Automated Insulin Delivery (AID) technology and investment research

Technical bottleneck

Why it's a moat — algorithm quality, CGM integration, and 510(k) regulatory pathway

AID is an algorithm business. The pump hardware is commoditizing; the competitive differentiation is in the control algorithm that determines how aggressively to dose insulin based on CGM data:

1. Time-in-Range (TIR) as the clinical metric: The key performance metric is TIR — the percentage of time a patient's glucose stays in the target range (70–180 mg/dL). Control-IQ+ (Tandem) and Omnipod 5 achieve TIR of 70–75% in clinical trials, vs. 50–60% for manual insulin delivery. Each percentage point of TIR improvement reduces HbA1c and long-term complication risk. The algorithms are FDA-cleared through the 510(k) pathway (substantial equivalence to a predicate device) — meaning new entrants can reference existing AID systems, but demonstrating non-inferior TIR requires clinical trials that take 2–3 years and $20–50M.

2. CGM integration lock-in: Omnipod 5 integrates exclusively with Dexcom G6/G7. Tandem integrates with Dexcom G6/G7 and is expanding to Abbott Libre 3 Plus (5 international markets, 2026). Once a patient is on a Dexcom CGM, switching to a Medtronic pump (which uses Guardian CGM) means switching both pump and CGM — a 2-for-1 switching cost that heavily favors incumbents. The CGM duopoly (Dexcom + Abbott) effectively determines the pump ecosystem: pumps that work with Dexcom capture the ~34% of CGM users on Dexcom; pumps that work with Libre capture the ~53%.

3. Form factor preference: Omnipod 5 is tubeless — the insulin reservoir and pump are integrated into a patch worn on the body, controlled by a smartphone. t:slim X2 and Mobi are tubed — the pump connects to the infusion site via thin tubing. Patient preference splits on form factor: some prefer tubeless (discretion, swimming, no snagging); others prefer tubed (can disconnect, larger reservoir, touchscreen on t:slim X2). This creates a natural duopoly — patients self-select into the form factor they prefer, reducing direct head-to-head churn.

4. The fully closed-loop frontier: Current AID systems are "hybrid closed-loop" — they automate basal insulin but still require manual meal boluses. The EVOLUTION/EVOLVE trials (Insulet, data at ATTD 2026) showed fully closed-loop performance with no meal boluses achieving ~68% TIR in Type 2 diabetes — comparable to hybrid closed-loop in some populations. If fully closed-loop becomes standard, the competitive advantage shifts to whoever has the best meal-agnostic algorithm. Insulet targets 510(k) submission in 2027, launch in 2028.

Adoption

Why it matters now

Insulet (PODD):

• Omnipod 5: tubeless patch pump AID + Dexcom G6/G7; Q1 2026 $762M (+34% YoY); FY2026 guide 21–23%
• International expansion: France reimbursement (Dexcom G7 + Omnipod 5, 2026)
• Pipeline: Omnipod 6 (ADA 2026 data, improved TIR, 2027 launch); fully closed-loop (EVOLVE trial, 2028 target)

Tandem Diabetes Care (TNDM):

• t:slim X2 (tubed, touchscreen) + Tandem Mobi (smallest durable AID, app-controlled)
• Control-IQ+ algorithm: automated basal + correction boluses
• CGM expansion: Dexcom G6/G7 + Abbott Libre 3 Plus (launched UK/Sweden, expanding 2026)
• Indications: Type 1 (ages 2+), Type 2 (ages 18+)

Key trends

• Type 2 diabetes expansion: AID historically for Type 1s; now expanding to insulin-dependent Type 2s — a 10× larger TAM
• Fully closed-loop: Eliminating meal boluses is the holy grail — Insulet's EVOLVE trial and Tandem's R&D both targeting this
• CGM multi-platform: Tandem adding Abbott Libre 3 Plus reduces single-CGM dependency; Insulet remains Dexcom-only (for now)
• Competitive pressure: Modular Medical's Pivot (tubeless patch pump, first commercial patients 2026) and Medtronic's MiniMed — incremental threats, not yet share-takers

Key players

Related Technologies

• Continuous Glucose Monitoring (CGM) — CGM is the sensor input for all AID systems; the two are an inseparable pair
• Sterile Fill-Finish — insulin is a biologic requiring sterile fill-finish; AID drives insulin consumption