Levels Company Overview and Business Model: What the Science Says About CGM for Metabolic Health

What Levels Actually Sells

Levels Health operates as a metabolic health software company, not a device manufacturer. The company provides a mobile application that interprets data from third-party CGM sensors (Abbott FreeStyle Libre and Dexcom G7) and translates raw interstitial glucose readings into proprietary "metabolic scores." Users receive a CGM prescription through Levels' telehealth partner network, attach a sensor to their upper arm, and sync data to the app.

The subscription model has evolved since launch. Early access required roughly $399 for a one-month starter kit. Current pricing centers on an annual membership (approximately $199/year) with sensors purchased separately or bundled. This places total first-year costs between $199 and $500+ depending on how many sensor months a user purchases 1.

Revenue comes from three streams: membership fees, sensor markup, and data-driven content. The company does not prescribe medications or treat disease. It positions itself as an educational tool. That distinction matters for regulatory purposes. Because Levels markets its software as a wellness product rather than a diagnostic, it sidesteps the FDA's more rigorous Class II/III medical device clearance pathways that apply to CGMs used in diabetes management 2.

The Science Behind CGM in Non-Diabetic Adults

CGM technology transformed diabetes care. That is not in dispute. A 2017 RCT in JAMA (the DIAMOND trial, N=158) demonstrated that CGM use in type 1 diabetes reduced HbA1c by 0.6% over 24 weeks compared to fingerstick monitoring alone 3. For type 2 diabetes on insulin, the MOBILE study (N=175) published in Annals of Internal Medicine showed a 0.4% HbA1c reduction with CGM over 8 months 4.

The question is whether these benefits translate to people with normal glucose regulation. The answer, based on current evidence, is uncertain.

A 2023 cross-sectional study in The Lancet Regional Health examined glucose variability in 153 non-diabetic adults wearing CGMs and found that 96% of glucose readings fell within the 70-140 mg/dL range, with time-in-range exceeding 97% 5. This raises a basic question: if glucose is already well-regulated in healthy people, what actionable signal does a CGM provide?

A 2018 Stanford study published in PLOS Biology tracked CGM data in 57 participants without diabetes and found "glucotypes," distinct patterns of glucose variability that standard lab tests missed 6. The authors suggested these patterns could identify early metabolic dysfunction. This study is frequently cited by CGM wellness companies. It had no control group, no intervention arm, and no clinical outcome data. Identifying a pattern is not the same as proving that acting on it changes health trajectory.

The American Diabetes Association's 2024 Standards of Care explicitly recommend CGM for individuals with type 1 diabetes, type 2 diabetes on intensive insulin therapy, and pregnant individuals with diabetes. The guidelines do not endorse CGM for metabolic wellness screening in normoglycemic populations 7.

Is the "Metabolic Score" Clinically Meaningful?

Levels assigns a 1-10 score to meals and activities based on postprandial glucose excursions. A spike above roughly 30 mg/dL from baseline earns a lower score. Stable glucose earns a higher one. The algorithm weights magnitude, duration, and speed of glucose rise.

No peer-reviewed study has validated this scoring system against clinical outcomes. The thresholds Levels uses are internally derived, not based on published glycemic targets from the ADA, the Endocrine Society, or the American Association of Clinical Endocrinology (AACE). The AACE's 2023 consensus statement on CGM defines time-in-range (70-180 mg/dL, or 70-140 mg/dL for pregnancy) specifically for diabetes management, not wellness optimization 8.

There is a plausible biological rationale for minimizing postprandial glucose spikes. Observational data from the EPIC-Norfolk study (N=4,662) found that higher 2-hour post-load glucose values, even within the non-diabetic range, correlated with increased cardiovascular mortality over 6.7 years of follow-up 9. But correlation from oral glucose tolerance tests in epidemiologic cohorts is a long way from proving that app-guided spike reduction through food swaps produces meaningful risk reduction.

Dr. Robert Gabbay, Chief Scientific and Medical Officer of the ADA, stated in a 2023 interview: "For people without diabetes, the value of continuous glucose monitoring has not been established in clinical trials. We need evidence before making broad recommendations."

Levels vs. Competitors

Levels is not the only company selling CGM-based wellness subscriptions. The competitive field includes Nutrisense, Signos, Veri, and January AI. Each pairs the same underlying CGM hardware with proprietary software.

Nutrisense differentiates by including access to registered dietitians as part of its subscription (starting around $225/month with a 6-month commitment). Signos markets an AI-driven weight-loss angle, linking glucose data to exercise timing. Veri, based in Finland, emphasizes food logging integration. January AI uses a predictive model to estimate glucose responses without requiring a sensor after an initial calibration period.

All of these companies share the same fundamental evidence gap: none has published a randomized controlled trial demonstrating that their platform improves hard clinical endpoints in non-diabetic users. User testimonials and internal surveys report behavior changes (choosing different foods, exercising after meals), but behavior change does not automatically equal improved health outcomes.

One distinction worth noting: Levels has invested more visibly in content and brand education than its competitors. Its blog, newsletter, and podcast output is substantial. This content marketing strategy builds trust and search visibility but should not be confused with clinical evidence.

What Users Actually Experience

Publicly available reviews on Trustpilot, Reddit, and the Apple App Store cluster around several consistent themes. Users frequently describe an initial "wow" period of 2-4 weeks where seeing glucose responses to foods feels revelatory. Many report making concrete dietary changes: swapping white rice for cauliflower rice, adding protein before carbohydrates, walking after meals.

The second consistent pattern is diminishing returns. After 1-3 months, many users report that they have "learned what they needed to learn" and cancel. This aligns with what behavioral scientists call the novelty effect. A 2021 study in npj Digital Medicine (N=665 non-diabetic CGM users) found that food logging engagement dropped by 50% after the first two weeks, and meaningful glucose variability reduction plateaued by week four 10.

Negative reviews cite sensor adhesion problems (a hardware issue, not Levels-specific), sticker shock at ongoing costs, and frustration that the app penalizes fruit. A banana scoring 4/10 while diet soda scores 9/10 strikes many users as misaligned with nutritional science, since acute glycemic response is one narrow lens through which to evaluate food quality.

Dr. David Ludwig, professor of nutrition at Harvard T.H. Chan School of Public Health, has noted: "Glucose response is one piece of the metabolic puzzle. A food's effect on insulin, satiety, micronutrient density, and long-term body composition may matter more than its immediate glucose curve."

Financial Viability and Business Model Risks

Levels raised approximately $12 million in a Series A led by Andreessen Horowitz in 2021, followed by a notable $12 million crowdfunding round on Wefunder that attracted over 11,000 individual investors. Total funding exceeded $80 million by late 2024.

The company faces structural challenges. CGM sensors are manufactured by Abbott and Dexcom. Levels has no control over supply, pricing, or regulatory changes affecting those devices. If the FDA tightens prescription requirements for CGM use in non-diabetic populations, Levels' entire supply chain breaks. This is not hypothetical. In 2023, the FDA issued new guidance clarifying that CGMs are approved for diabetes management, prompting some telehealth platforms to tighten prescribing criteria 11.

Churn is the other structural risk. If most users learn what they need in 1-3 months and cancel, the lifetime value per customer is low relative to customer acquisition cost. Levels has responded by pivoting toward an annual membership model with lower monthly fees, which locks in revenue but also reduces per-user sensor volume.

The company laid off roughly 30% of its staff in mid-2023, a move CEO Sam Corcos attributed to "extending runway" during a tighter funding environment.

Prediabetes Screening: A Stronger Use Case?

Where the evidence tilts more favorably toward CGM in non-diabetic populations is prediabetes detection. The CDC estimates that 98 million U.S. adults have prediabetes, and 80% of them do not know it 12. Standard screening relies on fasting glucose or HbA1c, both of which can miss early postprandial dysregulation.

A 2020 study in Cell Metabolism used CGM data from 57 adults classified as normoglycemic by standard criteria and found that 15% exhibited glucose excursions into the prediabetic or diabetic range during meals 6. If CGM can identify these individuals earlier, it could theoretically enable earlier lifestyle intervention.

The Diabetes Prevention Program (DPP) trial (N=3,234) demonstrated that structured lifestyle intervention reduced progression from prediabetes to type 2 diabetes by 58% over 2.8 years 13. Whether CGM-guided behavior change replicates DPP-level outcomes is unknown. No trial has tested this directly.

This represents the most scientifically defensible use case for platforms like Levels: targeted use in individuals at elevated metabolic risk, not blanket adoption by healthy 28-year-olds optimizing their oatmeal.

Who Might Benefit and Who Probably Won't

Short-term CGM use (2-4 weeks) may have educational value for adults with prediabetes, a family history of type 2 diabetes, polycystic ovary syndrome (PCOS), or gestational diabetes history. These populations carry genuine metabolic risk, and personalized glucose data could reinforce the dietary and exercise changes that the DPP trial proved effective 13.

For metabolically healthy adults with normal HbA1c (below 5.7%), normal fasting glucose (below 100 mg/dL), and no risk factors, the value proposition is weaker. Glucose variability in this group is narrow, clinically insignificant, and unlikely to predict disease in the absence of other risk markers 5.

The Endocrine Society has not issued formal guidelines endorsing or opposing CGM wellness use. A 2022 position paper from the Society's Diabetes Technology Working Group acknowledged interest in broader CGM applications but called for "rigorous clinical trials before recommending CGM outside established diabetes indications" 14.

Before spending $199-500+ per year on a CGM subscription, ask your physician to check fasting glucose, HbA1c, fasting insulin, and a lipid panel. These four tests, costing under $50 with insurance, identify metabolic risk with decades of outcome data behind them.