Diabetes on Autopilot: The Arrival of Fully Closed Loop

The race to create a fully closed-loop insulin pump, often called the "artificial pancreas", is entering its final sprint. For millions of people living with insulin-dependent diabetes, the dream of a device that requires zero manual input is closer than ever.

 

Here is a look at where the technology stands today, what is holding us back, and how close we are to ultimate freedom from diabetes management. This is also discussed in greater depth in the IDTechEx report: Diabetes Management Technologies 2027-2037: Markets, Players, and Forecasts.

 

 

For years, people managing type 1 and insulin-dependent type 2 diabetes relied on hybrid closed-loop (HCL) systems. These devices, while revolutionary, still require the user to "announce" meals, estimate carbohydrate grams, and manually press buttons for mealtime insulin boluses.

 

Advanced hybrid closed loop (AHCL) systems can take over the task of delivering extra correction doses when blood sugar spikes. AI and machine learning (ML) have completely transformed the creation of the predictive algorithms that enable Advanced Hybrid Closed-Loop (AHCL) systems to work by replacing rigid mathematical rules with dynamic forecasting. Traditional algorithms like Model Predictive Control (MPC) rely on fixed constraints, but AI has introduced breakthroughs that allow pumps to adapt to the unpredictable nature of human biology.

 

 

AHCL systems have three vital components:

 

Systems available today automatically adjust background (basal) insulin. They can even deliver small correction doses if blood sugar spikes. However, they are still "hybrid."

 

Users must still count carbohydrates. They must manually tell the pump before they eat. They must also announce upcoming exercise to prevent dangerous drops in blood sugar.

 

 

To eliminate manual inputs entirely, technology must overcome three major hurdles:

 

 

Subcutaneous insulin takes time to work. Even the fastest modern insulins take 15 to 30 minutes to enter the bloodstream and up to two hours to peak. Conversely, a biological pancreas secretes insulin directly into the blood, working instantly. Because of this lag, algorithms struggle to catch up to fast-acting carbohydrates without causing later crashes.

 

 

Without user input, a pump must guess when a meal starts based on rising glucose levels. By the time the continuous glucose monitor (CGM) registers a sharp spike, the blood sugar wave is already well underway. The pump is forced to play catch-up, which often leads to higher post-meal spikes than desired.

 

 

 

The human body uses two main hormones to regulate blood sugar: insulin (to lower it) and glucagon (to raise it). Most current pumps only deliver insulin. This means the algorithm must be conservative to avoid causing hypoglycemia (low blood sugar).

 

 

The gap between hybrid and fully closed-loop systems is shrinking rapidly due to three massive innovations:

 

 

 

 

Drug manufacturers are developing ultra-rapid insulins. Some formulations use warming agents or chemical additives to speed up tissue absorption. Accelerating insulin action allows algorithms to aggressively counter meal spikes without human warnings.

 

 

The introduction of bi-hormonal pumps is a gamechanger. These devices house both insulin and a stable form of liquid glucagon. If blood sugar drops quickly, the pump delivers a micro-dose of glucagon to lift it. This safety net allows the insulin algorithm to be far more aggressive, effectively neutralizing meals without user intervention.

 

 

 

Next-generation algorithms do not just react; they predict. By analyzing weeks of user data, AI can recognize personal habits, stress patterns, and hormonal cycles. Advanced meal-detection algorithms can now spot the exact signature of a post-meal glucose spike within minutes, deploying insulin long before the user could even open a bolus calculator.

 

 

We are no longer talking about decades; we are talking about years.

 

The dawn of the fully closed-loop insulin pump represents the final frontier in mechanical emulation of biology for diabetes. We are transitioning away from a world where patients are tethered to the emotional and mental anxieties of a chronic condition and moving into an era of truly automated management. While challenges regarding universal access and hardware optimization remain, the underlying technology has proven that a hands-free, autonomous artificial pancreas is no longer a distant futuristic dream, it is a reality changing lives today.

 

 

CamDiab has launched CamAPS Liberty, the world's first commercially available fully closed-loop automated insulin delivery system, featuring a self-learning algorithm that removes the need for carbohydrate counting or manual meal boluses. CamAPS Liberty is a feature within CamAPS FX. When activated, users are not required to count carbohydrates or bolus before meals and snacks. Given EU MDR approval in January 2026 and unveiled at the ATTD Congress 2026. This new feature within the CamAPS FX app is designed forperiodic use to reduce the management burden for type 1 diabetes users. A controlled clinical pilot began in early 2026 and a wider commercial rollout to existing CamAPS FX app users is ongoing in stages.

 

 

Commercial rollout is a matter of regulatory approval and manufacturing scale. IDTechEx estimate that fully closed-loop, hassle-free systems will begin hitting the market in a meaningful way within the next three to five years.

 

The finish line is in sight. The future of diabetes care is not just about better management, it is about giving people their time, mental energy, and freedom back.

 

For more information on this report, including downloadable sample pages, please visit www.IDTechEx.com/Diabetes, or for the full portfolio of healthcare-related research available from IDTechEx, see www.IDTechEx.com/Research/Healthcare.