One of NASA's fundamental principles is Dissimilar Redundancy, meaning, having (at least) two providers that share no critical path.
They do the same thing with all of their commercial partnerships:
So, as you can see, it is actually completely normal that NASA contracts more than one contractor for a specific service, and significantly more than one contractor for a development program. In fact, one could legitimately ask the question, why did NASA only select two contractors?
The "odd couple" in NASA's commercial services are actually the Gateway Logistics Services for cargo resupply of the Lunar Gateway and the Human Landing System. Both of these were awarded solely to SpaceX with Dragon XL and Starship HLS.
And specifically for the HLS, NASA calls out in the Source Selection Statement the fact that the main reason they only selected one contractor is that Congress only gave them enough money for one. (In fact, Congress didn't even give them enough money for one, but SpaceX was by far the cheapest bidder and NASA was able to re-negotiate payment terms with them to fit the budget.)
If it weren't for this financial restriction, NASA very likely would have awarded two contracts, if not all three.
You can see how important Dissimilar Redundancy is, if you consider that:
- After the SpaceX CRS-7 explosion, Falcon 9 was grounded, but NASA could keep on re-supplying the ISS using Cygnus (plus, there is also the Japanese HTV and the Russian Progress).
- After the AMOS-6 explosion, Falcon 9 was again grounded.
- After the Cygnus Orb-3 explosion, Antares was grounded, but Orbital Sciences could continue launching Cygnus missions to the ISS on ULA's Atlas V. This is an example of nested dissimilar redundancy: not only does NASA have two different contractors (soon three) re-supplying the ISS with cargo, but one of them can even use different launch vehicles.
- Cygnus is launching on the Antares rocket, which is a Ukrainian rocket using Russian engines, both which are currently hard to come by for obvious reasons. Therefore, two Cargo Resupply Missions using Cygnus will launch on a SpaceX Falcon 9 until the new Antares 330 is ready which no longer uses any Russian or Ukrainian parts. (This shows both the redundancy built into Cygnus by being able to launch on different launchers and the lack of redundancy in the US launch market caused by SpaceX's towering dominance.)
- Boeing's Starliner program is significantly delayed due to first valve problems, then software problems, then again valve problems.
- SpaceX's Crew Dragon program was also delayed due to parachute problems, then a valve problem which lead to a spacecraft exploding.
The most obvious example, though is the Space Shuttle, for which NASA had no redundancy, and thus had no capability of launching crew during the grounding of the fleet after the Challenger accident and the Columbia accident, and after the retirement of the fleet.
There is also an interesting pattern in the way NASA awards those contracts: in several of them, NASA chooses an "established player" and a "newSPACE" contender, for example with CRS (Northrop Grumman and SpaceX) and CCP (Boeing and SpaceX). It is also likely they wanted to do the same with the HLS (SpaceX and National Team).
They did the same here: Collins / ILC Dover is an established player, they already built the current ISS suits as well as the Apollo suits. And Axiom is a newSPACE company. What is somewhat interesting is that NASA awarded the contract for developing new suits to Collins only a couple of days after NASA put a stop on ISS spacewalks precisely because of unsolved problems with the current Collins spacesuits.