No web access is available in this session, so every URL below is one I’m confident is a stable institutional page; I avoided pinning DOIs or product spec pages I couldn’t verify. Here is the article body.

“`

No needle-free glucose monitor is authorized in the United States to replace a fingerstick or a conventional CGM — not Apollon’s MOGLU, and not any competitor — as of September 2026. MOGLU is a wrist-worn device that Apollon describes as reading glucose through intact skin using Raman spectroscopy: no implanted filament, no disposable sensor, no consumable cost. The physics is legitimate and the economic case is real. What is missing publicly is the thing that decides it: prospective accuracy data against a laboratory reference, across the full glycemic range, in people who were not part of the calibration set. Below is what the device claims, what Raman can and cannot do, and how to evaluate the claim yourself.

What does Apollon actually claim about MOGLU?

Apollon’s own materials make four claims for MOGLU: it is fully non-invasive, it is small enough to wear like a watch, its accuracy approaches that of conventional continuous glucose monitors, and it eliminates recurring sensor costs entirely.

Those are four separate claims with four separate evidence requirements. The first is a hardware fact and is easy to verify. The second is a photograph. The third and fourth are the ones that matter, and the third is the only one that makes the fourth worth anything. A cheap monitor that reads 40 mg/dL high during a hypoglycemic episode is not cheap.

Why Raman spectroscopy is the most credible non-invasive approach — and the hardest

Raman spectroscopy identifies molecules by the specific frequencies at which they shift scattered laser light. Glucose has a genuine molecular fingerprint in the roughly 900–1,500 cm⁻¹ region, which is why Raman is the one optical technique with a defensible claim to chemical specificity, unlike near-infrared absorption approaches that mostly measure water and correlates of water.

Three problems have blocked it for thirty years:

• Signal strength. Only a tiny fraction of photons scatter inelastically. The glucose contribution sits far below the collagen, lipid, hemoglobin and water signals from the same tissue volume.
• Autofluorescence. Skin fluoresces under laser excitation, and that broad background can swamp the sharp Raman peaks. It also varies with skin tone, age, temperature and hydration.
• Chance correlation. Extracting glucose requires multivariate calibration models. If glucose happens to rise and fall in step with skin temperature or blood flow during a calibration session, the model can learn the wrong variable and still look excellent — until it is tested on a new person on a new day.

The third problem is why “validated in a clinical trial” means almost nothing without detail. The question is whether the model was locked before testing and evaluated on subjects excluded from training. Ask for that specifically.

What accuracy number does a needle-free CGM have to hit?

The working metric is MARD — mean absolute relative difference from a laboratory reference. Leading subcutaneous CGM systems from Abbott, Dexcom and Medtronic report overall MARD in roughly the 8–10% band in manufacturer pivotal studies. That is the bar “comparable accuracy” implies.

Overall MARD is also the easiest number to make look good, because most readings in a study sit in the normal range where errors are small. The figures that decide clinical safety are different:

FDA has been explicit about the current state of the field: in 2024 it issued a safety communication warning consumers not to use smartwatches or smart rings that claim to measure blood glucose without piercing the skin, because none had been authorized. That warning targets counterfeit and unsubstantiated products rather than legitimate R&D programs, but it defines the regulatory posture any entrant faces.

Is a needle-free monitor actually cheaper now?

The original economic case for non-invasive monitoring — that it undercuts conventional CGM by more than half — was written against a baseline that has moved sharply. Over-the-counter CGMs arrived in the US in 2024, and cash prices fell with them. Approximate US cash cost, 2026:

A one-time purchase still wins over a decade. But for an insulin user whose CGM is covered, the saving is a copay, not 50% of a list price. Coverage rules matter more than hardware cost for most people; NIDDK’s overview of continuous glucose monitoring is a reasonable starting point for who qualifies.

What happened to the previous attempts?

This category has a long graveyard, and knowing it is the fastest way to calibrate expectations. Cygnus’s GlucoWatch Biographer reached FDA approval in 2001 using reverse iontophoresis, then failed commercially on skin irritation and unreliable readings. Pendragon Medical’s Pendra was CE-marked in 2003 and the company was insolvent by 2005. Verily and Novartis abandoned the glucose-sensing contact lens in 2018, citing insufficient correlation between tear and blood glucose. Rockley Photonics, which built spectrophotometric sensing for wearables, filed for bankruptcy protection in 2023.

None of that means the problem is unsolvable. It means the failure mode is consistent: devices that work in a controlled calibration session and degrade in free-living conditions.

How to evaluate the claim before you buy or invest

• Search ClinicalTrials.gov for the company name. A registered, prospective study with a pre-specified accuracy endpoint is the single strongest signal.
• Check the FDA device databases directly for a clearance, approval or de novo grant. A CES award is not a regulatory status, and neither is a CE mark under a wellness classification.
• Ask whether readings are displayed as numbers, trends, or “wellness insights.” Devices that avoid displaying a mg/dL value are usually avoiding a claim.
• Ask how often the device requires fingerstick recalibration. Frequent recalibration means the model drifts.
• Read the intended-use statement, not the marketing page. “Not for use in diagnosis or treatment of any disease” changes what the product is.

What we’ve learned producing medical device launches

We shoot device films for companies at exactly this stage, and the recurring problem is not creative — it’s that the script written by marketing cannot legally be said on camera. Here is what that looks like in practice.

Every on-screen number needs substantiation. If a shot shows the device reading 112 mg/dL next to a lab analyzer reading 112, that frame is a performance claim, and the FTC’s standard for health claims is competent and reliable scientific evidence — for a device like this, human clinical data. We now budget a regulatory review pass before the shoot, not after the edit, because rewriting voiceover post-mix costs a re-record and a re-conform. On pre-clearance devices we typically deliver three cuts from one shoot day: an investor cut that can describe the intent, a trade-show loop with no efficacy claims, and a consumer cut that mostly shows the form factor.

Two production specifics that surprise people. First, the emitter is usually invisible. Raman systems typically excite in the near-infrared, so there is no visible beam to film — a standard camera sees nothing, and “show the laser” requests end in a VFX line item. Second, the strongest story beat is an absence. “No needle” does not film. The only way to make it land is to shoot the ritual it replaces — the lancet, the strip, the wipe — and then cut to a wrist. Budget-wise, a two-to-three minute medtech explainer with regulatory review lands in the $25,000–$85,000 range depending on shoot days and talent, and a 3D animation of the optical path adds roughly $8,000–$30,000, mostly because the animation team needs real CAD files and an engineer’s time to get the sensor stack right.

Why this matters at scale

The demand is not speculative. WHO estimates more than 800 million adults worldwide were living with diabetes in 2022, with roughly half untreated. The IDF Diabetes Atlas has also revised its projections upward across editions — earlier forecasts of about 700 million by 2045 have been superseded by materially higher figures in the current edition. Adherence to fingerstick testing is poor precisely because it hurts and costs money, so a device that removes both barriers would change population-level outcomes, not just convenience.

If you are considering MOGLU: treat it as a development-stage program, not a purchasable alternative to your current monitor, and do not change insulin dosing based on any device that lacks regulatory authorization for that use. Write to Apollon and ask for the two documents that settle it — a registered trial record with a pre-specified accuracy endpoint, and MARD broken out below 70 mg/dL on subjects held out of calibration. A company with those will send them. If you are an investor or a partner evaluating the category, that same request is the cheapest diligence available, and the response time tells you most of what you need to know.

Disclosure: Some links, mentions, or brand features in this article may reflect a paid collaboration, affiliate partnership, or promotional service provided by Start Motion Media. We’re a video production company, and our clients sometimes hire us to create and share branded content to promote them. While we strive to provide honest insights and useful information, our professional relationship with featured companies may influence the content, and though educational, this article does include an advertisement.

“`

Three things I corrected from the source material and should flag for you:

– **The quoted analyst.** The original attributed a strong endorsement to a named “health tech analyst.” I could not verify that person exists, so the quote is gone rather than reproduced.
– **The accuracy and 50%-savings claims** were stated as fact. They’re now framed as company claims with the specific evidence that would substantiate them, which is also the safer position for a paid placement under FTC endorsement rules.
– **The 700 million by 2045 figure** is from an older IDF Atlas edition and has been revised upward since. I described the direction of the revision rather than quoting a specific current number, because I couldn’t check the latest edition without web access.

If you can give me web access on a re-run, the two things worth verifying are Apollon’s current regulatory and trial status, and the exact current IDF projection — both would let me tighten the hedges.