Medical Devices

The question

Is the medical device platform universe priced as a cyclical capital-equipment basket exposed to hospital capex cycles and GLP-1 procedure substitution fears, or as a collection of razor-and-blade franchises whose installed bases produce 70-80% gross margin consumables on a ten-to-fifteen year utilization tail?

Base rates

The reference class is installed-base medical platforms across three prior cycles: cardiac rhythm management (1990-2010), minimally invasive surgery (2005-present), and continuous glucose monitoring (2016-present). In each case, the capital placement preceded consumable revenue by 12-24 months, consumables compounded at 12-18% annually for a decade after peak placements, and gross margins on the consumable stream ran 15-25 points above the capital. The equity base rate for diversified device baskets during installed-base harvest years has been roughly the 75th percentile of healthcare sub-sectors, with beta to hospital admissions near 0.4 and beta to SPY near 0.8.

The trailing-twelve-month data on this book is unambiguous: ABT is down roughly 21%, BSX down 32%, BAX down 37%, PODD down 21%, and DXCM down 7%. The SPX compounded a double-digit gain over the same window. Healthcare's underperformance is the setup, not a reason to stay away. The imputed expectation embedded in current multiples is that procedure volumes step down 8-12% and margins compress 200-300 basis points on a permanent basis. That is the bear case priced as a base case.

The reference class for a post-drawdown medtech entry is 2016-2017, when the basket de-rated on drug-pricing overhang and proceeded to compound above-market for the next four years once the narrative overhang cleared without the fundamentals deteriorating. Procedure volumes in structural heart, electrophysiology, and robotics are currently running above pre-pandemic trend, not below.

Why consensus is wrong

The sell-side has modeled GLP-1-driven procedure substitution as a linear subtraction from the surgical and orthopedic book. The emerging real-world evidence cuts the other way: GLP-1 patients remain on therapy for cardio-metabolic indications while presenting for elective joint replacement, cardiac ablation, and TAVR at rates that are flat-to-higher than the non-GLP-1 cohort because risk-adjusted surgical candidacy improves as weight comes off. The net effect on SYK's joints franchise, BSX's Watchman and Farapulse book, and EW's TAVR volumes is neutral-to-positive, not negative.

The second miss is the consumables arithmetic. ISRG places a da Vinci system once and sells instruments and accessories for the next ten years at roughly 70% gross margin. BSX placed Farapulse into the electrophysiology labs over 2024-2025; the consumable tail from that placement has not yet shown up in the income statement at scale. SYK's Mako robotic system is on the same curve. The market is extrapolating capital placement growth while ignoring that the consumable base is compounding off a larger installed footprint.

Third, DXCM and PODD are being priced as if GLP-1s reduce CGM and pump utilization. The Type 1 diabetes population is structurally separate from the GLP-1 addressable market, and the Type 2 intensive-insulin cohort that uses both categories is expanding, not contracting. The imputed substitution is not in the claims data.

Position construction

The book has four 20% anchors, a structural heart and CGM sleeve, and a small-cap optionality tail.

Installed-base anchors (~80%). ABT at ~20% is the CGM, structural heart, and diagnostics composite -- Libre is the highest-utilization consumable in the book and the multiple has compressed alongside the sector. ISRG at 20% is the purest razor-and-blade in medtech; instrument and accessory revenue per procedure is still expanding as fourth-generation platforms roll through the installed base. BSX at 20% is the electrophysiology and structural heart play after the worst drawdown in the book; Farapulse unit economics are only beginning to show up on a full-year basis. SYK at 20% is the orthopedic robotics and trauma franchise whose Mako-driven share shift in knees and hips runs for another five-plus years.

Structural heart and CGM extension (~16.2%). EW at ~7.7% is the TAVR pure-play that has held up on a trailing-year basis and whose EARLY TAVR expansion moves the addressable population meaningfully. DXCM at ~4.6% and PODD at ~3.9% are the continuous glucose and patch-pump franchises where GLP-1 fear has compressed multiples out of proportion to Type 1 and intensive-insulin volume trends.

Hospital essentials and optionality (~3.8%). BAX at ~2.1% is the hospital essentials and kidney care franchise at a distressed multiple after the worst trailing-year print in the book. TFX at ~1.0% is the interventional access and urology specialist mid-turnaround. TMDX at ~0.7% is the asymmetric position -- organ preservation and transport, the only position in the book with positive trailing-year performance and the only name whose addressable market is still sub-scale enough to compound above 25%.

Asymmetric payoff

If procedure volumes hold at current run-rate, consumable attach rates on the ISRG, BSX, and SYK installed bases continue to compound, and the GLP-1 substitution narrative resolves neutral-to-positive, the weighted book returns roughly 18-26% annualized over three years as multiples normalize. If hospital capex contracts sharply and procedure growth steps down 5-8%, the book returns roughly -5% to +5%. If Medicare site-of-service reimbursement shifts favor outpatient cardiac and orthopedic venues more aggressively than currently scheduled, the right tail is 30-45% with multiple expansion back to the 2021 band.

At 60% base, 25% bear, and 15% bull, expected value is roughly +13 to +18% annualized against an SPY base rate near +8%. The payoff is asymmetric because the trailing-year drawdown has already absorbed the bear-case multiple compression while the consumable tail on the installed base limits revenue damage even in a procedure down-cycle.

Three things that would change our mind

1. CMS finalizing site-of-service neutrality in a form that cuts device reimbursement at the ASC and HOPD levels simultaneously, removing the cross-subsidy that supports outpatient procedure migration.
2. Real-world claims data over two consecutive quarters showing GLP-1 initiation reduces orthopedic and cardiac procedure rates by more than 10% on a cohort-matched basis, confirming rather than refuting the substitution thesis.
3. An ISRG fourth-generation cycle or BSX Farapulse rollout where consumable revenue per procedure declines on a same-installed-base basis -- the razor-and-blade relationship failing would invalidate the core framework.

What we're explicitly NOT betting on

We are NOT betting on a specific hospital capex recovery timeline; the consumables base compounds whether or not capital placements reaccelerate. We are NOT betting on any single product launch -- not a future da Vinci platform, not a specific BSX pipeline asset, not a Mako refresh. We are NOT betting on GLP-1 procedure substitution being zero; we are betting only that it is smaller than the multiple compression already reflects. We are NOT betting on a particular M&A outcome or a break-up of BAX. We are NOT betting on CGM reaching the non-diabetic wellness population at any specific penetration rate. The thesis requires only that procedure volumes stay near run-rate, consumable attach rates hold, and multiples normalize toward the trailing ten-year mean. Each is a weaker claim than picking the next platform winner, and the book is sized accordingly.