diff --git a/Project.toml b/Project.toml
index 8cdff5c..72afd49 100644
--- a/Project.toml
+++ b/Project.toml
@@ -18,25 +18,28 @@ Tensors = "48a634ad-e948-5137-8d70-aa71f2a747f4"
 
 [weakdeps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
+AMDGPU = "21141c5a-9bdb-4563-92ae-f87d6854732e"
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 Exodus = "f57ae99e-f805-4780-bdca-96e224be1e5a"
 
 [extensions]
 FiniteElementContainersAdaptExt = "Adapt"
+FiniteElementContainersAMDGPUExt = ["Adapt", "AMDGPU"]
 FiniteElementContainersCUDAExt = ["Adapt", "CUDA"]
 FiniteElementContainersExodusExt = "Exodus"
 
 [compat]
 AcceleratedKernels = "0.3"
 Adapt = "3, 4"
+AMDGPU = "1"
 Aqua = "0.8"
 Atomix = "1"
 CUDA = "5"
 DocStringExtensions = "0.9"
 Exodus = "0.13"
 JET = "0.9"
-Krylov = "0.9"
 KernelAbstractions = "0.9"
+Krylov = "0.9"
 LinearAlgebra = "1"
 Parameters = "0.12"
 Reexport = "1"
diff --git a/ext/FiniteElementContainersAMDGPUExt.jl b/ext/FiniteElementContainersAMDGPUExt.jl
new file mode 100644
index 0000000..91f3d30
--- /dev/null
+++ b/ext/FiniteElementContainersAMDGPUExt.jl
@@ -0,0 +1,26 @@
+module FiniteElementContainersAMDGPUExt
+
+using Adapt
+using AMDGPU
+using FiniteElementContainers
+using KernelAbstractions
+
+FiniteElementContainers.gpu(x) = adapt_structure(ROCArray, x)
+
+function AMDGPU.rocSPARSE.ROCSparseMatrixCSC(asm::SparseMatrixAssembler)
+  # TODO Not sure what the AMD Backend is called in KernelAbstractions
+  # I couldn't quite figure it out. This assert statement below though
+  # would be good for error checking and device consistency.
+  # @assert typeof(get_backend(asm)) <: CUDABackend "Assembler is not on a CUDA device"
+  @assert length(asm.pattern.cscnzval) > 0 "Need to assemble the assembler once with SparseArrays.sparse!(assembler)"
+  @assert all(x -> x != zero(eltype(asm.pattern.cscnzval)), asm.pattern.cscnzval) "Need to assemble the assembler once with SparseArrays.sparse!(assembler)"
+  n_dofs = FiniteElementContainers.num_unknowns(asm.dof)
+  return AMDGPU.rocSPARSE.ROCSparseMatrixCSC(
+    asm.pattern.csccolptr,
+    asm.pattern.cscrowval,
+    asm.pattern.cscnzval,
+    (n_dofs, n_dofs)
+  )
+end
+
+end # module