DXCM dominates US continuous glucose monitoring alongside Abbott, but the recent margin reset, sales-force fumble, and OTC pivot have shaved the multiple even as the install base keeps compounding. At $61 versus an $85 base IV, the price now pays you to wait while the Type-2 and OTC opportunities prove out.

Plain English

Dexcom makes a small sensor that sticks on your arm and reads your blood sugar every five minutes, sending the number to your phone. Diabetics who must take insulin rely on it to avoid dangerous lows and to dose correctly. Once a patient pairs the sensor with their insulin pump, phone, doctor, and pharmacy plan, switching to a competitor is a hassle. There are essentially two companies in the world that do this well — Dexcom and Abbott. Dexcom earns most of its money from repeat sensor sales, which is a recurring, subscription-like business. Bigger market over time, but pricing pressure too.

Thesis

Dexcom is a continuous glucose monitoring (CGM) pure play that has spent two decades turning a niche endocrinology product into a chronic-care subscription. Once a patient is paired with a transmitter, app, pharmacy benefit, and insulin-pump integration (Tandem, Insulet, Beta Bionics), the recurring sensor stream becomes high-margin annuity revenue with strong gross margins and 70% FCF conversion (FCF/NI 0.70 over five years per the scorecard).

The scorecard tells two stories. The good: ROIIC of 30.6% over five years on a base of single-digit historical ROIC (10y avg 2.2%) shows that recent incremental capital is finally earning real returns; net cash balance sheet (net debt/EBITDA -1.30x) and 42x interest coverage make the business unbreakable. The bad: a 29.8% rise in share count over a decade (option-heavy med-tech compensation), TTM P/E of 46.7x against a 10y average of 80x — multiple compression is well underway — and a reverse-DCF that requires only ~9.9% growth, which is well below the company's historical and consensus trajectory.

The IV bracket is $57 / $85 / $92, with the current $61.35 near the low end. Owner earnings of $0.67B against an enterprise multiple of 41.5x EV/FCF say you are paying for growth, but the price/IV ratio of 0.72 means even mid-case outcomes deliver a 39% upside. Buy below $70, trim above $92. The asymmetry is real because Dexcom is the rare med-tech where a wide TAM (>500M diabetics globally, only ~1% on CGM today) collides with a near-duopoly structure and pharmacy-benefit-locked recurring revenue.

Moat

Dexcom's moat sits in the narrow-but-real bucket and rests on four reinforcing pillars: switching costs, intangibles (FDA + integration), scale-driven cost advantage, and a duopoly structure that limits price competition.

Switching costs. Like Damodaran's Microsoft Office example [6], Dexcom has spent a decade making it easier to switch IN and harder to switch OUT. Once a Type-1 diabetic — or increasingly a Type-2 on insulin — has paired their G7 with their pump (Tandem t:slim, Insulet Omnipod 5, Beta Bionics iLet), their iPhone, their Apple Watch, their endocrinologist's data dashboard (Clarity), and their pharmacy benefit, the friction to switch to Abbott's Libre is meaningful. The patient must re-train, re-pair, lose pump integration, and often change pharmacy reimbursement tier. This is the classic 'multiple gauntlets' Damodaran describes [6]. Stress test — a $10B competitor with 5 years would not break this; Abbott has $40B+ revenue and has competed for a decade, yet Dexcom still holds roughly half the US CGM market by revenue.

Intangibles. Two FDA Class III device approvals (G7, Stelo OTC), 15+ years of clinical evidence in peer-reviewed journals, and pump-partner exclusivity give Dexcom regulatory moat. As Damodaran notes [2], legal protection 'may not lead to value enhancement' if regulators control prices — and indeed Medicare and PBM pricing pressure is the central bear case. But unlike a pharma patent that cliffs, the CGM moat is the system (sensor + algorithm + app + integrations + reimbursement codes), and each generation (G6 → G7 → G8) refreshes the IP stack.

Cost advantage / scale. Dexcom and Abbott together produce >90% of global CGM sensors. Sensor manufacturing is a high-fixed-cost, high-yield-curve business — the second billion sensors are dramatically cheaper than the first. This mirrors the GEICO economics Buffett describes [5]: scale lets the leader 'enlarge the price advantage we offer customers' rather than just widen margins. Dexcom's gross margins (mid-60s) sit well above any sub-scale entrant could achieve from a standing start.

Duopoly structure / brand. This is not Coca-Cola brand [2] — patients don't ask for Dexcom by name — but prescribing endocrinologists do, and that physician brand is sticky. New entrants (Medtronic Simplera, Roche, Sinocare in China) have launched but failed to dent US share materially.

What the moat is NOT. Dexcom does not have a Mayo Clinic moat [3] — there is no irreplaceable intangible. It also does not have See's Candy economics [3]: it must reinvest heavily in R&D and salesforce. The 10y avg ROIC of 2.2% is the smoking gun — capital reinvested in tooling, biosensor R&D, and inventory has only recently begun earning excess returns (ROIIC 30.6% over five years). Per Damodaran [4], excess returns 'will undoubtedly draw in new competitors over time' — Abbott Libre 3 already matches Dexcom on most clinical specs at a lower list price, and Chinese players are 5–7 years behind but coming.

Erosion risks. (a) OTC commoditization — Stelo (Dexcom's OTC product for non-insulin Type-2 and prediabetics) ironically may narrow the moat by training consumers that CGM is a $50/month commodity. (b) GLP-1 demand destruction — if Ozempic/Mounjaro keep Type-2 patients off insulin, the addressable insulin-using CGM TAM shrinks. (c) PBM consolidation — three PBMs control >75% of US scripts; one bad formulary cycle wiped 20% off the stock in 2024. (d) Apple/Samsung non-invasive sensors — a 5-10 year tail risk, but binary if it lands.

Verdict. The switching cost is real but narrower than bulls argue (Abbott proves you CAN compete head-to-head); the regulatory moat is real but renews each generation rather than compounding; the scale moat is real but shared with Abbott. Net: this is a textbook duopoly economic castle [5] — protective but not impregnable.

Moat verdict: NARROW

Management

Dexcom is run by Kevin Sayer (Chair) and CEO Jacob Leach (promoted 2025), with Jereme Sylvain as CFO. The cultural DNA is bio-medical engineering: capital is allocated as if every dollar must be re-invested into the next sensor generation. This is appropriate for a company still in the early innings of a 500M-patient TAM, but it creates the capital-allocation profile of a growth company, not a compounder in the See's-Candy [3] sense.

Reinvestment — A. R&D as a % of sales runs ~12-13%, producing G6, G7, Stelo, and pump integrations. The 5-year ROIIC of 30.6% is the receipt: incremental capital is now earning real excess returns even though the 10-year average ROIC is only 2.17% (the legacy capital base from pre-G6 era was poorly productive, and the recent franchise economics are masked by averaging). Direction of travel matters more than the average here.

Acquisitions — B. Dexcom has been disciplined; few large M&A misadventures. The TypeZero (closed-loop algorithms) tuck-in and Pacific Diabetes Technologies stake have been sensible. No empire-building.

Debt — A. Net cash position (net debt/EBITDA -1.30x), interest coverage 42.5x. The 0.25% convertible notes were issued opportunistically when the stock was 4x today's price — a textbook example of issuing equity-linked paper at peak valuation. This is good capital allocation timing.

Buybacks — C. Here the report card gets ugly. Share count is up 29.8% over 10 years. Dexcom has authorized buybacks ($750M in 2024) but the diluted-share-count reduction has been modest and — critically — the buybacks were heaviest when the stock traded at 80-100x earnings. Buying back at 4-5x today's price is wealth destruction. The company has not articulated a price-disciplined buyback philosophy à la Buffett. Even now, with the stock 70% below its 2021 high, the buyback pace is described as 'opportunistic' rather than aggressive.

Dividends — N/A. No dividend, appropriate for a company still reinvesting at 30% ROIIC.

Stock-based compensation. This is the heart of the C grade. SBC runs $300-400M annually against ~$700M of owner earnings — i.e. ~50% of true free cash flow is being paid to employees in stock. The 29.8% 10-year share count growth reflects this. Buffett would call this 'real money.' Bulls argue it is necessary to retain medical-device engineering talent in a hot San Diego biotech labor market; bears argue it inflates GAAP earnings and the buybacks are largely SBC mop-up rather than genuine return of capital.

Communication. Investor day cadence is regular; segment disclosure is reasonable; guidance has been missed twice in the last 18 months (the 2024 sales-force restructuring and the 2025 channel-mix surprise) which dented credibility. Management has been transparent about the misses, which is to their credit, but the operational execution is no longer the metronome it was 2018-2022. Sayer's transition out of CEO into Chair is a reasonable handoff, though the new CEO is unproven on Wall Street.

Insider ownership. Modest — under 1% combined for executives. This is a hired-hand culture, not a founder/owner culture. Munger and Buffett would prefer the latter.

Net. Operationally above-average, capital-allocation discipline mediocre, SBC structure problematic. The recent ROIIC suggests the franchise can outrun the dilution, but only if growth re-accelerates.

Capital allocator: B-

Industry

Continuous Glucose Monitoring — Porter's Five Forces.

Threat of new entrants — Low to Moderate. Capital intensity is modest by med-tech standards (~$500M to build a global sensor line) but regulatory intensity is brutal: FDA Class III approval requires multi-year clinical trials, and pump-integration partnerships are negotiated 3-5 years in advance. Medtronic and Roche have full diabetes franchises and still trail. Apple has been rumored for a decade and has not shipped non-invasive glucose. Chinese OEMs (Sinocare, Microtech) are gaining share in EM markets but absent in US/EU because of reimbursement gates. Threat is real but slow-moving.

Bargaining power of suppliers — Low. Sensor inputs (platinum wires, enzyme chemistry, plastic housings, Nordic BLE chips) are commodity. Contract manufacturing (Flex, Jabil, in-house Mesa AZ) is competitive. No supplier has Dexcom over a barrel.

Bargaining power of buyers — High and rising. This is the central problem. Three PBMs (CVS Caremark, Express Scripts, OptumRx) control >75% of US prescription volume. They negotiate annually on rebates, formulary tier, and step-therapy. In 2024, Dexcom flagged 'channel mix' (more pharmacy, less DME) as a margin headwind precisely because PBM margin is structurally below DME. CMS/Medicare is also a powerful buyer, having repeatedly cut CGM reimbursement rates. International buyers (NHS, German statutory insurance) are even more price-sensitive. Patients themselves have very little price elasticity (insulin users cannot do without CGM, period), but the insurer in between captures most of that surplus.

Threat of substitutes — Moderate and shifting. Fingerstick BGM (Roche, LifeScan) is the cheap legacy substitute, declining ~5%/year. Flash glucose (Abbott Libre) is technically a CGM but historically positioned as a substitute; that line has now blurred. The wild card is therapeutic substitution: GLP-1s (semaglutide, tirzepatide) reduce A1c so dramatically in Type-2 patients that some never escalate to insulin and therefore never need CGM. Long-tail substitute: implantable 1-year sensors (Senseonics Eversense), non-invasive optical (still vaporware after 20 years).

Rivalry — High and intensifying. Effectively a duopoly with Abbott Libre 3+, but a price-competitive duopoly. Abbott has consistently undercut Dexcom on list price by 30-40% and won the basal-only Type-2 segment. Both are now racing into the OTC / wellness segment (Stelo vs Lingo). The lack of meaningful product differentiation — both are 14-day, factory-calibrated, pump-integrated, Bluetooth — pushes the competition toward price and reimbursement deals.

Value pool location and trajectory. The TAM expands (Type-2 non-insulin, prediabetes, wellness, gestational, hospital). Per-patient revenue compresses (PBM pressure, OTC commoditization). Net dollars: probably grow at a 10-15% CAGR for the duopoly through 2030, but with the value pool shifting from a 65% gross-margin pharmacy product to a possibly 50% gross-margin OTC product. The number of patients multiplies; the profit per patient erodes. Dexcom must run hard to stay in place at the per-unit level.

Industry Verdict: Good (but trending from Excellent toward Average over the next decade as PBM power and OTC dynamics compress per-unit economics).

Inversion

I am short DXCM. Here is why the bulls are wrong.

1. The single event that kills this: a PBM step-therapy edict that mandates Abbott Libre as first-line for the entire Type-2 basal cohort. This already happened in pieces in 2024-2025. The full-fat version — CVS Caremark or OptumRx writing 'Libre 3 first, Dexcom only on documented failure' across their commercial book — would strip 25-35% of Dexcom's growth pipeline overnight. The Type-1 base is loyal, but Type-1 is biologically capped; all the growth math the bulls cite assumes Type-2 share. PBMs do not negotiate on clinical merit — they negotiate on net price per member per month. Abbott has structurally lower COGS (factory-calibrated since launch, single-piece sensor architecture) and can underbid Dexcom indefinitely.

2. Why the moat is narrower than bulls think. Bulls cite switching costs. Reality: in 2024 Dexcom lost meaningful share to Libre 3 in basal Type-2 despite the supposedly sticky pump-integration moat — because basal Type-2 patients aren't on pumps. The pump moat protects the high-end Type-1 pump cohort (~1.5M US patients), not the 30M+ Type-2 TAM that justifies the multiple. Bulls also cite physician brand. Reality: prescribing endocrinologists are increasingly fungible because PBM step-therapy makes the physician's preference irrelevant. Bulls cite scale. Reality: Abbott has 3x Dexcom's overall revenue and 2x its diabetes-care revenue — Abbott has the scale advantage. Dexcom is the smaller player in a duopoly where the larger player chooses to compete on price.

3. Why management is worse than it appears. SBC at ~50% of owner earnings is a structural transfer from shareholders to employees. The 29.8% 10-year share count growth quantifies this. Buyback authorizations are reactive — heaviest when the stock was at $130, modest at $61 — the textbook value-destroying pattern. The 2024 sales-force reorganization was botched (acknowledged by the company); the 2025 'channel mix' miss was foreseeable to anyone watching pharmacy-vs-DME trends. The new CEO transition adds execution risk during the most competitive year in the company's history. Management is competent at engineering, mediocre at capital allocation, and unproven at navigating a margin-compression cycle.

4. What bulls are extrapolating that won't hold.

• Linear OTC scaling. Stelo will hit a wall: most healthy people who try a CGM don't continue past three months because the data is repetitive. OTC churn will be severe and the unit economics will look more like a fitness tracker than a chronic-care subscription.
• International growth at US margins. Germany, France, UK, and Japan reimburse CGM at 40-60% of US prices. Each new international patient is at structurally lower gross profit dollars.
• GLP-1 as tailwind. The bull pivot is 'GLP-1 patients also want CGM.' Possible, but GLP-1s reduce A1c so dramatically that many patients never progress to insulin — the very transition that makes CGM sticky and high-value. The net effect on the value pool over 10 years is more likely negative than positive.
• Margin recovery. Bulls model gross margin recovery to 65%+. With OTC mix shift, international mix shift, and PBM pricing, the structural ceiling is more like 58-60%.

5. Valuation trap — multiple compression in progress. P/E TTM 46.7x against 10y average 80x looks like a discount. It is not. The 80x average reflects a different company — one growing 25%+ with expanding margins. The current company is growing 11-14% with compressing margins. Comparable medical-consumables peers (Insulet, Inspire) trade at 30-35x earnings; commodity device peers (Hologic, Edwards) trade at 20-25x. There is no fundamental reason DXCM should not converge toward 25-30x earnings, which on $1.45-$1.65 EPS is a $40-50 stock. Reverse-DCF implied growth of 9.87% sounds undemanding, but if growth lands at 7-8% with margin compression, the present value collapses fast.

If I am right, the stock could be worth $42 within 2 years.

Lollapalooza Bias Check

Authority bias. Dexcom is a celebrated growth name with a 25-year founder/operator narrative; the analyst community has lionized Kevin Sayer for over a decade. I am leaning on the narrative that 'CGM is going to be in every diabetic's hand' because authoritative voices (JPM Healthcare Conference, ADA scientific sessions) have repeated it. I am partially controlling for this by noting that Abbott — a far larger and more diversified company — is the price leader, which complicates the standard Dexcom-as-category-king narrative.

Anchoring. The stock traded at $160 in 2021. Anchoring to that prior peak makes today's $61 feel cheap. But the 2021 peak was a 90x earnings COVID-tailwind valuation that was never the right baseline. I am explicitly using the IV bracket ($57-$92) rather than 'percentage off the high' as the anchor, but the temptation to view this as a 'fallen angel' bargain is real and present.

Recency bias. The 2024-2025 sales-force miss and channel-mix surprise are vivid; I may be over-weighting them and assuming the trajectory continues. Equally, I may be over-weighting the recent ROIIC of 30.6% as a signal of franchise strength when in fact it reflects post-COVID operating leverage that has now normalized. Both directions of recency are active.

Confirmation bias toward 'too hard.' Med-tech with regulatory exposure, payer concentration, and rapidly evolving technology (GLP-1, non-invasive sensors, OTC) is where I have a low base rate of insight. I am tempted to call this 'Too Hard' and walk away — which would be intellectually honest but might cause me to miss a 30% margin of safety. The honest position is: I know enough to size small, not enough to bet big.

Social proof. Several respected investors (notable names omitted, but visible in 13F filings) have been adding to DXCM at these prices. That signal should be weighted, but it is also a classic Cialdini trap — the same crowd was buying at $130 18 months ago.

Deprival super-reaction. Having identified what looks like a real margin of safety (px/IV 0.72), I will be tempted to defend that view against subsequent disconfirming data points (next PBM contract season, next quarterly print). I should pre-commit to the bear-case price targets now so I do not move them later.

Incentive bias I do NOT have. No positional bias, no career risk, no client demand. This is one of the cleaner inversions I can do.

10-Year Outlook

Same business in 10 years? Mostly yes — Dexcom will still sell wearable glucose biosensors with associated software and reimbursement. The product form factor will compress (smaller, longer-wear, possibly multi-analyte) but the business model is recognizable. This is not a Munger 'fundamentally different shape in 10 years' company.

Customer base larger? Probably yes. Type-1 grows at biological rates (~2%/year). Type-2 insulin-using grows ~3-4%/year if GLP-1s do not fully arrest progression, less if they do. OTC/wellness adds a long tail of low-ARPU users. Net users: 3-5x today. Profit per user: 0.5-0.7x today. Net revenue: 2-3x. Net profit: less clear; depends on whether scale offsets mix.

Profit per customer higher? No — almost certainly lower. PBM compression, international mix, OTC commoditization all cut the same direction. The only offset is sensor cost reduction from yield curves and longer sensor life.

Moat wider? Probably similar or narrower. Pump integration moat is durable but applies to a static cohort. The OTC channel is structurally moat-less.

Single biggest threat. Two-headed: (1) Apple/Samsung non-invasive optical glucose sensing in the watch — binary, low probability per year but compounding over 10 years. (2) GLP-1 induced therapeutic substitution that shrinks the high-value insulin-using TAM.

Confidence assessment. This is a knowable business with a non-trivial number of moving parts (regulation, reimbursement, GLP-1 interaction, OTC behavior, Apple). I can model the duopoly. I cannot model with confidence the OTC adoption curve, the GLP-1 behavioral end-state, or the Apple sensor probability. Net: medium. Not high enough to size aggressively, not low enough to walk away.

CONFIDENCE: medium

Position Guidance

• Recommendation: Buy (small initial position, scale on weakness)
• Conviction: medium
• Target buy price: $58 (at or below the IV-low of $57.26 = full margin of safety)
• Target trim price: $92 (above IV-high of $91.85 = bull-case fully priced)
• Position sizing: 1.5-2.5% of portfolio at $61. Add another 1-1.5% if it trades below $50. Cap at 4% — narrow moat plus PBM concentration plus med-tech regulatory risk plus 50%-of-FCF SBC say this should never be a top-five position.
• Sell triggers: (a) Two consecutive quarters of share loss to Abbott in commercial channel, (b) PBM step-therapy edict that mandates Libre first-line, (c) Stock crosses $92 with no fundamental upgrade, (d) Apple ships a credible non-invasive glucose sensor.