diff --git a/pyproject.toml b/pyproject.toml
index 71baee3982..fdb771cb49 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -67,7 +67,7 @@ gpu_102 = ["pycuda", "cupy-cuda102", "cufinufft==1.3"]
 gpu_110 = ["pycuda", "cupy-cuda110", "cufinufft==1.3"]
 gpu_111 = ["pycuda", "cupy-cuda111", "cufinufft==1.3"]
 gpu_11x = ["pycuda", "cupy-cuda11x", "cufinufft==1.3"]
-gpu_12x = ["pycuda", "cupy-cuda12x", "cufinufft==1.3"]
+gpu_12x = ["pycuda", "cupy-cuda12x", "cufinufft==2.2.0"]
 dev = [
     "black",
     "bumpversion",
diff --git a/src/aspire/nufft/cufinufft.py b/src/aspire/nufft/cufinufft.py
index dba4cdf4a5..84c757a786 100644
--- a/src/aspire/nufft/cufinufft.py
+++ b/src/aspire/nufft/cufinufft.py
@@ -28,12 +28,6 @@ def __init__(self, sz, fourier_pts, epsilon=1e-8, ntransforms=1, **kwargs):
         # Passing "ntransforms" > 1 expects one large higher dimensional array later.
         self.ntransforms = ntransforms
 
-        # Workaround cufinufft A100 singles issue
-        # ASPIRE-Python/703
-        # Cast to doubles.
-        self._original_dtype = fourier_pts.dtype
-        fourier_pts = fourier_pts.astype(np.float64, copy=False)
-
         # Basic dtype passthough.
         dtype = fourier_pts.dtype
         if dtype == np.float64 or dtype == np.complex128:
@@ -102,12 +96,8 @@ def transform(self, signal):
         `(ntransforms, num_pts)`.
         """
 
-        # Check we're not forcing a dtype workaround for ASPIRE-Python/703,
-        #   then check if we have a dtype mismatch.
-        # This avoids false positive complaint for the workaround.
-        if (self._original_dtype == self.dtype) and not (
-            signal.dtype == self.dtype or signal.dtype == self.complex_dtype
-        ):
+        # Check dtype mismatch.
+        if not (signal.dtype == self.dtype or signal.dtype == self.complex_dtype):
             logger.warning(
                 "Incorrect dtypes passed to (a)nufft."
                 " In the future this will be an error."
@@ -143,8 +133,6 @@ def transform(self, signal):
         self._transform_plan.execute(result_gpu, signal_gpu)
 
         result = result_gpu.get()
-        # ASPIRE-Python/703
-        result = result.astype(complex_type(self._original_dtype), copy=False)
 
         return result