diff --git a/kitty/cursor_trail.c b/kitty/cursor_trail.c
index 68525b2e7..1240407f3 100644
--- a/kitty/cursor_trail.c
+++ b/kitty/cursor_trail.c
@@ -1,50 +1,59 @@
 #include "state.h"
+#include "float.h"
 
 inline static float
 norm(float x, float y) {
     return sqrtf(x * x + y * y);
 }
 
-inline static bool
-get_cursor_edge(float *left, float *right, float *top, float *bottom, Window *w) {
+static void
+update_cursor_trail_target(CursorTrail *ct, Window *w) {
+#define EDGE(axis, index) ct->cursor_edge_##axis[index]
 #define WD w->render_data
-    *left = WD.xstart + WD.screen->cursor_render_info.x * WD.dx;
-    *bottom = WD.ystart - (WD.screen->cursor_render_info.y + 1) * WD.dy;
+    float left = FLT_MAX, right = FLT_MAX, top = FLT_MAX, bottom = FLT_MAX;
     switch (WD.screen->cursor_render_info.shape) {
         case CURSOR_BLOCK:
         case CURSOR_HOLLOW:
-            *right = *left + WD.dx;
-            *top = *bottom + WD.dy;
-            return true;
         case CURSOR_BEAM:
-            *right = *left + WD.dx / WD.screen->cell_size.width * OPT(cursor_beam_thickness);
-            *top = *bottom + WD.dy;
-            return true;
         case CURSOR_UNDERLINE:
-            *right = *left + WD.dx;
-            *top = *bottom + WD.dy / WD.screen->cell_size.height * OPT(cursor_underline_thickness);
-            return true;
+            left = WD.xstart + WD.screen->cursor_render_info.x * WD.dx;
+            bottom = WD.ystart - (WD.screen->cursor_render_info.y + 1) * WD.dy;
         default:
-            return false;
+            break;
+    }
+    switch (WD.screen->cursor_render_info.shape) {
+        case CURSOR_BLOCK:
+        case CURSOR_HOLLOW:
+            right = left + WD.dx;
+            top = bottom + WD.dy;
+            break;
+        case CURSOR_BEAM:
+            right = left + WD.dx / WD.screen->cell_size.width * OPT(cursor_beam_thickness);
+            top = bottom + WD.dy;
+            break;
+        case CURSOR_UNDERLINE:
+            right = left + WD.dx;
+            top = bottom + WD.dy / WD.screen->cell_size.height * OPT(cursor_underline_thickness);
+            break;
+        default:
+            break;
+    }
+    if (left != FLT_MAX) {
+        EDGE(x, 0) = left;
+        EDGE(x, 1) = right;
+        EDGE(y, 0) = top;
+        EDGE(y, 1) = bottom;
+        ct->screen = WD.screen;
     }
 }
 
-bool
-update_cursor_trail(CursorTrail *ct, Window *w, monotonic_t now, OSWindow *os_window) {
+static bool
+update_cursor_trail_corners(CursorTrail *ct, Window *w, monotonic_t now, OSWindow *os_window) {
     // the trail corners move towards the cursor corner at a speed proportional to their distance from the cursor corner.
     // equivalent to exponential ease out animation.
-
     static const int ci[4][2] = {{1, 0}, {1, 1}, {0, 1}, {0, 0}};
     const float dx_threshold = WD.dx / WD.screen->cell_size.width;
     const float dy_threshold = WD.dy / WD.screen->cell_size.height;
-    bool needs_render_prev = ct->needs_render;
-    ct->needs_render = false;
-
-#define EDGE(axis, index) ct->cursor_edge_##axis[index]
-
-    if (!WD.screen->paused_rendering.expires_at && !get_cursor_edge(&EDGE(x, 0), &EDGE(x, 1), &EDGE(y, 0), &EDGE(y, 1), w)) {
-        return needs_render_prev;
-    }
 
     // the decay time for the trail to reach 1/1024 of its distance from the cursor corner
     float decay_fast = OPT(cursor_trail_decay_fast);
@@ -55,7 +64,8 @@ update_cursor_trail(CursorTrail *ct, Window *w, monotonic_t now, OSWindow *os_wi
             ct->corner_x[i] = EDGE(x, ci[i][0]);
             ct->corner_y[i] = EDGE(y, ci[i][1]);
         }
-    } else if (OPT(cursor_trail) < now - WD.screen->cursor->position_changed_by_client_at && ct->updated_at < now) {
+    }
+    else if (ct->updated_at < now) {
         float cursor_center_x = (EDGE(x, 0) + EDGE(x, 1)) * 0.5f;
         float cursor_center_y = (EDGE(y, 0) + EDGE(y, 1)) * 0.5f;
         float cursor_diag_2 = norm(EDGE(x, 1) - EDGE(x, 0), EDGE(y, 1) - EDGE(y, 0)) * 0.5f;
@@ -83,24 +93,34 @@ update_cursor_trail(CursorTrail *ct, Window *w, monotonic_t now, OSWindow *os_wi
             ct->corner_y[i] += dy * step;
         }
     }
+
     ct->updated_at = now;
     for (int i = 0; i < 4; ++i) {
         float dx = fabsf(EDGE(x, ci[i][0]) - ct->corner_x[i]);
         float dy = fabsf(EDGE(y, ci[i][1]) - ct->corner_y[i]);
         if (dx_threshold <= dx || dy_threshold <= dy) {
-          ct->needs_render = true;
-          break;
+            return true;
         }
     }
+    return false;
+}
 
-    if (ct->needs_render) {
-        ColorProfile *cp = WD.screen->color_profile;
-        ct->color = colorprofile_to_color(cp, cp->overridden.cursor_color, cp->configured.cursor_color).rgb;
+bool
+update_cursor_trail(CursorTrail *ct, Window *w, monotonic_t now, OSWindow *os_window) {
+    if (!WD.screen->paused_rendering.expires_at && OPT(cursor_trail) < now - WD.screen->cursor->position_changed_by_client_at) {
+        update_cursor_trail_target(ct, w);
+    }
+
+    bool needs_render_prev = ct->needs_render;
+    ct->needs_render = false;
+
+    if (update_cursor_trail_corners(ct, w, now, os_window)) {
+        ct->needs_render = true;
     }
 
-#undef EDGE
     // returning true here will cause the cells to be drawn
     return ct->needs_render || needs_render_prev;
 }
 
 #undef WD
+#undef EDGE
diff --git a/kitty/screen.h b/kitty/screen.h
index 71e78ecfb..e71c6ac6a 100644
--- a/kitty/screen.h
+++ b/kitty/screen.h
@@ -97,6 +97,7 @@ typedef struct {
     struct {
         unsigned int cursor_x, cursor_y, scrolled_by;
         index_type lines, columns;
+        color_type cursor_bg;
     } last_rendered;
     bool is_dirty, scroll_changed, reload_all_gpu_data;
     Cursor *cursor;
diff --git a/kitty/shaders.c b/kitty/shaders.c
index 38a5cc3ab..e64440169 100644
--- a/kitty/shaders.c
+++ b/kitty/shaders.c
@@ -385,6 +385,9 @@ cell_update_uniform_block(ssize_t vao_idx, Screen *screen, int uniform_buffer, c
 #undef COLOR
     rd->url_color = OPT(url_color); rd->url_style = OPT(url_style);
 
+    // store last rendered cursor color for trail rendering
+    screen->last_rendered.cursor_bg = rd->cursor_bg;
+
     unmap_vao_buffer(vao_idx, uniform_buffer); rd = NULL;
 }
 
@@ -1154,7 +1157,7 @@ draw_cursor_trail(CursorTrail *trail) {
     glUniform2fv(trail_program_layout.uniforms.cursor_edge_x, 1, trail->cursor_edge_x);
     glUniform2fv(trail_program_layout.uniforms.cursor_edge_y, 1, trail->cursor_edge_y);
 
-    color_vec3(trail_program_layout.uniforms.trail_color, trail->color);
+    color_vec3(trail_program_layout.uniforms.trail_color, trail->screen->last_rendered.cursor_bg);
 
     glDrawArrays(GL_TRIANGLE_FAN, 0, 4);
 
diff --git a/kitty/state.h b/kitty/state.h
index 08a592001..a121073c7 100644
--- a/kitty/state.h
+++ b/kitty/state.h
@@ -218,7 +218,7 @@ typedef struct {
 typedef struct {
     bool needs_render;
     monotonic_t updated_at;
-    color_type color;
+    Screen* screen;
     float corner_x[4];
     float corner_y[4];
     float cursor_edge_x[2];