Globus Medical, Inc. [GMED] — Tier-2 Research

Analysis date: 2026-07-29 · Framework: Criteria, 2026-07-29 (investment-memo v1.7.0) Price used: $80.64 (Alpaca IEX daily close, 2026-07-28) This document issues no position verdict. It scores the Criteria and reports evidence. Whether the analysis justifies a position is a question for a particular book.

0. Why this memo was written

GMED topped the Tier-1 screen at +34.6pp of implied-path margin (42.2% demonstrated CAGR vs 7.6% required). It is simultaneously a short in a professional healthcare long/short book used as a reference, on the thesis "NVRO impaired, spine market rolling", target $62 against a then-price of $80.58 (−23%).

Both cannot be right in the form stated. The purpose of Tier 2 here was to resolve the disagreement on primary evidence. The answer is that neither is right as stated, and the correct view is closer to the reference book's diagnosis than to the screen's — but it points the other way on the conclusion. Section 6 states this plainly.

1. Screen-input validation — five errors found, all material

Every Tier-1 input was re-derived from primary filings (SEC EDGAR XBRL companyfacts CIK 0001237831, the FY2025 10-K filed 2026-02-24, and the Q1 2026 10-Q filed 2026-05-07). Data recency: the latest filing is 83 days old as of the analysis date — current, not stale.

The TTM error is the exact "TTM that skipped Q4" defect flagged in the batch brief. The screen summed Q1'25 + Q2'25 + Q3'25 + Q1'26 — four tagged quarterly periods, but not four consecutive ones. US registrants do not file a Q4 10-Q, so Q4 was absent from the tag set and the window silently spanned five quarters with one duplicated. Q4'25 revenue is recoverable as FY2025 ($2,938.9m) minus the disclosed nine-month figure ($2,112.5m) = $826.4m.

The gross-margin error is a period mismatch. GMED stopped tagging us-gaap:GrossProfit after Q2 2024. The last tagged annual value is FY2023's $1,020.3m. Divided by FY2025 revenue of $2,938.9m that gives 34.72% — which is exactly the screen's 34.7%. The number is a ratio of two different years. GMED's actual gross margin is 67.4%.

Verified share count. Class A 113,236,099 + Class B 22,430,097 = 135,666,196 at 2026-03-31; the 10-Q cover reports 135,744,626 at 2026-05-05. The screen's 135,744,626 is correct. No 4.1x or 54% error here.

One judgement, disclosed: GMED carries $100.3m of "business acquisition liabilities" (contingent consideration, current + non-current). Treated as debt-like, net cash falls to $699.0m and EV rises to $10,247.5m. The headline figures use the $799.3m definition; the reverse DCF was not re-run on the alternative because a $100m move on a $10.1bn EV shifts the required CAGR by under 20bp.

2. Accounting quality — the finding

2.1 Is the reported growth real? Mostly not.

GMED's 42.2% three-year revenue CAGR (FY2022 $1,022.8m → FY2025 $2,938.9m) is almost entirely acquired.

Two mergers sit inside that window: - NuVasive, closed 2023-09-01, all-stock, aggregate consideration $2.604bn (net identifiable assets $1.394bn, goodwill $1.210bn). Roughly doubled the company. - Nevro, closed 2025-04-03, all-cash, total purchase price $252.5m ($5.85/share).

The company itself publishes the organic series, and it is not 42%:

Sources: FY2025 8-K Ex-99.1 (2026-02-24) "Net Sales Reconciliation of the Nevro Acquisition Table"; Q1 2026 8-K Ex-99.1 (2026-05-07); Q3 2025 8-K Ex-99.1 (2025-11-06). Nevro Q2'25 and Q3'25 are derived from the disclosed nine-month Nevro total of $293.6m less the disclosed Q3 and Q4 figures.

Trailing-twelve-month organic growth is +8.5% (Q2'25–Q1'26 base $2,724.3m vs prior four quarters $2,510.8m). That is the number this memo treats as demonstrated, not 42.2%.

2.2 The margin expansion is 62% purchase accounting

This is the finding that matters most, and no screen could see it. Operating income rose from $165.99m (FY2024) to $479.82m (FY2025), +$313.8m — a +9.7pp margin swing that the screen recorded as evidence of operating leverage. The segment note (FY2025 10-K, Note 20) discloses the driver:

The NuVasive inventory step-up burned off. That single line swings +$195.9m year over year — 62% of the entire $313.8m operating-income increase. MD&A states it directly: "The $77.7 million, or 7.5%, decrease in cost of sales for the year ended December 31, 2025 was primarily driven by the NuVasive amortization of inventory fair value step-up of $215.4 million included within the December 31, 2024 balance as compared to the Nevro amortization of inventory fair value step-up of $19.3 million."

Restating gross margin for it:

Ex purchase accounting, gross margin is essentially flat at ~67–68%. The apparent 8.5pp gross-margin expansion in FY2025 is a reversal of a FY2024 accounting charge, not a manufacturing gain.

2.3 A $117.7m non-cash, largely non-taxable gain runs through net income

The Nevro purchase price ($252.5m) came in below the fair value of net assets acquired ($370.3m), producing a $117.7m bargain purchase gain. It arises almost entirely from recognising $144.9m of Nevro deferred tax assets (pre-acquisition federal NOLs with an indefinite carryforward). "The majority of the bargain purchase gain is non-taxable."

It sits in other income, below the operating-income line — so operating income is clean of it. Net income is not. FY2025 GAAP net income of $537.9m and diluted EPS of $3.92 include it in full:

Ex-bargain-purchase GAAP diluted EPS is ~$3.06, not $3.92. Management's own non-GAAP EPS of $3.98 excludes it, so the non-GAAP series is not affected — but any screen reading GAAP EPS growth of +425% off this year is reading an accounting event.

2.4 What did NOT turn up

Checked explicitly, against the precedents named in the brief: - No non-recurring settlement income inflating revenue (the 10x Genomics pattern). The only settlement items run through SG&A: a $43.1m Pimenta litigation accrual in Q3 2025 (a headwind), partly offset by $5.7m of net settlements received. FY2025 "provision for litigation" was $37.7m vs $0.3m in FY2024. - No concentrated distributor (the AAOI pattern). GMED sells through a direct US sales force plus distributors; no single customer is disclosed as material. Receivables $686.4m on TTM revenue $3,100.6m = DSO ~81 days, normal for spine implants (consignment inventory, hospital payers). Not the ~264-day AAOI signature. - Inventories $772.6m are high relative to revenue (91 days) but that is structural for spine — sets are consigned to hospitals. Inventory grew 1.8% QoQ against revenue +27% YoY: no channel build. - Auditor (Deloitte) named the Nevro business combination a critical audit matter and excluded Nevro from the ICFR assessment (8.5% of assets, 10.0% of revenues). Standard for a first-year acquisition; disclosed here because it is the one control-scope gap in the file.

3. Mechanism — the specific, evidenced driver

The mechanism is the recovery of the NuVasive sales force, now visible as five consecutive quarters of monotonic organic acceleration, plus a second, later leg from Enabling Technologies coming off a trough.

This is not a narrative. It is a series:

Base-business YoY:  Q1'25  −1.4%  →  Q2'25  +3.3%  →  Q3'25  +7.0%  →  Q4'25 +10.6%  →  Q1'26 +13.2%

Five quarters, monotone, each figure disclosed by the company. The mechanism underneath it:

1. US Spine is the engine and it is now compounding at 10%. Management: "US Spine continued to lead the way in growth... posting the third straight 10% growth quarter" (Q1 2026 release). Q4 2025: "our US Spine business capped-off 2025 by growing revenue 10% over the prior-year quarter." Q2 2025 was +5.7% as reported / +7.4% day-adjusted, and described as "its highest sequential revenue growth since the second quarter of 2022." The 2024 organic decline was NuVasive integration disruption — competitive rep attrition and overlapping territory consolidation — and it has been worked through.

2. Enabling Technologies (imaging/navigation/robotics) inflected in Q4 2025 after a two-year decline. FY2024 $154.0m → FY2025 $141.0m (−8.4%), with US ET down $17.3m on lower unit placements. Then Q4'25 +19% and Q1'26 $26.9m vs $22.2m, +21.1%, international ET up $4.6m on increased placement. A capital business coming off a trough is the second leg and it is far earlier in its cycle than spine implants.

3. Portfolio broadening beyond spine is now being disclosed as a growth line. Trauma and joint reconstruction appear as named contributors in the MD&A for the first time from Q2 2025 (see §5, mention-frequency: both terms are first-ever mentions in the 2025Q2 release and recur every quarter since).

4. Operating leverage on the base is real and separable from the accounting. The company publishes base-business Adjusted EBITDA excluding Nevro:

+4.2pp on the full year, +5.7pp in Q4 — on the base business, with Nevro carved out and with no inventory step-up in either the numerator or the comparison. This is the operating-leverage evidence that survives §2.2.

Corroboration that the mechanism is being funded, not just described: the FY2026 non-GAAP EPS guide has been raised twice inside six months — $4.30–4.40 (initial) → $4.40–4.50 (Feb 2026) → $4.70–4.80 (May 2026) — while the revenue guide of $3.18–3.22bn was reaffirmed both times. Guidance is being raised on margin, not mix.

4. Product-cycle intelligence

Named products and where they are in the cycle:

• Excelsius robotic platform (assembled Methuen, MA) — the ET franchise. Placements fell through 2024–early 2025, recovered from Q4 2025. Management frames the next step as the "closed-loop surgical intelligence ecosystem" linking patient selection, planning, implants and robotics — first used in the Q4 2025 release and repeated in Q1 2026 (both first-ever mentions; see §5). No dated launch is disclosed, so none is asserted here.
• HFX / Senza spinal cord stimulation (Nevro), manufactured in Costa Rica — plus Nevro's minimally invasive sacroiliac joint offering. This is the asset the reference book calls impaired. See §6.
• Spine implantables, biologics (San Antonio, TX and Audubon, PA), interbody, and intraoperative neuromonitoring — IONM contributed +$10.1m of the US increase in Q1 2026 on its own, a services line that is under-discussed.
• Trauma, hip/knee and extremities — small, but newly named as contributors from Q2 2025.

What is not in the file: GMED does not disclose unit placements for Excelsius, procedure volumes, or franchise-level margins. Anything of that kind would have to come from a source this memo does not have, so it is omitted rather than estimated.

5. Transcript mention-frequency (required core metric)

Corpus: SEC EDGAR 8-K Exhibit 99.1 quarterly earnings releases, 10 consecutive quarters, 2023Q4–2026Q1. Alpha Vantage EARNINGS_CALL_TRANSCRIPT was unavailable — the shared 25/day quota was exhausted by other work on 2026-07-29 (the API returned the rate-limit notice, not data). Per references/mention-frequency.md, the first-party 8-K Ex-99.1 corpus is the designated substitute: complete, near-constant length, and one source used consistently across the whole series. Pre-announcement releases (Jan 2024, Jan 2025, Jul 2025) are excluded — they are a different document type and mixing them would manufacture artifacts.

Counts are normalised per 10,000 words (word counts 3,627–4,740; raw counts also shown).

Read — emerging: operating leverage, margin, ecosystem, surgical intelligence, trauma, joint reconstruction, US Spine. Read — decaying: NuVasive, synergies, integration. Read — U-shaped: Enabling Technologies — troughs at 2.1/10k in 2024Q4 exactly when ET revenue was falling, and reaches its series high of 7.8/10k in 2026Q1 exactly as ET revenue turns +21%.

The generative finding. Run before the valuation work, the anomaly the language flags is not Nevro (which is merely a large acquisition being talked about). It is that the vocabulary of repair — NuVasive, synergies, integration — has been fully replaced by the vocabulary of compounding — operating leverage, margin, ecosystem, surgical intelligence — inside two quarters, and that trauma and joint reconstruction entered the disclosure set from nothing in 2025Q2 and have persisted every quarter since. The first is corroborated independently by the base-EBITDA-margin series (§3, 29.2% → 33.4%) and by the twice-raised EPS guide. The second is corroborated by the MD&A naming them as contributors. Neither claim rests on the mention counts alone.

Honesty constraints on this table, stated because they are real: - The raw counts are small (1–3 for most emerging terms). A single-mention change is noise. The only terms with a robust shape at raw-count scale are Nevro (0→10–12), NuVasive (10→0) and restructuring (0→9, stable). Everything flagged as "emerging" above is a hypothesis corroborated elsewhere, never a standalone finding. - A press release is a curated document, more so than a transcript. It shows what management chose to headline, which is the intended signal, but it carries none of the Q&A that reveals what the sell side is asking. - Window is 10 quarters. Nothing here is a claim about "all history."

6. The disagreement with the reference book — resolved on evidence

The reference book is short GMED at a $62 target on "NVRO impaired, spine market rolling." Taking the two legs separately.

Leg 1: "NVRO impaired" — CORRECT on the business, WRONG on the consequence.

Nevro's revenue is eroding under Globus ownership:

Trailing twelve months −7.9% versus Nevro's last standalone year, and Q1'26 is −18.8% against Q1'24 on a two-year stack. Nevro's own history was already flat-to-down (FY2022 $406.4m → FY2023 $425.2m → FY2024 $408.5m, −4.0%). The CFO's "Nevro performed as expected as we continue to diligently work through the integration" is a managed way of saying the same thing. The reference book's read on the asset is right.

But the impairment consequence does not exist, for a specific accounting reason. GMED paid $252.5m for $370.3m of fair-valued net assets. That produced a bargain purchase gain, not goodwill. The only Nevro intangibles on the balance sheet are $56.0m (developed technology $36.0m / 7yr, customer relationships $11.5m / 10yr, trade names $8.5m / 15yr). There is no Nevro goodwill to write down. Total company goodwill of $1,438.7m is NuVasive-and-legacy, not Nevro.

Against that $252.5m price, Nevro delivered $36.0m of Adjusted EBITDA in nine months, of which $21.2m came in Q4 2025 alone — an exit rate of roughly $85m annualised on a $252.5m purchase price. The deal moved from dilutive to accretive inside two quarters; the Q3 2025 release states "The Company now expects its Nevro acquisition to be accretive to earnings in 2025." A declining asset bought at 0.6x sales that yields ~$85m of run-rate EBITDA against a $252.5m price is a good outcome, not an impairment. The GAAP net loss of $37.5m over the period is after step-up amortisation, intangible amortisation and allocated integration cost.

The honest cost of Nevro is dilution to the group growth rate, not a write-down: it is ~12% of revenue shrinking mid-single-digit against a base growing 13%.

Leg 2: "spine market rolling" — NOT SUPPORTED, and contradicted by both companies' filings.

If the US spine market were rolling, the two largest challengers should both be decelerating and the incumbent's organic growth should be compressing. Neither is happening in the direction the thesis needs:

ATEC data: FY2025 $764.2m, FY2024 $611.6m; quarterly from EDGAR XBRL, Q4 derived as FY less nine months.

GMED is accelerating into a market the thesis says is rolling. ATEC is decelerating — but from +30% to +13.5%, which is share-gain maturation, not market contraction. The two are converging at ~13%, both far above any plausible ~4–6% US spine market growth rate. That combination is only consistent with the large incumbents (Medtronic, Stryker, J&J spine) losing share to both — which is a share story, not a market story.

Corroborating detail the thesis has to explain away: GMED's US Spine business has grown 10% for three consecutive quarters (management-stated, Q3'25 / Q4'25 / Q1'26), and Enabling Technologies — the most capital-cycle-sensitive line, the first to break if hospital budgets were rolling — turned +19% in Q4'25 and +21% in Q1'26 after a −8.4% year.

What the reference book may be seeing that this memo is not

Stated because it is the honest counterweight, not to be even-handed for its own sake:

1. GMED's own guidance implies the acceleration stops. FY2026 revenue guidance of $3.18–3.22bn, with Nevro at the Q1'26 run rate (~$331m), implies base-business revenue of ~$2,869m, i.e. ~+8.5% organic — well below Q1'26's +13.2%. Either management is sandbagging (they raised EPS twice while holding revenue, which is what sandbagging looks like) or they expect the comp to get harder. This is the single most falsifiable point in the long case, and Q2 2026 results resolve it.
2. Q1'26's +13.2% is against Q1'25's −1.4% — the easiest comp in the series. The two-year stack is $677.2m vs Q1'24 $606.7m = +11.6% over two years, ~5.6% annualised. The acceleration is real; its level is flattered by the base.
3. Momentum agrees with the short on the near term. 6-1 momentum is −7.2%, three-month −11.1%, and the stock is below its 200-day ($80.64 vs $83.61) despite 12-1 momentum of +46.6%. The tape has been rolling over since spring even as the fundamentals accelerated.
4. Relative value is not on the long's side. GMED trades at a +0.99x EV/Sales premium to ATEC, the 77th percentile of that spread's 3.5-year history (§Valuation). Cheap absolutely, rich relatively.

Verdict on the disagreement

The evidence supports the reference book's diagnosis of Nevro and rejects its diagnosis of the spine market — and it rejects the screen's characterisation of GMED as a 42% grower entirely.

The correct description of GMED is neither: it is a ~9–13% organic grower with genuine and separately evidenced operating leverage, carrying a ~12%-of-revenue declining asset it bought for 0.6x sales, at 3.27x EV/Sales (13th percentile of its own five-year range) and 17.0x guided FY2026 non-GAAP EPS. The screen's +34.6pp margin is manufactured; the honest margin is +2.8pp (§Valuation), which is a pass, but a thin one that inverts below a 17.5x exit multiple.

7. Criteria scorecard

8. Downside Criteria — the permanent-loss case with a named cause

MEASURED. Does not block. Logged to the ledger.

Named cause: the organic acceleration is a NuVasive-disruption rebound, not a durable share gain, and it reverts to the ~5% base-business rate of FY2025 while Nevro continues to shrink.

The mechanism, specifically: GMED's 2024 organic decline was self-inflicted (integration disruption). Recovering lost territory produces a 12–18-month growth bulge that ends when the territory is recovered. Q1'26's +13.2% sits on a −1.4% comp; the two-year stack is +5.6% annualised. The company's own FY2026 guidance implies ~+8.5% organic, decelerating from Q1's 13.2% — which is either conservatism or management's actual view of the fade.

Quantification. At a 5% terminal organic rate and no Nevro recovery, the terminal EBIT margin thesis (20.0%) is unlikely to be reached — an implanted-device business without volume leverage stalls nearer 16–17%, GMED's FY2025 level. Re-solving the reverse DCF at a 16.5x exit multiple (RMD, the lowest medtech anchor) and a 17% terminal margin gives a required CAGR the business would not clear, and the price that clears it at a 5% delivered CAGR is roughly $58–62 per share, −24% to −28%.

That range brackets the reference book's $62 target. Their target is a coherent output of a scenario that is not the base case here but is entirely reachable, and it is reachable on the same arithmetic — the disagreement is about the organic growth rate, not about the valuation method.

Probability assigned: 30%. Logged to the recommendation ledger for Brier scoring.

Going concern: not applicable, and stated explicitly. GMED has zero financial debt, $799.3m of net cash, $588.7m of FY2025 free cash flow and $4.7bn of book equity. There is no solvency path to permanent loss. The permanent-loss case is a multiple-and-growth case only.

Second-order downside, named but not sized: the Pimenta litigation, for which $43.1m was accrued in Q3 2025. The accrual is disclosed; the ultimate exposure is not, and no number is asserted here.

9. Disclosed limitations

1. No Street consensus. Alpha Vantage quota was exhausted on the analysis date. Every forward figure in this memo is either company guidance (labelled) or a house estimate (labelled). No consensus number is quoted, and none is invented.
2. Mention-frequency corpus is press releases, not transcripts (see §5). Different instrument, same method, one source used consistently. The Alpha Vantage transcript series should be built when quota permits.
3. No FY2027 guidance exists. The FY2027 revenue figure used in the 12-month target is a house extension, marked as such at the point of use.
4. Pro-forma NuVasive combined revenue is not disclosed in the FY2025 10-K (Nevro was below the ASC 805 materiality threshold for pro-forma disclosure at ~10% of revenue), so a clean FY2022→FY2025 organic CAGR spanning both mergers cannot be constructed. The TTM organic figure (+8.5%) is used instead and its window is stated.
5. Consolidated ADV is not available from the IEX feed; only the partial-tape figure is reported, labelled.
6. No Excelsius placement counts, procedure volumes or franchise margins are disclosed by the company. None are estimated.

Sources

• SEC EDGAR XBRL companyfacts, CIK 0001237831 (GMED), 0001444380 (Nevro Corp), 0001350653 (Alphatec Holdings)
• GMED FY2025 Form 10-K, filed 2026-02-24 (accession 0001628280-26-011209)
• GMED Q1 2026 Form 10-Q, filed 2026-05-07 (accession 0001628280-26-032290)
• GMED Form 8-K Ex-99.1 earnings releases: 2026-05-07, 2026-02-24, 2025-11-06, 2025-08-07, 2025-05-08, 2025-02-20, 2024-11-05, 2024-08-06, 2024-05-07, 2024-02-20
• Nevro Corp Forms 10-K/10-Q through FY2024
• Alphatec Holdings Forms 10-K/10-Q through Q1 2026
• Alpaca Markets daily bars and options chain/snapshots (prices, volatility, open interest, Greeks)