26), no drawn debt (the $150M term loan that financed the Flex distribution was repaid in FY25; the new credit agreement is an undrawn revolver), net cash +$1,095M, current ratio ~2.45x, and a negative cash-conversion cycle (~−2 days — supplier-financed working capital). No liquidity or runway concern; NXT achieved an investment-grade rating in FY26. FCF: OCF $562.9M − capex $49.3M = $514M (capex ~1.4% of revenue — genuinely asset-light). Note FY26 OCF/NI fell to 0.96x (from 1.29x FY25, 1.40x FY24): the culprit is a $137M build in the “Section 45X credit receivable” to $352.6M — NXT recognizes the credit in earnings ahead of collecting the cash — plus a ~$127M contract-asset (unbilled) build. Reported ROIC (54.3% → 36.9% → 28.9%) and ROE (~30% FY26) look elite, but are 45X-inflated: ex-45X net income is ~$274M, implying ex-45X ROE ~14% and ex-45X ROIC in the mid-teens — respectable, not exceptional.

Verdict. Mixed, and lower-quality than the headline. This is a net-cash, capital-light, cash-generative business — but the reported margin expansion, the 30% ROIC, and half of operating income are a government manufacturing subsidy, not durable operating improvement. Ex-45X, gross profit is flat, gross margin is ~22% and falling, and operating income is shrinking. Economics do not clearly improve with scale once the credit is stripped, and a growing on-balance-sheet §45X receivable means earnings are running ahead of cash. Quality of earnings is below what the income statement advertises.

7. Capital Allocation

The cash-generation profile. NXT throws off ~$500–620M of FCF against a ~$1.1B net-cash balance sheet and no debt — a genuine capital-allocation opportunity set. What management does with it is, so far, reinvestment plus a string of tuck-ins, with minimal return of capital and one first-of-its-kind larger deal now in flight.

Reinvestment. R&D nearly tripled in two years ($42.4M → $79.4M → $120.9M; ~3.4% of revenue) and SG&A rose 86% — funding the “end-to-end solar platform” repositioning behind the November-2025 rebrand from Nextracker to Nextpower. Whether this spend earns a return is unproven: it coincides with declining ex-subsidy operating profit, so the market is being asked to fund an investment cycle on faith.

M&A — serial small tuck-ins, now stepping up in size. All deals to date are cash-funded and small:

• FY25 — Ojjo (June 2024, driven-pile foundations): goodwill $105.9M, intangibles $49.7M, ~$145M cash.
• FY26 — Bentek, OnSight, Origami Solar, Fracsun (May–Nov 2025; power electronics, monitoring/software, US steel, soiling sensors): goodwill $117.9M, intangibles $25.8M, ~$129M cash total.
• Post-FY26 (announced, not in FY26 accounts) — Prevalon Energy (BESS/storage, up to ~$365M) and Zimmermann PV-Steel Group (up to €330M, June 2026 — European steel/foundations, +15 countries).

The pattern is vertical integration (foundations, steel, electronics) plus adjacency diversification (storage, software) — buying growth and optionality beyond the core single-axis tracker. Multiples are undisclosed but the deals are small; no impairments have been taken. Zimmermann (~€330M) is the first sizeable acquisition and a genuine step-up in integration and capital-at-risk — worth watching, given that growth is now partly acquired and the platform thesis is unproven.

Return of capital is minimal. No dividend (10-K: none intended in the foreseeable future). A $500M buyback was authorized January 27, 2026 (3-year term) but only ~$0.4M had been used by March 31, 2026 ($499.6M remaining) — a token authorization, not yet a program. With the stock having run +100% in a year, the absence of execution is defensible, but capital return is not yet a real lever.

Incentive alignment. Per the 2026 proxy, FY26 short-term incentive metrics are revenue + adjusted operating income (switched from FY25’s adjusted EBITDA + adjusted FCF), and the LTI/PSU plan uses EBITDA + FCF (one- and three-year) with a three-year relative-TSR modifier and service vesting. The rTSR modifier is a positive — it imposes relative-performance discipline. But there is no return-on-capital (ROIC) metric and no EPS metric anywhere in the plan. That is a mild but real negative: the plan rewards scale (revenue, EBITDA, operating income) — precisely the metrics that the §45X subsidy and bolt-on acquisitions inflate — rather than capital efficiency or per-share value.

Insider behavior. Across a wide sample of the 184-filing Form 4 corpus (recent May–June 2026 plus a stratified scan), transaction codes are grants (A), option/RSU conversions (M), Up-C/exchange dispositions (J) and open-market sales (S) — zero code-P open-market purchases. Named sellers include founder-CEO Daniel Shugar, CFO Charles Boynton, and Chairman/former-CEO Howard Wenger. This is a routine grant-vest-sell pattern for a founder-led post-IPO company — neutral-to-mildly-negative (no conviction buying). The heavy Flex/TPG stake exit ran through registered secondaries in 2023–24, before this window, and is now complete.

Verdict. Prudent but unproven. Management has kept the balance sheet clean (net cash, no leverage), avoided value-destructive megadeals, and taken no impairments — all good. But the record is short, the reinvestment is running into declining ex-subsidy profitability, return of capital is a placeholder, incentives ignore returns on capital, and insiders are net sellers. NXT has not yet demonstrated it is a skilled capital allocator; the Zimmermann deal and whether the platform spend earns a return are the tests ahead.

8. Changes and Headwinds — Last Two Years

The last two years reshaped the company from a Flex-controlled carve-out into an independent, acquisitive platform — and swapped one set of overhangs (parent control, tax-receivable complexity) for another (policy-cliff demand timing, roll-up integration risk).

Full independence from Flex. The Up-C structure that made GAAP book equity negative through FY23–FY25 has unwound. Flex completed its exit via the January 2, 2024 distribution, and TPG exchanged its remaining LLC units in February 2025; the company now trades as a single-class, widely-held C-corp. (Interpretation: this removes the parent-overhang discount and simplifies the equity story — book equity has since flipped positive, ~$15/sh BVPS FY26 — even if it added ~140M public Class A shares to the count.)

The rebrand — Nextracker → Nextpower (November 2025). The name change is thesis-central: it marks the deliberate strategic shift from a single-product “solar tracker” company to an “integrated utility-scale energy technology platform” spanning trackers + foundations (Ojjo/Zimmermann) + eBOS (Bentek) + steel frames (Origami) + robotics (OnSight) + storage/BESS (Prevalon) + power conversion/inverters. Bull read: TAM expansion and cross-sell. Bear read: dilution of a focused, high-ROIC tracker moat into adjacencies where the advantage is unproven.

From pure-play tracker vendor toward an adjacency roll-up. Management has spent the period buying its way down the balance-of-system stack and into storage (see Growth and Capital Allocation for the full deal list). The most thesis-relevant moves are Prevalon (storage, ~May 2026) — the entry into the utility-scale storage TAM — and Zimmermann (up to €330M, June 2026) — the first sizeable deal, adding European steel/foundation capability and ~15 countries.

New products. June 2026 brought the NX Gemini 2P (two-portrait tracker) and NX Anchor integrated foundation (European-portfolio expansion), on top of continued NX Horizon-XTR terrain-following and TrueCapture control-software upgrades.

Raised guidance and a sell-side re-rating. With the Q4/FY26 print (May 12, 2026), management raised FY27 guidance (~May 29), triggering a wave of PT increases into the $142–182 range (JPMorgan $174, BNP $182, KeyBanc $164, Jefferies $159, Wells Fargo $151, RBC $149; Mizuho lifted to $142 while staying Neutral). Record backlog >$5.25B; investment-grade rating achieved.

The policy pivot — the swing variable (Fact + Interpretation). The IRA→OBBBA (2025) transition cut two ways. §45X manufacturing credits were preserved — central to US utility-solar economics and to NXT’s domestic-content edge (and, as the Financial Quality section shows, to more than half its operating income). But the §48E/45Y project ITC/PTC credits were put on an accelerated phase-out, which pulls demand forward and creates a post-cliff demand-timing air-pocket risk for utility-scale solar. A late tailwind: on June 30, 2026, solar names rose on reports the administration may ban foreign inverters — a domestic-content positive, and directly relevant to NXT’s accelerated entry into power conversion. Module AD/CVD tariffs raise system costs but are largely a module, not a tracker, issue.

Verdict. On balance these developments strengthen the operating thesis — independence, a storage TAM via Prevalon, European reach via Zimmermann, raised guidance, preserved §45X — while adding two real risks: a policy-driven demand-timing cliff (the §48E/45Y phase-out) and roll-up integration risk across seven-plus acquisitions in ~24 months. The business is bigger and broader; the tape now prices much of the upside.

9. Risk Analysis

Catastrophic / total-loss risk: Low. Net cash ~$1.1B, no meaningful debt, sustained GAAP profitability, ~29% ROIC (albeit subsidy-supported), ~$5.25B backlog. A permanent impairment of capital would require a near-total, durable collapse of US and international utility-solar demand — not the base case. The realistic downside is a valuation-plus-cyclical de-rate (~−50%), not insolvency. The cluster that matters is 1 + 2 + 3 acting together: a policy-driven demand air-pocket that both cuts numbers and collapses the growth multiple simultaneously.

10. Valuation Discussion — Embedded Expectations

No price target, no recommendation; this section reads the price as a set of embedded assumptions.

The multiple, at spot. At $112.84 (2026-07-02 close, ~28% below the $163.13 ATH, +114% off the $52.61 52-week low), on ~156M diluted shares, NXT carries a ~$17.7B market cap. With net cash of ~$1.095B (no meaningful debt), EV ≈ $16.6B — one of the cleaner balance sheets in the group, and an underappreciated differentiator versus leveraged peers.

Source: ROIC.ai; NXT Q4/FY26 press release (Ex-99.1 to 8-K, 2026-05-12); AZI valuation_index 2026-07-01. FY26 adj-EBITDA ~$830M and adj-EPS ~$4.5 are built from quarterly prints.

Against its own (short) history. AZI’s own-history percentiles put NXT at the 84.5th composite percentile — P/E 82nd, P/B 79th, P/S 92.9th (near the richest since the Feb-2023 IPO). This is a ~3.4-year history spanning a policy-driven boom, so the percentile is context, not a verdict — but on sales, the market has essentially never paid more for this revenue stream (FACT).

The FY27 guide the market is looking through. Management raised FY27 guidance on May 12–29, 2026 (FACT): revenue $3.8–4.1B (from $3.6–3.8B), adjusted EBITDA $825–900M (from $800–900M), adjusted diluted EPS $4.21–4.59, GAAP EPS $3.19–3.56, on a record >$5.25B backlog. Superficially bullish. But decompose it:

• Revenue midpoint (+11%) far outruns adjusted EBITDA midpoint (+~4% vs ~$830M) and adjusted EPS (~flat vs FY26); GAAP EPS is guided down (~$3.38 vs $3.83). (INTERPRETATION.)
• The Q4 FY26 print already shows the compression: adj-EBITDA margin fell to 22.9% from 26.2% a year earlier, and adjusted EPS fell 19% y/y ($1.05 vs $1.29) and sequentially ($1.10 → $1.05). (FACT.)
• Two mechanical drags: ~$50M of incremental power-conversion (inverter) entry costs baked into FY27, and a declining §45X vendor-rebate tailwind (~$47M in Q4 FY26 vs $67M a year earlier). (FACT/INTERPRETATION.)

So the market is not paying ~25–33x forward for the guided year — the guided year delivers roughly flat per-share earnings. It is paying for re-acceleration beyond FY27 (international ramp, storage via Prevalon, power conversion maturing). That is the load-bearing assumption.

Reverse-DCF / what’s embedded. Two triangulating reads:

1. FCF-yield (Gordon). At a ~2.9% adjusted-FCF yield (company’s $513.6M — itself down 17% y/y on the working-capital/§45X-receivable build) and a ~9–10% cost of equity, the price embeds ~6–7% perpetual FCF growth. That is undemanding if — and only if — FY26 FCF is a durable, through-cycle base. It becomes very demanding if OBBBA’s §48E/45Y phase-out resets the US demand base downward in FY28.
2. EBITDA multiple. ~19–20x forward EV/adj-EBITDA for a steel-and-motors hardware maker is a growth-equity multiple. The guide itself supports maybe 12–15x on flat-to-mid-single-digit EBITDA growth; the gap to 19–20x is the market underwriting sustained mid-teens growth well past the guided year, with margins holding low-20s despite commoditizing hardware. That is the bet.

Is the premium to ARRY deserved? Largely yes. Array trades at ~EV/Sales 1.78x on an ~8.6% EBITDA margin (vs NXT ~20%), with net debt plus a preferred overhang and slower growth — NXT’s ~2.6x EV/Sales premium is margin- and balance-sheet-justified. The harder comparison is FSLR, at just ~8.5x EV/EBITDA / ~5.9x P/FCF despite an equally policy-levered profile: NXT trades at ~2.7x FSLR’s EBITDA multiple. The bull defense is that FSLR’s earnings are fused to §45X (strip the subsidy, ~10% gross margin) and its bookings turned negative, so the market caps its multiple; NXT is asset-light, higher-ROIC, and still booking records. The bear reading is that both ride the same US policy stack — and NXT’s own earnings are also ~half §45X — yet the market extends NXT far more benefit of the doubt.

Scenario analysis (3-year, to ~FY29 exit; ~158M shares; net cash builds).

The bull band coincides with the top of the sell-side range ($151–182). The distribution is asymmetric to the downside: the base case roughly re-earns today’s price over three years (a mediocre IRR for the risk), while the bear case is a policy-triggered halving. ASSUMPTIONS throughout; no single point target.

Verdict. A genuinely high-quality, net-cash, ~29%-ROIC grower — priced for the quality and for a growth re-acceleration the guided year does not yet show. At ~25–33x forward earnings and ~19–20x forward adjusted EBITDA, with adjusted per-share earnings flat-to-declining and a policy-driven demand cliff plausibly landing in FY28, the price embeds the bull path with little margin of safety. The market is underwriting NXT’s quality correctly and its policy/cyclical risk generously.

11. Variant Perception

Consensus. Sell-side is broadly constructive post-Q4: price targets $142–182 (WF $151, RBC $149, Jefferies $159, KeyBanc $164, JPM $174, BNP $182; Mizuho Neutral $142). The consensus narrative: “domestic-content secular winner” — #1 global tracker share, record $5.25B backlog, raised FY27 guide, net-cash balance sheet, ~29% ROIC, and a policy stack (domestic content, FEOC, potential foreign-inverter ban) that advantages US-content suppliers. In this framing, the ~28% pullback from the ATH is a buyable consolidation in a structural grower.

The strongest bull case. NXT is the category leader in the lowest-cost incremental generation source, riding AI/data-center-driven electricity demand. It converts that into industry-leading unit economics (~20% EBITDA margin vs ARRY’s 8.6%), net cash, and high ROIC, and is extending the platform into foundations, eBOS, robotics, storage (Prevalon), and power conversion (inverters) — each a TAM expansion cross-sold into the same developer relationships. Domestic content and FEOC insulate it from the tariff pressures crushing import-dependent rivals. If international scales and storage/power-conversion mature, high-teens revenue growth with stable margins re-rates the stock toward the top of the sell-side range (bull scenario, ~+45–65%).

The strongest bear case. The FY27 guide the bulls celebrate embeds flat-to-declining per-share earnings: adjusted EBITDA +4%, adjusted EPS ~flat, GAAP EPS down, adj-EBITDA margin already compressing (26.2%→22.9% y/y), and the §45X tailwind fading. Half of reported operating income is §45X; ex-subsidy operating income is shrinking. The record backlog is partly safe-harbor pull-forward ahead of OBBBA’s §48E/45Y phase-out — which sets up a US demand-timing cliff in FY28. The core product is commoditizing steel-and-motors hardware, and the “platform” pivot (rebrand to Nextpower) pushes into lower-margin, unproven adjacencies — dilution of a focused, high-ROIC moat into empire-building. Paying 25–33x forward earnings / ~20x forward EBITDA for that, on a ~3.4-year public track record and a beta-1.36 stock that is the solar factor, offers no margin of safety; a de-rate to hardware-cyclical multiples is ~−50% with no operational catastrophe required.

The 3–5 assumptions that matter most.

1. Does the US ITC/PTC phase-out create a real FY28 demand cliff, or is it backfilled? — the single swing variable for both numbers and multiple.
2. Is the ~20% adj-EBITDA margin (and ~22% ex-45X gross margin) defensible, or does hardware commoditization + mix grind it toward the mid-teens? — Q4’s compression is the tell.
3. Does growth re-accelerate beyond FY27 (international + storage + power conversion), which the ~20x forward EBITDA multiple requires — or does the guided flat-per-share year extend?
4. Is the moat “spec-in + scale + software,” or just first-mover share that ARRY/Chinese entrants erode on price?
5. Is the platform diversification value-accretive or margin-dilutive empire-building? — Prevalon/inverters are the near-term test.

What falsifies each side.

• Falsifies the bull: two-plus quarters of negative book-to-bill or backlog decline, adj-EBITDA margin breaking below ~20% (or ex-45X gross margin below ~20%), or a hard US demand air-pocket as safe-harbored projects clear.
• Falsifies the bear: durable positive bookings and re-accelerating organic revenue through the ITC phase-out, adjusted margins stabilizing at ~22%+, and international/storage revenue proving genuinely additive at group-level margins.

Factor-positioning read (where consensus may be offsides). FactorsToday: beta 1.36, dominant loading Solar-Energy industry +2.2–2.5 (this stock effectively is the solar factor), strongly negative LowVol and negative InterestRate, modest +Quality/+SmallSize; y1 +102% then m3 −13.3%, now ~28% off the ATH; factor-cousins FSLR/TAN/ICLN/QCLN. This is a high-beta clean-energy momentum name that ran hard and is consolidating — not a falling knife, not abandoned value. Consensus is crowded into the “secular domestic-content winner” thesis, and the position is a leveraged bet on two exogenous variables (solar-policy durability and interest rates) that the fundamental bull case tends to underweight. That crowding cuts both ways — it amplifies upside on a favorable policy print and downside on a rate or ITC-phase-out shock — but it is precisely the setup where a consensus anchored on backlog and share can be offsides on the margin trajectory and the FY28 demand base that the tape (the m3 −13% wobble) has only begun to price.

12. Fact vs. Interpretation

13. Open Questions

1. §45X survival into FY28–30. What is the exact solar-component step-down (25%/yr from CY2030, ending 2032) applied to NXT’s specific eligible components, and how much of the $379.9M FY26 benefit survives each year? This is the single biggest driver of normalized earnings — the reported ~30x P/E sits on subsidized EPS; ex-45X operating income is $317M and falling.
2. Is the $352.6M §45X credit receivable fully collectible, and on what timeline? It grew +$137M in FY26 and is why cash conversion fell below 1.0x. Assigned credits self-collect via reduced federal tax; vendor-rebate credits depend on supplier solvency. Any impairment/delay hits both earnings and FCF.
3. What is the true ex-subsidy gross-margin trajectory, and why is it falling (27% → 22%) despite a richer US mix — pricing competition, input cost, or lower-margin acquired/storage revenue? This determines whether the core tracker business has residual pricing power.
4. How much of the >$5.25B backlog is safe-harbor pull-forward, and what does the FY28 US utility-solar demand base look like once those projects clear?
5. Zimmermann (€330M) and roll-up economics — multiple, return hurdles, and whether the platform (foundations + steel + storage + software) genuinely cross-sells or is empire-building masked by 45X-inflated headline profit with no ROIC in comp.
6. Software (TrueCapture/NX Navigator) scale — is there a real, growing, high-margin recurring line, or is it a bundled feature? Undisclosed.
7. Customer concentration — “Customer G” hit ~12% in FY26; is this a one-project spike or a structural dependence?

14. What Must Be True

For the bull case (the stock compounds from here):

1. US demand does not fall off a cliff in FY28. The §48E/45Y phase-out is backfilled by safe-harbored project execution plus international and storage demand — book-to-bill stays positive through the transition. Falsification test: two consecutive quarters of backlog decline or negative net bookings, or a visible FY28 US revenue step-down as safe-harbored projects clear.
2. The core margin holds without the subsidy expanding. Ex-§45X gross margin stabilizes at ~22%+ and adjusted EBITDA margin holds ~20%+ — i.e., the price competition denting ex-credit profit abates and mix/scale offset the fading credit. Falsification test: ex-45X gross margin breaks below ~20%, or adjusted EBITDA margin below ~20%, in FY27.
3. Growth re-accelerates beyond the guided-flat FY27, driven by genuine international attach (RoW returns to growth after the FY26 −11%) and accretive storage/power-conversion revenue at group margins. Falsification test: RoW revenue declines again in FY27, or the platform adjacencies dilute rather than lift group margin.

For the bear case (the stock de-rates ~50%):

1. OBBBA’s project-credit phase-out triggers a real US demand air-pocket in FY28, and the record backlog proves to be pulled-forward demand rather than incremental. Falsification test: US bookings and revenue grow through FY28 with no discernible safe-harbor cliff.
2. The subsidy fades faster than the core recovers. §45X steps down and/or the credit-sharing terms tighten while ex-credit gross margin keeps sliding, exposing a ~22%-and-falling underlying business dressed up as a 32% one. Falsification test: §45X per-unit benefit and ex-45X margin both stabilize, keeping reported margins near 32%.
3. The market re-rates a policy-levered, commoditizing-hardware maker from ~20x EBITDA toward hardware-cyclical multiples (10–15x) as the growth-multiple thesis meets flat per-share earnings. Falsification test: adjusted EPS resumes double-digit growth in FY27–28, defending the multiple.

The bull and bear share one fulcrum: the FY28 US demand base and the ex-subsidy margin. Everything else — share leadership, net cash, asset-light returns, the storage TAM — is agreed. The disagreement is whether today’s ~25–33x forward earnings are the entry point to a decade of compounding or the top of a policy-and-momentum-fueled re-rating on earnings that are half government-issued.

Source appendix follows as Appendix B in the combined report.

APPENDIX A — Standard Diligence Questionnaire

Nextpower Inc. (formerly Nextracker Inc.) — NASDAQ: NXT. As of 2026-07-02. Grounded in the FY26 10-K (nxt-20260331), ROIC.ai, AZI, and FactorsToday. Fact / Interpretation / Assumption labels applied where they matter. Supplemental to the memo; not counted toward the length standard.

General

What thoughtful questions have other investors asked about this company? The five that recur: (1) How much of the reported 32–34% gross margin and ~30% ROIC is the §45X manufacturing credit versus durable operating economics? (Answer, this report: roughly half of operating income; ex-45X gross margin ~22%.) (2) How big is the FY28 US demand air-pocket once OBBBA’s §48E/45Y safe-harbor window closes? (3) Is the record >$5.25B backlog incremental demand or pulled-forward demand? (4) Does the platform pivot (Nextracker→Nextpower: storage, inverters, foundations, eBOS) create value or dilute a focused, high-return tracker moat? (5) Can the ~20% adjusted-EBITDA margin survive tracker-hardware commoditization and a fading subsidy?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: closer to a policy-supported high than a low. Reported margins are inflated by a §45X credit that grew from $121M (FY24) to $380M (FY26) and steps down from 2030; US demand is being pulled forward by the OBBBA project-credit deadlines. Underlying (ex-45X) operating profit is already declining, so the reported peak coexists with an eroding core.

Driven by the external environment or internal actions? Both, but the swing factors are external: IRA/OBBBA policy, interest rates (project IRRs), and tariffs. Internal execution (share leadership, asset-light scaling, backlog conversion) is genuinely strong; it operates on a demand base set by policy.

How stable are revenues? Lumpy and project-cyclical, recognized as product ships to named sites (88% over-time). A >$5.25B backlog and multi-year VCAs provide 12–18-month visibility, but backlog is cancellable and subject to timing/scope changes.

Outlook for products/services? Trackers remain the standard for utility-scale solar; secular demand (AI/data-center load, cheapest incremental generation) is strong. The debate is entirely timing/policy, not product relevance.

How big will this market be — growing, shrinking, domestic or international? Global utility-solar is a multi-hundred-GW/year, growing market; tracker attach is high in mature markets and under-penetrated internationally (the structural growth vector). NXT is 77% US today, and its FY26 international revenue fell 11% — so the “global” story is aspiration ahead of actuals.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Consolidating at the top (a two-strong NXT/ARRY oligopoly with a fragmented, FEOC-constrained tail) — favorable versus the flooded module industry. But the underlying hardware is commoditizing, and ex-45X margins are sliding under price competition.

How profitable is the business (ROIC, ROE)? Reported ROIC 28.9% / ROE ~30% (FY26) — but ~half is §45X; ex-subsidy ROIC is mid-teens, ROE ~14%. Still above cost of capital, not exceptional.

How profitable is the industry — competitors, barriers? Bifurcated: NXT earns real (subsidy-supported) margins; the #2 (Array) has lost money in four of five years. Barrier to entry is bankability/track record (lenders won’t finance unproven 30-year structures), reinforced by spec-in switching costs and steel-fab scale — a real but narrow moat.

Can the business be easily understood? Yes — it sells steel trackers plus control software to solar developers. The one genuinely complex item is the §45X credit accounting (vendor rebates + IRC §6418 assignment booked as COGS reduction), which is where the analysis lives.

Can it be undermined by foreign low-cost labor? Partly — trackers are steel fabrication. But domestic-content ITC adders and FEOC rules deliberately advantage US-content suppliers, and NXT’s asset-light model already localizes fabrication near project sites. The subsidy structure is, in effect, the defense against low-cost foreign competition.

Do brands matter? Not as consumer brands; the “brand” is bankability/track record — an intangible that functions like one (independent engineers and lenders trust the NXT name on a 30-year structure).

What is the nature of competition? Price + bankability + domestic-content compliance + terrain/software features. NXT competes on the value-add layer, not the commodity steel.

Customers’ switching costs? Moderate. EPCs standardize crews, software, and O&M around a tracker vendor and sign multi-year VCAs, but multi-source across projects; no per-plant lock-in.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The installed-base field-performance dataset and bankability credential (intangible, unbooked). Conversely, a large recognized-but-uncollected $352.6M §45X credit receivable sits on the balance sheet — earnings ahead of cash.

Off-balance-sheet liabilities? A Tax Receivable Agreement payable to Flex’s affiliate and TPG (85% of realized step-up tax benefits) — carried at $393.2M (FY26), a multi-decade cash drag (~$27M paid in FY26). Standard operating leases and supplier commitments otherwise.

How conservative is the accounting? Mixed. Revenue recognition is standard percentage-of-completion/over-time. The aggressive-looking item is recognizing the §45X credit in earnings ahead of cash collection (growing receivable) and presenting the subsidy inside gross margin rather than as other income — technically compliant, but it flatters the headline margin and ROIC.

How CapEx-hungry is the business? Very light — capex ~1.4% of revenue; components are made by a 100±facility contract-manufacturing network “with close to no capital investment.” This is the source of the high headline returns.

Capital Allocation & Management

How much FCF, and how is it used? ~$514M FCF (FY26). Uses: R&D/SG&A reinvestment (R&D tripled in two years), serial small M&A (7+ tuck-ins), and a token buyback. Net cash still built to $1.1B. Return of capital is minimal.

Significant acquisitions recently? Yes — Ojjo, Bentek, OnSight, Origami, Fracsun (all small, cash), plus announced Prevalon (storage, ~$365M) and Zimmermann (European steel, €330M — the first sizeable deal). Goodwill $371M → $489M in FY26.

Buying back shares? Only nominally — $500M authorized Jan-2026, ~$0.4M used by FY26-end.

Issuing large amounts of new shares to insiders? SBC is $120.3M (~3.4% of revenue), driving ~2–3%/yr dilution — material but not extreme. The big historical share increase (46M→141M) was the Up-C collapse, not insider issuance.

Compensation policy? STI on revenue + adjusted operating income; LTI on EBITDA + FCF with a 3-year relative-TSR modifier. No ROIC and no EPS metric — rewards scale, which the subsidy and M&A inflate. Mild negative.

Motivations of management? Founder-led (Dan Shugar, CEO); building an “energy technology platform.” Insiders are net sellers (grants/vest/sell; zero open-market buys) — neutral-to-mildly-negative on conviction signaling.

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — Delaware C-corp, Class A common, standard 1099. (Historically an Up-C, now collapsed; a residual TRA remains but the equity is a plain single-class common.)

Dividend policy? None; no dividend intended in the foreseeable future.

How profitable is the business? Reported: highly (19–20% net margin, ~29% ROIC). Ex-45X: respectable (mid-teens ROIC, ~22% gross margin). The gap is the whole point.

Is net income diverging from cash from operations? Yes, newly so — FY26 OCF/NI fell to 0.96x (from 1.4x) as the §45X credit receivable and unbilled contract assets built ~$264M combined. A QoE flag to watch.

Risks & Downside

What would cause the stock to decline? (1) OBBBA §48E/45Y phase-out producing an FY28 US demand air-pocket; (2) multiple compression from ~20x EBITDA toward hardware-cyclical levels; (3) ex-45X margin sliding further / §45X stepping down; (4) a rate back-up (strong negative interest-rate factor loading); (5) a bookings/backlog decline revealing pulled-forward demand.

Risk of a catastrophic loss? Low. Net cash ~$1.1B, no debt, sustained profitability, >$5.25B backlog. Realistic downside is a ~−50% valuation/cyclical de-rate, not insolvency.

Chance of a total loss? Remote — would require a durable, near-total collapse of global utility-solar demand.

Recent News & Events

Has the business environment changed recently? Yes: OBBBA (2025) preserved §45X but accelerated the §48E/45Y project-credit phase-out; June-2026 reports of a possible foreign-inverter ban (domestic-content tailwind, relevant to NXT’s inverter entry). The Nov-2025 rebrand to Nextpower formalized the platform pivot.

Significant acquisitions? Prevalon (storage, ~May-2026) and Zimmermann (€330M, June-2026) — see above.

Change in accounting policies? Direct §45X claiming (by IRC §6418 assignment) began CY2025, adding to the supplier vendor-rebate route — increasing the credit flowing through COGS.

Recent changes — new markets, facilities, management? European expansion (Zimmermann + NX Gemini 2P/NX Anchor launches); accelerated entry into power conversion (inverters, ~$50M FY27 cost); a COO/exec transition (Vinje, mid-2026). Founder Dan Shugar remains CEO; Chuck Boynton CFO.

APPENDIX B — Source Appendix

Nextpower Inc. (formerly Nextracker Inc.) — NASDAQ: NXT. Research date 2026-07-02. Primary sources first. Every material claim in the memo traces to a source below.

Primary — SEC filings (EDGAR, CIK 0001852131)

Primary — quantitative data services

Secondary — industry & comparative context

Notes on data quality & reconciliation

• §45X is the central QoE item. The $379.9M (FY26) / $224.9M (FY25) / $121.4M (FY24) reduction-of-cost-of-sales figures are taken directly from the 10-K MD&A and Note 13; the ex-45X margin bridge is arithmetic on those disclosed figures. All ROIC/AZI-derived margins are reconciled to the filing.
• Reported ROIC/ROE are subsidy-inflated. ROIC.ai’s return_on_cap prints negative (an artifact of the historical Up-C negative book-capital structure) and was disregarded; return_on_inv_capital (28.9% FY26) is used, with the ex-45X adjustment noted in the memo.
• Enterprise value: ROIC.ai’s FY26 EV prints negative (it sets market cap to zero for the period) and was disregarded; EV is computed from spot market cap (~$17.7B at $112.84 on ~156M diluted shares) less ~$1.095B net cash ≈ $16.6B.
• Name change: the company was renamed from Nextracker Inc. to Nextpower Inc. in November 2025 (10-K cover + MD&A); ticker NXT unchanged. Third-party feeds (ROIC.ai, FactorsToday) still label it “Nextracker.”
• Management commentary (guidance, backlog, “platform” framing) is treated as hypothesis and validated against filings and competitor data per the research protocol.