Making more ventilators in time to help coronavirus victims is hard enough. But what about manufacturing the sophisticated equipment needed to clinically test them? Or hospital beds for the patients using them?

Those are the sorts of questions occupying John Carlson, a former Johnson & Johnson executive who now oversees medical manufacturing for multinational Flex. As a contract manufacturer for medical-device companies (among other industries), Flex makes ventilators and other in-demand health care equipment at 22 locations globally, including several in the U.S. and Mexico.

Now it’s also working with some unspecified “nontraditional” companies (think automakers), to help them switch over their production lines to the highly regulated world of medical-device manufacturing. “They don’t even know what the basic rules and regulations are to do this,” Carlson says. “But they have a different set of supply networks than we do, so they can tap into different people with different capabilities. There’s a collaborative spirit right now that’s really about saving lives and getting product out there.”

In the following edited and condensed conversation with Fortune, Carlson explains some of the other challenges for companies racing to fix the global ventilator shortage.

Fortune: Where are you seeing specific constraints in the supply chain for making ventilators?

Carlson: Inside any medical device, you’ll have a bunch of common components that you could buy off the shelf, and then a few very custom components that we’ll spend a lot of time making sure we have a supply of. For ventilators, it’s typically components relating to the pump system and the valve where you’ll find the [supply-chain] bottleneck.

What else is holding back manufacturers from producing medical equipment in time to meet U.S. demand?

The final testing of the products usually requires very specific equipment. Being able to rapidly scale that test is another area that can be a constraint. For something like a ventilator, you’ll have an artificial-type lung that the device has to be tested against, to make sure that it has the right flow rate, the right pressure control. Those pieces of the test equipment can be pretty elaborate. Duplicating them is something that we jump on very quickly to make sure that not only can we get the parts, but can we assemble them and [make sure] they perform as they’re intended.

Shortages of ventilators and N95 masks have gotten a lot of attention. What other kinds of medical devices are you seeing particular need for?

Oxygen concentrators, or a [basic] portable system to generate oxygen for patients with compromised lung function. When you’re in a hospital there’s a bunch of infrastructure built into it: You’ve got vacuum systems in the wall. You’ve got oxygen. You’ve got IV systems. But now they’re trying to put patients into hotel rooms and dorm rooms and any place you can to keep them out of the hospital—and none of that infrastructure exists in a hotel room. So some of those basic supplies, as well as simple things like hospital beds, are products where we’re seeing an increase in demand.

Flex manufactures medical devices around the world, but this pandemic has had equally global reach. How have your operations—and your employees—been affected?

We screen their temperatures before they come into the plant. We issued them masks. We check on them regularly. We’re putting in systems and processes to allow them to maintain appropriate distances. We learned a lot with our manufacturing environment in China in February. All this started over the Chinese New Year, where many of our employees [were traveling], and then they were locked down and prevented from coming back. But over the last several weeks, we’ve been able to bring it back, and we’re back to basically full production across all of our China facilities.

The U.S. has less than 20% of the ventilators it might need for hospitalized coronavirus patients, according to some estimates. Can you and your partners ramp up in time?

[Long pause] I don’t know how to answer that. Typically a program like this would take anywhere from 12 to 24 months, and in what used to be months we’re now getting done in days and weeks. We will do enough that it can have a significant impact on the lives of many people. I just don’t know how to quantify if it’s fast enough. It’ll never be fast enough—because if there’s someone today who doesn’t have a product, then we’re not fast enough.