The Number on Your Arm
In the spring of 2018, a Type 1 diabetic named Nick Jonas — the pop star, the youngest Jonas Brother, diagnosed at age thirteen — stood on a concert stage in Dallas with a small gray disc adhered to the back of his upper arm. The device, roughly the size of a half-dollar coin, was sending a glucose reading to his iPhone every five minutes. Between songs, he glanced at his phone. The crowd, twenty thousand strong, had no idea that a $350 sensor, manufactured by a company most of them had never heard of, was the reason Jonas could perform a two-hour set without risking hypoglycemic collapse. The device was a Dexcom G6 continuous glucose monitor, and it represented the quiet culmination of a twenty-year bet that the future of diabetes management was not a better finger stick but the elimination of finger sticks entirely.
That bet — placed in a San Diego garage in 1999 by a radiation oncologist and a biomedical engineer, funded initially by the kind of capital that backs ideas too early to have a market — would produce a company that, by 2024, generated $3.9 billion in annual revenue, commanded a market capitalization exceeding $30 billion, and had attached its sensors to the bodies of more than two million active users worldwide. Dexcom didn't invent continuous glucose monitoring. It didn't even commercialize the first CGM. What it did, with a persistence that borders on monomania, was make CGM work — accurate enough for insulin dosing, small enough for daily wear, connected enough to transform raw electrochemical signals into actionable data that could be shared with physicians, caregivers, and — crucially — the algorithms inside insulin pumps that were learning to function as artificial pancreases.
The story of Dexcom is, at one level, a medical device success story: a narrow clinical technology that gradually expanded its addressable market through relentless iteration. But it is also a case study in something rarer — a company that achieved dominance not by inventing a new category but by being so obsessively better at executing within an existing one that it became synonymous with the category itself. CGM is Dexcom in the minds of most endocrinologists, most Type 1 patients, most insulin pump manufacturers. The question now — the tension that defines Dexcom's next decade — is whether the company can extend that identity beyond the intensive insulin-using population that built it and into the vastly larger, far less clearly defined market of Type 2 diabetics and, more controversially, the "metabolic health" consumers who don't have diabetes at all.
By the Numbers
Dexcom at a Glance (FY 2024)
$3.9BAnnual revenue
~65%Gross margin
$30B+Market capitalization
2M+Active CGM users globally
10,000+Employees worldwide
~85%U.S. CGM market share (intensive insulin users)
350M+Estimated global diabetic population
A Radiation Oncologist's Side Project
The founder's biography contains the kind of detail that, in retrospect, feels inevitable but at the time was merely eccentric. John Burd was a practicing radiation oncologist in Madison, Wisconsin — a physician who spent his days calibrating ionizing radiation to kill tumors without killing patients, a discipline that demands fanatical precision in sensor technology and dosimetry. In the mid-1990s, Burd became fascinated by the problem of continuous glucose sensing, which had defeated dozens of academic labs and corporate R&D departments for a generation. The core challenge was electrochemical: glucose molecules in subcutaneous interstitial fluid could be detected by an enzyme-coated electrode, but the signal degraded rapidly. Biofouling — the body's immune response coating the sensor in proteins and cells — meant that any implanted sensor lost accuracy within hours or days. Burd, accustomed to thinking about how radiation interacts with tissue at the molecular level, saw the biofouling problem differently. He filed his first CGM-related patent in 1999.
He recruited Scott Glenn, a biomedical engineer, and together they incorporated Dexcom — a portmanteau of "dextrose" and "com" (for communication) — in May 1999 in San Diego, a city already thick with medical device companies and the particular species of venture capital that understood FDA timelines. The initial funding was modest: a $3 million Series A in 2000 from Kingsbury Associates and other local investors. The pitch was simple and, to most medtech VCs, familiar to the point of exhaustion: we will build a sensor that measures glucose continuously, without finger sticks, and transmits the data wirelessly. The unfamiliar part was Burd's specific approach to the sensor chemistry — a proprietary membrane technology that would extend sensor life by managing how glucose and oxygen reached the electrode.
Early Dexcom was a sensor company in the purest sense: a team of fewer than thirty people in a rented lab space, running benchtop experiments on polymer membranes and electrochemical cells, burning through cash at roughly $200,000 per month, with no product, no revenue, and an FDA approval timeline measured in years. The company went public in April 2005 — an IPO that raised $34 million at a $200 million valuation, notable less for its size than for its timing. Dexcom had exactly zero products on the market. The IPO was a bet, by both the company and its underwriters, that the clinical data from ongoing trials would support approval of its first-generation system, the STS (Short-Term Sensor), which had received a CE mark in Europe but was still winding through the FDA's premarket approval process.
The STS received FDA approval in March 2006. It was, by any modern standard, terrible. The sensor required calibration every twelve hours via finger-stick blood glucose readings. It was approved only as an adjunctive device — meaning patients still had to confirm readings with a traditional glucometer before making treatment decisions. The transmitter was bulky. The receiver was a standalone device with no smartphone connectivity. Sensor life was three days. Insertion was painful. And it cost roughly $250 per month out of pocket, because insurance coverage was virtually nonexistent.
It also worked. Not perfectly, not conveniently, but meaningfully. For the first time, a Type 1 diabetic wearing the STS could see their glucose trend — not just a single point-in-time number from a finger prick, but a directional arrow showing whether they were rising, falling, or stable. That trend arrow, a seemingly trivial piece of UI, would turn out to be the most consequential design decision in the history of diabetes technology. It shifted diabetes management from a reactive discipline (test, see a number, respond) to a predictive one (see a trend, anticipate, preempt).
The Decade of the Sensor
What followed the STS launch was not a breakthrough but a grind — a decade of incremental sensor improvements that, compounded, amounted to a revolution. Each generation attacked a specific dimension of the user experience:
📈
The Dexcom Sensor Generations
From painful prototype to invisible infrastructure
2006STS (1st gen): 3-day sensor life, 12-hour calibration, standalone receiver. FDA adjunctive use only.
2007STS-7: Extended sensor life to 7 days. Slightly improved accuracy (MARD ~16%).
2009SEVEN PLUS: Improved membrane chemistry. First system to gain pediatric indication.
2012G4 PLATINUM: Breakthrough in accuracy — MARD of ~13%. Receiver redesigned. First system with smartphone-like display.
2015G5 Mobile: First CGM with direct Bluetooth-to-smartphone connectivity. Eliminated the standalone receiver for many users. FDA-approved for non-adjunctive use (insulin dosing decisions).
2018G6: Factory-calibrated — no finger sticks required. 10-day sensor life. One-touch applicator. Integrated with insulin pumps for automated insulin delivery.
2023
The metric that matters most in CGM is MARD — Mean Absolute Relative Difference — which measures the average percentage deviation of the sensor reading from a laboratory reference. A MARD below 10% is generally considered accurate enough for non-adjunctive insulin dosing. Dexcom's progression from ~16% MARD (STS-7) to ~8.2% (G7) represented not a marketing improvement but a genuine change in the clinical utility of the device. Below 10% MARD, a CGM becomes trustworthy enough that patients — and, critically, the FDA — accept it as a primary glucose measurement, not merely a supplement to finger sticks.
The G5 Mobile, launched in 2015, was the first CGM approved by the FDA for non-adjunctive use, meaning patients could make insulin dosing decisions based solely on the CGM reading. This was the regulatory inflection point that unlocked everything that followed: integration with insulin pumps, coverage by Medicare and most commercial insurers, and the gradual obsolescence of the traditional blood glucose meter. When Kevin Sayer, who had become CEO in 2015 after serving as CFO and then president, described the G5 approval, he called it "the moment we went from being a nice-to-have to a standard of care."
The G6 changed the conversation from 'should I use CGM?' to 'which CGM should I use?' And increasingly, the answer is all of them — all patients with diabetes should be on CGM.
Sayer himself deserves the compressed portrait. A BYU-trained accountant who had spent a decade at Conductus, a superconductor company, before joining Dexcom in 2007 as CFO, Sayer was not a physician-founder or a sensor scientist but a financial operator who understood that Dexcom's real product was not a piece of hardware but a recurring revenue stream attached to a patient's body. His strategic intuition — that the disposable sensor, not the reusable transmitter, was the economic engine of the business — shaped Dexcom's pricing, manufacturing investment, and partnership strategy for the next decade. He pushed relentlessly to reduce the cost of sensor manufacturing, knowing that margin expansion on the consumable was the path to both profitability and insurance coverage.
The Partnership That Built the Standard
No medical device achieves standard-of-care status alone. Dexcom's ascent was inseparable from its relationship with insulin pump manufacturers — a symbiosis that, by the early 2020s, had hardened into something closer to a duopoly controlling the automated insulin delivery (AID) ecosystem.
The critical partnership was with Tandem Diabetes Care, a San Diego neighbor that had been building touchscreen insulin pumps since 2012. In 2018, Tandem received FDA clearance for its t:slim X2 pump with Control-IQ technology — an algorithm that used Dexcom G6 CGM data to automatically adjust basal insulin delivery. This was, functionally, the first widely available hybrid closed-loop system: a machine that could both raise and lower insulin delivery based on predicted glucose trends, with the Dexcom sensor serving as the system's eyes.
The impact on both companies was immediate and compounding. Tandem's pump sales surged because Control-IQ was the most user-friendly AID system on the market. Dexcom's sensor sales surged because every Tandem pump required a Dexcom sensor to function. By 2023, approximately 40% of Dexcom's U.S. sensor volume was attached to an insulin pump — a percentage that represented not just revenue but stickiness. A patient using a Tandem pump with a Dexcom sensor was locked into both ecosystems simultaneously: switching CGMs meant switching pump algorithms, retraining, and accepting clinical risk during the transition. The switching costs were not financial but physiological.
Dexcom also maintained integration partnerships with Insulet (the Omnipod 5 tubeless pump), Beta Bionics (the iLet bionic pancreas), and various international pump manufacturers. The company's strategy was deliberate: be the CGM that works with every pump, making Dexcom the default sensor in the AID ecosystem regardless of which pump a patient chose. This was, in effect, a platform strategy executed in the language of medical devices — own the data layer, let others compete on the hardware.
But the partnership dynamic contained a tension that would eventually surface. Dexcom's dominance as the AID sensor of choice gave it enormous leverage over pump manufacturers, who depended on Dexcom integration for their clinical differentiation. Tandem, Insulet, and others were, in a structural sense, building their businesses on Dexcom's platform. If Dexcom ever decided to build its own pump — or if a competitor like Abbott achieved AID integration with its FreeStyle Libre — the entire partnership ecosystem could destabilize. Sayer repeatedly insisted that Dexcom would remain "sensor-focused," but the strategic temptation was obvious: the company that controlled both the sensor and the algorithm could capture the full margin stack of automated insulin delivery.
The Abbott Problem
For most of Dexcom's history, its competitive landscape was comfortably sparse. Medtronic, the largest medtech company in the world, had its own CGM — the Guardian Sensor — but it was widely regarded as less accurate than Dexcom's and was bundled exclusively with Medtronic's own insulin pumps, limiting its addressable market. The real competitive threat arrived not from a traditional medtech rival but from a diagnostics giant with a fundamentally different go-to-market philosophy.
Abbott Laboratories launched FreeStyle Libre in Europe in 2014 and in the U.S. in 2017. Libre was not, technically, a continuous glucose monitor — it was a "flash" glucose monitor, meaning it stored data continuously but required the user to scan the sensor with a phone or reader to see the current reading. No automatic alerts, no real-time transmitter, no Bluetooth push to a smartphone app. This distinction, which Dexcom emphasized relentlessly in its marketing, was clinically meaningful: a device that required user initiation to display data could not wake a sleeping patient during a dangerous nocturnal hypoglycemic event.
But Libre had something Dexcom didn't: price. At roughly $75 per month in the U.S. (less than half the cost of a Dexcom G6), with a sensor that required no prescription in many international markets, Libre was positioned not as a precision diabetes management tool but as a mass-market glucose awareness product. Abbott's strategy was volume over margin: get the sensor on as many arms as possible, at the lowest possible price point, and build an installed base so large that it would become the default CGM through sheer ubiquity.
By 2023, Abbott's FreeStyle Libre had surpassed $5 billion in annual revenue globally — larger than Dexcom's entire top line — driven primarily by enormous uptake in Europe, where national health systems favored Libre's lower per-unit cost. In the U.S., the picture was more nuanced: Dexcom dominated the intensive insulin-using market (Type 1 and insulin-dependent Type 2), while Abbott carved out the broader Type 2 and "CGM-curious" segments. But Abbott was not standing still. The FreeStyle Libre 3, launched in the U.S. in 2023, added real-time continuous monitoring and smartphone alerts, eliminating the "flash" limitation that had been Dexcom's primary competitive talking point. And Libre 3's MARD was reported at ~7.9% — statistically comparable to, and by some measures slightly better than, the Dexcom G7.
The competitive dynamic with Abbott would define Dexcom's strategic choices for the next decade. Dexcom could defend its premium position through clinical superiority and AID integration, accepting a smaller but more profitable share of the CGM market. Or it could go downmarket, competing on price and accessibility to capture the much larger Type 2 and wellness populations. The company, under Sayer's leadership, chose to do both — a strategic ambition that carried significant execution risk.
The Stelo Gambit
In March 2024, the FDA cleared Dexcom Stelo — the first over-the-counter continuous glucose monitor in U.S. history. The clearance was a regulatory milestone that Dexcom had pursued for years, and its significance extended far beyond the company's own product roadmap. Stelo was cleared for adults eighteen and older who do not use insulin — a population that included not just the estimated 25 million Americans with non-insulin-treated Type 2 diabetes but also the much hazier category of metabolically interested consumers: the biohacker wearing a CGM to optimize their diet, the prediabetic seeking early intervention, the CrossFit devotee who wanted to see how their glucose responded to a sweet potato.
Stelo used the same G7 sensor platform as Dexcom's prescription product but was packaged and marketed differently: no prescription required, available on Dexcom's website and eventually at pharmacies, priced at approximately $99 per month (later reduced to $89) with no insurance billing. The user experience was simplified — no urgent low glucose alerts (because the target population was not at risk of insulin-induced hypoglycemia), a streamlined app focused on food logging and glucose pattern recognition, and a gentler onboarding flow designed for people who had never worn a medical device on their body.
The strategic logic was compelling. Dexcom's core market — approximately 1.5 million intensive insulin users in the U.S. — was approaching saturation. Penetration among Type 1 diabetics with commercial insurance was estimated at 50-60% by 2024. Growth in that population required either converting the remaining holdouts (many of whom were uninsured, underinsured, or simply resistant to wearing a device) or waiting for the incident rate of Type 1 diagnosis to grow, which it does at only 1-2% annually. The Type 2 market, by contrast, was vast and largely untapped: roughly 37 million Americans with diabetes, of whom only a small fraction used any CGM.
But Stelo also represented a profound brand risk. Dexcom had spent two decades positioning itself as the most accurate, most clinically validated, most physician-trusted CGM on the market. Its brand was built on precision and medical necessity. Stelo — an over-the-counter wellness product sold directly to consumers — risked diluting that brand into the same territory as fitness trackers and sleep monitors, consumer health devices with high churn rates and questionable clinical utility. If Stelo's users experienced sensor failures, inaccurate readings, or simply grew bored after three months and stopped subscribing, the resulting consumer dissatisfaction could taint the Dexcom brand in ways that affected its core prescription business.
The early results were mixed. Dexcom reported that Stelo exceeded internal launch targets in Q3 and Q4 2024, but the company declined to break out Stelo-specific revenue or subscriber numbers, suggesting the figures were not yet large enough to carry their own narrative. Wall Street analysts estimated Stelo revenue at $50-80 million in its first partial year — meaningful for a new product but negligible relative to Dexcom's $3.9 billion total revenue.
Stelo is not a distraction from our core business. It is the front door. Every Stelo user who sees their glucose data for the first time and realizes they need to make changes — that person is a future Dexcom customer, whether they stay on Stelo or eventually move to a prescription CGM.
The Revenue Wobble
On July 25, 2024, Dexcom reported Q2 2024 earnings that sent its stock down 40% in a single session — the worst single-day decline in the company's history. Revenue of $1.004 billion grew 15% year-over-year but missed the consensus estimate by $31 million. More troublingly, the company cut its full-year 2024 revenue guidance from $4.20-$4.35 billion to $4.00-$4.05 billion, citing lower-than-expected new patient starts in the U.S. and a slower-than-anticipated ramp in its direct-to-consumer channel.
The stock, which had traded above $130 before the earnings call, cratered to $75. Overnight, $18 billion in market capitalization evaporated.
What happened? The diagnosis was contested, but the proximate causes were identifiable. First, Dexcom had restructured its U.S. sales force earlier in 2024, shifting resources away from endocrinologist offices (where Dexcom was already deeply penetrated) toward primary care physicians and the broader Type 2 market. The transition was disruptive: the new sales reps were less experienced, the primary care physicians were less familiar with CGM reimbursement, and the sales cycle was longer. The result was a temporary hole in new patient acquisitions — a classic execution stumble during a go-to-market reorganization.
Second, the rebate environment shifted. Pharmacy benefit managers and Medicare Advantage plans pushed for larger rebates on Dexcom sensors, compressing net revenue per user even as gross volume grew. Dexcom had historically enjoyed pricing power commensurate with its clinical superiority, but the arrival of FreeStyle Libre 3 — now functionally comparable in accuracy and approved for real-time monitoring — gave payers a credible alternative to use as negotiating leverage.
Third, and most speculatively, some analysts suggested that the "GLP-1 effect" was beginning to dampen CGM demand. Drugs like Ozempic and Mounjaro, prescribed to millions of Type 2 diabetics, were achieving glucose control so effectively that some patients and physicians questioned whether CGM was necessary. If a patient's A1C dropped from 8.5% to 6.2% on semaglutide, the marginal value of a $300/month CGM became harder to justify. Dexcom pushed back aggressively on this narrative, arguing that CGM and GLP-1s were complementary — the CGM provided the data to optimize drug titration and dietary behavior — but the market's anxiety was palpable.
The Q2 2024 stumble was, in microcosm, every growth-stage medtech company's nightmare: the moment when the story shifts from "massive TAM expansion" to "show me the execution." Dexcom's bull case depended on the company's ability to simultaneously defend its premium position in the intensive insulin market, expand into the much larger Type 2 population, and launch an entirely new consumer product — all while a well-capitalized competitor was attacking from below on price and closing the accuracy gap from above.
The Manufacturing Bet
Behind the sensor chemistry and the software algorithms, Dexcom's competitive position rested on something less glamorous but equally critical: manufacturing scale. A CGM sensor is a sophisticated electrochemical device — enzyme coatings, membrane layers, flexible circuits, sterile packaging — that must be produced in quantities of hundreds of millions per year at a consistency level measured in parts per million. Each sensor is a single-use disposable that generates revenue only when it is produced, shipped, and applied to a patient's body. The economics of CGM are, at their core, the economics of high-precision consumable manufacturing.
Dexcom operated three primary manufacturing facilities by 2024: its headquarters campus in San Diego, a facility in Mesa, Arizona (opened in 2022), and an expanding operation in Malaysia. The Malaysia facility, which began production in 2023, was Dexcom's most significant manufacturing investment — a bet that the company could achieve the cost reductions necessary to compete in the Type 2 and OTC markets only by moving a substantial portion of sensor production to a lower-cost geography.
The Mesa facility represented a different strategic logic: domestic manufacturing capacity for the U.S. market, insulated from trade policy risk, and co-located with Dexcom's growing software and data science teams in the southwestern U.S. Between Mesa and Malaysia, Dexcom invested over $500 million in manufacturing capital expenditure between 2021 and 2024, a figure that reflected both the scale of the company's demand growth and its conviction that sensor unit cost reduction was the single most important enabler of its Type 2 expansion strategy.
Sayer repeatedly framed the manufacturing buildout in terms of gross margin trajectory. Dexcom's gross margins, which had hovered around 63-65% in 2022-2023, were compressed by the startup costs of the new facilities. The company guided toward 65%+ gross margins by 2025 as Malaysia ramped and sensor yields improved, with a longer-term target of 70%+ as the fixed costs of the new plants were absorbed across a much larger volume base. The math was straightforward: at $3 per sensor manufacturing cost (estimated) and $10-12 per sensor average selling price (blended across channels), every percentage point of yield improvement or cost reduction flowed almost directly to gross margin.
Software Eats the Sensor
If the first two decades of Dexcom's history were defined by sensor chemistry — making the electrochemical measurement accurate enough to trust — the emerging chapter was defined by software. The glucose number itself was becoming a commodity. What differentiated Dexcom was increasingly not the raw measurement but what the company's software did with it.
The Dexcom Clarity platform, launched initially as a data management tool for physicians, evolved into something more ambitious: a cloud-based analytics engine that aggregated glucose data from millions of sensors, generated population-level insights, and — through its integration with electronic health records — enabled physicians to manage diabetic patients remotely. By 2024, Clarity was processing over 30 billion glucose data points annually, a dataset of extraordinary depth that no competitor could replicate because no competitor had Dexcom's installed base of high-frequency, real-time CGM users.
The software strategy had three layers. First, Clarity itself: the physician-facing analytics platform that improved clinical outcomes (demonstrated in peer-reviewed studies showing that patients whose physicians used Clarity had better A1C control). Second, the consumer app: the Dexcom G7 and Stelo apps, which translated raw glucose data into behavioral nudges — "your glucose spiked after that meal," "you were in range 78% of the time this week," "your nighttime glucose has been trending higher." Third, and most speculatively, the data asset: the anonymized, aggregated glucose dataset that could be used for drug development partnerships, population health research, and the training of AI/ML models for predictive glucose management.
The sensor is the razor. The software is the blade. But the data — the data is the shaving cream, the aftershave, and the subscription box. We are building the most comprehensive picture of human glucose metabolism that has ever existed.
The AI angle was not merely aspirational. In 2023, Dexcom demonstrated a prototype glucose prediction algorithm that could forecast a user's glucose level 60 minutes into the future with clinical accuracy — a capability that, if integrated into AID systems, would enable insulin pumps to act preemptively rather than reactively. The company also partnered with Type Zero Technologies (acquired by Dexcom in 2018) to develop next-generation algorithms for fully closed-loop insulin delivery systems, where the pump would make all insulin dosing decisions autonomously with no user input.
The strategic implication was clear: as the hardware sensor became commoditized and competitors closed the accuracy gap, Dexcom's durable competitive advantage would shift from electrochemistry to data science. The company that had the most sensors on the most bodies, generating the most glucose data, feeding the best algorithms, integrated with the most insulin pumps, would own the platform layer of automated diabetes management. Everything else was a commodity input.
The International Puzzle
By 2024, international revenue accounted for approximately 30% of Dexcom's total — a percentage that reflected both the company's growing global presence and the uncomfortable reality that its largest non-U.S. markets were structurally different from its domestic stronghold in ways that complicated the playbook.
In the United States, Dexcom benefited from a fragmented insurance market where clinical superiority could command premium pricing and where direct-to-consumer marketing influenced both patient demand and physician prescribing. Endocrinologists and certified diabetes educators served as high-influence channels that Dexcom had cultivated for years. Medicare coverage, secured in 2017, had been transformative for the over-65 population.
In Europe, the dynamics were inverted. National health systems — NHS in the UK, AOK and TK in Germany, Assurance Maladie in France — made purchasing decisions based on health technology assessments that weighed clinical benefit against cost-effectiveness. In these systems, Abbott's FreeStyle Libre had a structural advantage: its lower per-unit cost made it easier to justify at the population level, even if individual clinical superiority was acknowledged. Libre's market share in several European countries exceeded 70%, a dominance that Dexcom struggled to contest.
Germany was the exception and the template. Dexcom had achieved meaningful market share in Germany through a strategy of partnering with the German diabetes registry (DPV), generating real-world evidence demonstrating that Dexcom's accuracy translated into better clinical outcomes specifically in the German healthcare context, and working with physician key opinion leaders to establish Dexcom as the CGM of choice for complex patients. This evidence-based market access approach — expensive, slow, but durable — was being replicated in the UK, France, and Japan, with varying degrees of progress.
The Japanese market, entered in 2022 through a distribution partnership, represented Dexcom's most significant bet in Asia. Japan had the second-highest per-capita spending on diabetes care globally and an aging population with rising Type 2 prevalence. But the Japanese regulatory environment was slow, the reimbursement process opaque, and Abbott had arrived first.
The Culture of the Sensor
Dexcom's San Diego headquarters — a sprawling campus in the Sorrento Valley biotech corridor — operated with the peculiar intensity of a company that knew it was both a category leader and perpetually one sensor generation away from competitive parity. The engineering culture was defined by two seemingly contradictory imperatives: fanatical reliability (because a malfunctioning CGM could lead to a diabetic emergency) and aggressive iteration speed (because Abbott was shipping new products on 18-month cycles).
The quality control regime was aerospace-grade. Each sensor underwent automated optical inspection, electrochemical testing, and sterile packaging verification. Dexcom's manufacturing defect rate, while not publicly disclosed, was described by former employees as "sub-ppm" — sub-parts-per-million — a standard that would be unremarkable for semiconductor fabrication but was extraordinary for a wet-chemistry biomedical device. When a sensor lot failed testing, it was destroyed, not reworked. The cost of that discipline was real: Dexcom's manufacturing yields, while improving, remained lower than Abbott's simpler sensor architecture, contributing to the cost structure differential between the two companies.
Sayer cultivated a leadership team that blended sensor scientists with consumer technology executives — a deliberate hybrid that reflected the company's strategic ambiguity about whether it was a medical device company or a health technology platform. Jake Leach, the CTO, had been at Dexcom since 2005 and embodied the sensor side: a biomedical engineer who had personally designed membrane chemistry for three sensor generations. Teri Lawver, the chief commercial officer hired from Livongo in 2020, represented the consumer technology side: a digital health executive who thought about user acquisition funnels, retention metrics, and net promoter scores. The tension between these two orientations — clinical precision versus consumer delight — was productive when managed and corrosive when not.
In 2024, Dexcom employed over 10,000 people — a number that had doubled in three years and that carried the growing pains typical of any company scaling through its second or third organizational phase. The sales force restructuring that contributed to the Q2 2024 revenue miss was, in part, a symptom of organizational complexity: too many new hires, insufficient institutional knowledge, and a go-to-market strategy that was evolving faster than the field organization could absorb.
The GLP-1 Question
No discussion of Dexcom's strategic position in 2024-2025 was complete without addressing the elephant in the endocrinology waiting room: semaglutide. Novo Nordisk's Ozempic and Wegovy, along with Eli Lilly's Mounjaro and Zepbound, had by 2024 generated combined annual sales exceeding $50 billion and were being prescribed to millions of patients with Type 2 diabetes and obesity. The drugs achieved glucose control so effectively — typical A1C reductions of 1.5-2.0 percentage points — that a genuine question emerged: if GLP-1 receptor agonists could normalize glucose levels pharmacologically, did patients still need to monitor glucose continuously?
Dexcom's answer was emphatically yes, and the company marshaled data to support it. Studies presented at the ADA Scientific Sessions showed that patients on GLP-1s who also used CGM achieved tighter glucose control — higher time-in-range, fewer hypoglycemic events, lower glycemic variability — than those on GLP-1s alone. The argument was intuitive: even if a drug lowers average glucose, the minute-to-minute variability of glucose in response to meals, exercise, stress, and sleep still matters for long-term outcomes. CGM provided the feedback loop that allowed patients to understand how their body responded to specific behaviors, enabling dietary and exercise modifications that worked synergistically with the drug.
But the bull case for CGM-GLP-1 synergy ran headlong into a payer reality: insurance companies, already grappling with $1,000+/month GLP-1 costs, were reluctant to add $300/month CGM on top. The marginal reimbursement case for CGM in a patient whose A1C was already at 6.5% on semaglutide was genuinely weak, at least by the evidence standards that payers applied. Dexcom needed to generate randomized controlled trial data demonstrating that CGM plus GLP-1 produced outcomes superior to GLP-1 alone, and that the incremental benefit justified the incremental cost. That data was being generated — Dexcom had multiple studies underway — but was not yet available at the scale needed to shift payer behavior.
The GLP-1 dynamic also created a strategic opportunity that was easy to underestimate. The millions of patients newly prescribed GLP-1s were, by definition, metabolically engaged — they had sought treatment, they were motivated to improve their health, and they were already spending significant sums on pharmacotherapy. This was exactly the population most likely to adopt CGM for behavioral optimization. Stelo, priced at $89/month with no prescription required, was explicitly designed for this demographic: the motivated, metabolically aware consumer who wanted data-driven feedback on their diet and lifestyle.
Whether the GLP-1 wave would lift Dexcom or swamp it remained, in early 2025, an open question — the kind of question that made the stock a battleground between conviction and doubt.
The Arm of the Future
In January 2025, Dexcom previewed its next-generation sensor platform at the JPMorgan Healthcare Conference — a sensor that was, according to the company, roughly 50% smaller than the G7, with a 15-day wear period and the ability to measure not just glucose but additional analytes including ketones and lactate. The multi-analyte sensor, if successfully commercialized, would represent a category expansion from glucose monitoring to metabolic monitoring — a shift in the device's identity from a diabetes tool to a general-purpose metabolic sensor.
The implications were significant. A sensor that measured ketones could detect diabetic ketoacidosis (DKA) before it became clinically dangerous — a potentially life-saving application for Type 1 diabetics, who are at chronic risk of DKA. A sensor that measured lactate could provide real-time exercise intensity feedback for athletes, opening a consumer wellness market that dwarfed the diabetes population. And a sensor that measured multiple analytes simultaneously could serve as the input layer for next-generation AI health models that correlated glucose, ketones, lactate, heart rate, and activity data to generate genuinely personalized health recommendations.
This was the long-term vision that animated Dexcom's R&D investment — roughly $550 million in 2024, or approximately 14% of revenue. The company was betting that the subcutaneous sensor it had spent twenty-five years perfecting was not the endpoint but the platform: a minimally invasive, body-worn biochemistry lab that could measure an expanding array of analytes and transmit the data to algorithms that would, over time, manage not just diabetes but metabolic health broadly.
Whether the market for general-purpose metabolic sensing would materialize — whether the motivated wellness consumer would pay $89/month indefinitely for data about their ketones and lactate — was unknowable. What was knowable was that Dexcom, having mastered the hardest version of the problem (clinical-grade glucose sensing for insulin-dependent diabetics), was better positioned than any company on earth to extend into the easier versions.
The disc on Nick Jonas's arm was, in this reading, not an endpoint but a prototype — the earliest iteration of a body-worn sensor platform that would, in the company's most ambitious projections, become as ubiquitous as the smartphone it reported to. A small gray circle, barely visible under a rolled-up sleeve, measuring what the blood already knows.
