Stellantis has signed a memorandum of understanding with Roshel, a Canadian armoured- and defence-vehicle manufacturer, covering the potential sale of its idled Brampton Assembly plant in Ontario. Unifor declared talks at an impasse on 17 September. The collective agreement covering more than 9,000 Stellantis workers in Canada expires at 11:59 p.m. on 20 September 2026.
Three facts, in that order, describe how a car plant stops being a car plant.
What Happened
Brampton has been idled since 2023. It was retooled to build the next-generation Jeep Compass. That production is now going to Illinois instead — a decision Stellantis has tied to changes in US trade policy.
The company's stated position is blunt: it no longer has a sustainable business case for building the Compass at Brampton.
With no product and no business case, the plant becomes real estate. Roshel — which builds armoured vehicles, not passenger cars — has signed an MOU to potentially acquire it, and has said it would prioritise hiring affected Stellantis workers if the deal progresses.
Bargaining opened on 1 September 2026. Unifor National President Lana Payne wrote to Industry Minister Mélanie Joly asking the federal government to block the sale and to clarify whether the MOU breaches commitments Stellantis made to Ottawa. Joly's response was that the government wants to see "a new model from Stellantis" and will apply "maximum pressure."
Unifor has not ruled out a strike. Payne's framing of the union's position is the cleanest summary of the dispute: the options are not Roshel or nothing.
The Real Dispute Is Not About Brampton
A retooled assembly plant that has sat empty for three years is not a factory. It is a commitment the company made and a bill it has decided not to pay.
Strip out the politics and the disagreement is about what a public commitment is worth once the economics move.
Stellantis took Canadian policy support and made Canadian footprint commitments. Then US trade policy changed the arithmetic of building a Compass for the North American market in Ontario versus Illinois. The company followed the arithmetic.
That is the pattern worth noting, because it is not confined to Canada:
- Tariff policy is now a production-location instrument. Where trade measures make cross-border assembly expensive, capacity migrates toward the largest protected market. Canada is on the wrong side of that gradient for North American volume.
- Idled capacity converts, and rarely back into cars. A plant that sits out a product cycle loses its workforce, its supplier proximity and its place in the network. Defence and logistics buyers can use the shell. Automotive rarely comes back.
- Government leverage is retrospective. Ottawa's instrument is pressure over past commitments, not a mechanism to compel future production. "Maximum pressure" is a statement about political cost, not contractual obligation.
What Unifor Is Actually Bargaining For
The strike threat is leverage over the sale, not over wages — which is unusual, and important.
A conventional round of bargaining trades pay and benefits against job security language. This one is about whether the employer retains an asset it has already decided is surplus. The union is using an expiring contract covering 9,000 workers across multiple Canadian facilities to contest the disposal of one idled site.
That gives Unifor real short-term power and weak long-term position. A strike can impose cost on running plants. It cannot create a business case for a product Stellantis has already allocated to Illinois.
The more achievable outcome is a different commitment attached to the sale — product allocation elsewhere in Canada, investment guarantees, or transition and hiring terms binding on Roshel — rather than a reversal.
The Roshel Angle Deserves Attention
A defence manufacturer buying an idled passenger-car plant is not an accident of timing. Armoured-vehicle demand has been structurally strong, and the skills, floor space, paint capability and supplier base of automotive assembly transfer reasonably well.
What does not transfer is scale of employment. Volume automotive assembly is one of the most labour-dense manufacturing activities there is. Specialist armoured-vehicle production is not. "Prioritise hiring affected workers" is a meaningful commitment at Roshel's headcount, not at Brampton's.
That gap — between a plant saved and a workforce retained — is what the union is really contesting.
What to Watch
- Whether the MOU converts to a definitive agreement, and what employment terms are attached if it does.
- Whether Ottawa's "maximum pressure" acquires an instrument. Clawback of past support would be a materially different signal to every automaker holding Canadian commitments.
- Whether strike action, if taken, is aimed at Brampton or at Windsor-scale leverage. The first is symbolic. The second is expensive.
Why It Matters
This is the second time in a month that a legacy automaker's European or North American footprint decision has been made on trade-policy grounds rather than demand. Volvo split its model line-up by region because one car no longer sells everywhere. Stellantis is moving assembly across a border because one tariff regime no longer treats North America as one market.
The common thread: the single integrated production footprint that defined the industry for thirty years is being unbundled by policy, one plant at a time. Brampton is what that looks like on the ground — a retooled plant, a reallocated product, and 9,000 people negotiating over what the commitment was worth.
