Phasing a 30-Robot Hospital Tender: Start With Three

What does a phased hospital robot rollout look like?

A phased rollout means agreeing to a fleet ceiling but committing money and staff time only to a first group of robots. Each next batch is released against results the hospital measured on its own corridors, elevators and shifts.

Sheba Medical Center in Israel is a working example. Its public tender permits up to 30 robots, and according to Pulse Equity, citing company data reported in October 2026, three are deployed now. The tender structure is the lesson here, more than any single performance figure.

The practical questions are which routes go first, what evidence unlocks the next batch, and when a rollout should stop. In short: pick one high cost of delay route with a fixed path, set measurable triggers before go live, and write the exit criteria before the first robot moves.

What do the Sheba figures actually show?

Pulse Equity's release reports that average chemo delivery time fell from about 40 minutes to about 14 minutes over 11 months, a reduction the release puts at roughly 65 percent. Gix's own website lists the same 40 to 14 minute result for the chemo use case, and it states that the figures are company reported and not independently audited. We found no independent audit in the sources reviewed.

Gix also says the figures describe the chemo use case only, not a group wide average. That distinction matters. A hospital reading the headline should ask what the comparison baseline was, which shift it covered and how many runs sat inside the sample.

The timeline needs care as well. Several sources date commercial operations to December 2025, while one headline frames the story as beginning in October 2026. The 9,000 plus autonomous missions figure also comes from company data. Use these numbers as a vendor's account to test, not as proof that your floors will behave the same way.

Which routes should go first?

Pick the route where a delay costs the most and where the path repeats. Sheba started with chemotherapy medication runs for the oncology department, a trip with a clear origin, a clear destination and time pressure that nurses already feel every day. Older coverage of Sheba's pilot described a refrigerated, locked compartment with alerts, which is the kind of controlled load that suits a first batch.

Expansion came later and in sequence. On October 12, 2026, Gix announced that robots had begun carrying blood and bone marrow specimens. Routes named in its coverage include bone marrow from hematology to the lab, and donated blood from the donation center to the blood bank.

Those specimen routes are new handoffs with their own custody steps, so they belong in a second wave. Once the first route runs on a predictable schedule with few exceptions, the team can add a route that shares a corridor or elevator bank without changing everything at once.

What result should trigger the next batch?

A trigger is a number agreed before go live and measured against a baseline captured from the existing courier or nurse run on the same route, hours and days. Without that baseline, every later claim is a guess. Set the measures in writing, and keep the definitions fixed for the whole pilot.

Most facility teams should track at least three measures for the first route:

• Delivery time against the manual baseline, measured door to door on the same route and hours
• Completed runs against scheduled runs, with each failure logged by cause such as elevator wait, blocked door or docking error
• Handoff accuracy at the receiving end, including temperature or custody checks where the load requires them

When should a rollout pause or stop?

Exit criteria keep a phased program from becoming a sunk cost. They should be as specific as the triggers. A route should pause, not limp along, when a problem persists after the vendor has had a fair chance to correct it.

Common stop conditions include the following:

• A safety event on the route that the vendor cannot explain and close with a documented fix
• Sustained performance below the trigger after workflow adjustments have been tried
• Staff time on exceptions, rescues and re-sends that exceeds the time the manual route used to take
• Repair response that misses the agreed window on faults that take the route offline

How does a leased fleet fit a phased program?

Service Robot Co. is a full service commercial robot integrator, and we lease robots only. A customer takes each robot on a three year lease with one fixed payment quoted after a site assessment, so the first batch can be sized to the first route rather than to the full tender.

We work across manufacturers, so the robot for a short, repeatable corridor can differ from the one suited to a longer or heavier run. Finance, deployment, integration, staff training and servicing come from one vendor, delivered through a nationwide US engineer network, which keeps each phase gate in one place.

A phased program still needs written triggers and exit criteria. A lease simply makes each batch a decision rather than a commitment to the whole fleet.