diff --git a/include/PICA/shader.hpp b/include/PICA/shader.hpp
index 7136312f..195733e0 100644
--- a/include/PICA/shader.hpp
+++ b/include/PICA/shader.hpp
@@ -17,7 +17,8 @@ namespace ShaderOpcodes {
 		MUL = 0x08,
 		MOVA = 0x12,
 		MOV = 0x13,
-		END = 0x22
+		END = 0x22,
+		LOOP = 0x29
 	};
 }
 
@@ -25,6 +26,13 @@ class PICAShader {
 	using f24 = Floats::f24;
 	using vec4f = OpenGL::Vector<f24, 4>;
 
+	struct Loop {
+		u32 startingPC; // PC at the start of the loop
+		u32 endingPC;   // PC at the end of the loop
+		u32 iterations; // How many iterations of the loop to run
+		u32 increment;  // How much to increment the loop counter after each iteration
+	};
+
 	int bufferIndex; // Index of the next instruction to overwrite for shader uploads
 	int opDescriptorIndex; // Index of the next operand descriptor we'll overwrite
 	u32 floatUniformIndex = 0; // Which float uniform are we writing to? ([0, 95] range)
@@ -35,9 +43,12 @@ class PICAShader {
 	std::array<u32, 128> operandDescriptors;
 	std::array<vec4f, 16> tempRegisters; // General purpose registers the shader can use for temp values
 	OpenGL::Vector<s32, 2> addrRegister; // Address register
-	u32 pc = 0; // Program counter: Index of the next instruction we're going to execute
-	
 	u32 loopCounter;
+
+	u32 pc = 0; // Program counter: Index of the next instruction we're going to execute
+	u32 loopIndex = 0; // The index of our loop stack (0 = empty, 4 = full)
+	std::array<Loop, 4> loopInfo;
+
 	ShaderType type;
 
 	vec4f getSource(u32 source);
@@ -46,6 +57,7 @@ class PICAShader {
 	// Shader opcodes
 	void add(u32 instruction);
 	void dp4(u32 instruction);
+	void loop(u32 instruction);
 	void mov(u32 instruction);
 	void mova(u32 instruction);
 	void mul(u32 instruction);
diff --git a/src/core/PICA/shader_interpreter.cpp b/src/core/PICA/shader_interpreter.cpp
index b1b1b6d7..af75f47e 100644
--- a/src/core/PICA/shader_interpreter.cpp
+++ b/src/core/PICA/shader_interpreter.cpp
@@ -11,11 +11,25 @@ void PICAShader::run() {
 			case ShaderOpcodes::ADD: add(instruction); break;
 			case ShaderOpcodes::DP4: dp4(instruction); break;
 			case ShaderOpcodes::END: return; // Stop running shader
+			case ShaderOpcodes::LOOP: loop(instruction); break;
 			case ShaderOpcodes::MOV: mov(instruction); break;
 			case ShaderOpcodes::MOVA: mova(instruction); break;
 			case ShaderOpcodes::MUL: mul(instruction); break;
 			default:Helpers::panic("Unimplemented PICA instruction %08X (Opcode = %02X)", instruction, opcode);
 		}
+
+		// Handle loop
+		if (loopIndex != 0) {
+			auto& loop = loopInfo[loopIndex - 1];
+			if (pc == loop.endingPC) { // Check if the loop needs to start over
+				loop.iterations -= 1;
+				if (loop.iterations == 0) // If the loop ended, go one level down on the loop stack
+					loopIndex -= 1;
+
+				loopCounter += loop.increment;
+				pc = loop.startingPC;
+			}
+		}
 	}
 }
 
@@ -153,4 +167,19 @@ void PICAShader::dp4(u32 instruction) {
 			destVector[3 - i] = dot;
 		}
 	}
+}
+
+void PICAShader::loop(u32 instruction) {
+	if (loopIndex >= 4) [[unlikely]]
+		Helpers::panic("[PICA] Overflowed loop stack");
+
+	u32 dest = (instruction >> 10) & 0xfff;
+	auto& uniform = intUniforms[(instruction >> 22) & 3]; // The uniform we'll get loop info from
+	loopCounter = uniform.y();
+	auto& loop = loopInfo[loopIndex++];
+
+	loop.startingPC = pc;
+	loop.endingPC = dest + 1; // Loop is inclusive so we need + 1 here
+	loop.iterations = uniform.x() + 1;
+	loop.increment = uniform.z();
 }
\ No newline at end of file