diff --git a/src/concurrent/concurrent_queues.zig b/src/concurrent/concurrent_queues.zig
new file mode 100644
index 0000000..b407ad1
--- /dev/null
+++ b/src/concurrent/concurrent_queues.zig
@@ -0,0 +1,323 @@
+const std = @import("std");
+const AtomicOrder = std.atomic.Ordering;
+const Cache_Line_Size = 64;
+
+/// Constructs a new multi-producer, multi-consumer queue for the given `T` type.
+/// The queue is lock-free and wait-free for both producers and consumers.
+/// The queue has a fixed capacity that is rounded up to the nearest power of two.
+/// The queue is thread-safe and can be safely accessed from multiple threads concurrently.
+pub fn MpmcQueue(comptime T: type) type {
+    return struct {
+        const Self = @This();
+
+        const Node = struct {
+            value: T,
+            sequence: usize,
+        };
+
+        const Buffer = struct {
+            entries: []Node,
+            mask: usize,
+        };
+
+        head: usize align(Cache_Line_Size) = 0,
+        tail: usize align(Cache_Line_Size) = 0,
+        buffer: Buffer,
+
+        /// Initializes a new `MpmcQueue` with the given `capacity`.
+        /// The actual capacity of the queue will be rounded up to the nearest power of two.
+        /// If the requested capacity is too large, this function will return an error.
+        pub fn init(allocator: std.mem.Allocator, capacity: usize) !Self {
+            const real_capacity = std.math.ceilPowerOfTwo(usize, capacity) catch return error.CapacityTooLarge;
+
+            const entries = try allocator.alloc(Node, real_capacity);
+            for (entries, 0..) |*entry, i| {
+                entry.* = .{
+                    .value = undefined,
+                    .sequence = i,
+                };
+            }
+
+            return Self{
+                .buffer = .{
+                    .entries = entries,
+                    .mask = real_capacity - 1,
+                },
+            };
+        }
+
+        /// Deinitializes the `MpmcQueue` and frees the underlying memory used by the queue.
+        /// This function should be called when the queue is no longer needed to avoid memory leaks.
+        pub fn deinit(self: *Self, allocator: std.mem.Allocator) void {
+            allocator.free(self.buffer.entries);
+        }
+
+        /// Adds the given `value` to the queue.
+        /// This operation is thread-safe and can be called concurrently from multiple threads.
+        /// Returns `true` if the value was successfully added to the queue, or `false` if the queue is full.
+        pub fn push(self: *Self, value: T) bool {
+            var tail = @atomicLoad(usize, &self.tail, .monotonic);
+            while (true) {
+                const entry = &self.buffer.entries[tail & self.buffer.mask];
+                const seq = @atomicLoad(usize, &entry.sequence, .acquire);
+                const diff = @as(isize, @intCast(seq)) - @as(isize, @intCast(tail));
+
+                if (diff == 0) {
+                    if (@cmpxchgWeak(
+                        usize,
+                        &self.tail,
+                        tail,
+                        tail + 1,
+                        .monotonic,
+                        .monotonic,
+                    )) |_| {
+                        tail += 1;
+                        continue;
+                    }
+                    entry.value = value;
+                    @atomicStore(usize, &entry.sequence, tail + 1, .release);
+                    return true;
+                } else if (diff < 0) {
+                    return false;
+                }
+                tail = @atomicLoad(usize, &self.tail, .monotonic);
+            }
+        }
+
+        /// Removes and returns the next value from the queue, or `null` if the queue is empty.
+        /// This operation is thread-safe and can be called concurrently from multiple threads.
+        pub fn pop(self: *Self) ?T {
+            var head = @atomicLoad(usize, &self.head, .monotonic);
+            while (true) {
+                const entry = &self.buffer.entries[head & self.buffer.mask];
+                const seq = @atomicLoad(usize, &entry.sequence, .acquire);
+                const diff = @as(isize, @intCast(seq)) - @as(isize, @intCast(head + 1));
+
+                if (diff == 0) {
+                    if (@cmpxchgWeak(
+                        usize,
+                        &self.head,
+                        head,
+                        head + 1,
+                        .monotonic,
+                        .monotonic,
+                    )) |_| {
+                        head += 1;
+                        continue;
+                    }
+                    const value = entry.value;
+                    @atomicStore(usize, &entry.sequence, head + self.buffer.entries.len, .release);
+                    return value;
+                } else if (diff < 0) {
+                    return null;
+                }
+                head = @atomicLoad(usize, &self.head, .monotonic);
+            }
+        }
+    };
+}
+
+test "MpmcQueue.push - basic push operation" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+
+    try std.testing.expect(queue.push(1));
+    try std.testing.expect(queue.push(2));
+    try std.testing.expect(queue.push(3));
+}
+
+test "MpmcQueue.push - queue full" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 2) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+
+    try std.testing.expect(queue.push(1));
+    try std.testing.expect(queue.push(2));
+    try std.testing.expect(!queue.push(3));
+}
+
+test "MpmcQueue.push - wrap around" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+
+    try std.testing.expect(queue.push(1));
+    try std.testing.expect(queue.push(2));
+    _ = queue.pop();
+    _ = queue.pop();
+    try std.testing.expect(queue.push(3));
+    try std.testing.expect(queue.push(4));
+}
+
+test "MpmcQueue.push - different types" {
+    var queue = MpmcQueue(f32).init(std.testing.allocator, 2) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+
+    try std.testing.expect(queue.push(1.5));
+    try std.testing.expect(queue.push(2.5));
+}
+
+test "MpmcQueue.pop - basic pop operation" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+    try std.testing.expect(queue.push(1));
+    try std.testing.expect(queue.push(2));
+    try std.testing.expect(queue.push(3));
+    try std.testing.expectEqual(queue.pop(), 1);
+    try std.testing.expectEqual(queue.pop(), 2);
+    try std.testing.expectEqual(queue.pop(), 3);
+}
+
+test "MpmcQueue.pop - queue empty" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+    try std.testing.expectEqual(queue.pop(), null);
+}
+
+test "MpmcQueue.pop - wrap around" {
+    var queue = MpmcQueue(i32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+    try std.testing.expect(queue.push(1));
+    try std.testing.expect(queue.push(2));
+    try std.testing.expect(queue.push(3));
+    try std.testing.expectEqual(queue.pop(), 1);
+    try std.testing.expect(queue.push(4));
+    try std.testing.expect(queue.push(5));
+    try std.testing.expectEqual(queue.pop(), 2);
+    try std.testing.expectEqual(queue.pop(), 3);
+    try std.testing.expectEqual(queue.pop(), 4);
+    try std.testing.expectEqual(queue.pop(), 5);
+}
+
+test "MpmcQueue.pop - different types" {
+    var queue = MpmcQueue(f32).init(std.testing.allocator, 4) catch unreachable;
+    defer queue.deinit(std.testing.allocator);
+    try std.testing.expect(queue.push(1.5));
+    try std.testing.expect(queue.push(2.5));
+    try std.testing.expectEqual(queue.pop(), 1.5);
+    try std.testing.expectEqual(queue.pop(), 2.5);
+}
+
+test "MpmcQueue.pop - concurrent push and pop" {
+    const TestQueue = MpmcQueue(u64);
+    const ThreadCount = 4;
+    const ItemsPerThread = 10_000;
+
+    var queue = try TestQueue.init(std.testing.allocator, ThreadCount * 2);
+    defer queue.deinit(std.testing.allocator);
+
+    const Producer = struct {
+        fn run(q: *TestQueue, thread_id: u64) !void {
+            var i: u64 = 0;
+            while (i < ItemsPerThread) : (i += 1) {
+                const item = thread_id * ItemsPerThread + i;
+                while (!q.push(item)) {
+                    try std.Thread.yield();
+                }
+            }
+        }
+    };
+
+    const Consumer = struct {
+        fn run(q: *TestQueue, results: []std.atomic.Value(u64)) !void {
+            var count: u64 = 0;
+            while (count < ItemsPerThread) {
+                if (q.pop()) |value| {
+                    const index = value % ThreadCount;
+                    _ = results[index].fetchAdd(1, .acq_rel);
+                    count += 1;
+                } else {
+                    try std.Thread.yield();
+                }
+            }
+        }
+    };
+
+    var threads: [ThreadCount * 2]std.Thread = undefined;
+    var results: [ThreadCount]std.atomic.Value(u64) = undefined;
+
+    for (&results) |*result| {
+        result.* = std.atomic.Value(u64).init(0);
+    }
+
+    // Start producer threads
+    for (0..ThreadCount) |i| {
+        threads[i] = try std.Thread.spawn(.{}, Producer.run, .{ &queue, i });
+    }
+
+    // Start consumer threads
+    for (0..ThreadCount) |i| {
+        threads[ThreadCount + i] = try std.Thread.spawn(.{}, Consumer.run, .{ &queue, &results });
+    }
+
+    // Wait for all threads to complete
+    for (threads) |thread| {
+        thread.join();
+    }
+
+    // Verify results
+    for (results) |result| {
+        try std.testing.expectEqual(ItemsPerThread, result.load(.acquire));
+    }
+}
+
+test "MpmcQueue.pop - concurrent push and pop with wrap around" {
+    const TestQueue = MpmcQueue(u64);
+    const ThreadCount = 4;
+    const ItemsPerThread = 10_000;
+
+    var queue = try TestQueue.init(std.testing.allocator, ThreadCount * 2);
+    defer queue.deinit(std.testing.allocator);
+
+    const Producer = struct {
+        fn run(q: *TestQueue, thread_id: u64) !void {
+            var i: u64 = 0;
+            while (i < ItemsPerThread) : (i += 1) {
+                const item = thread_id * ItemsPerThread + i;
+                while (!q.push(item)) {
+                    try std.Thread.yield();
+                }
+            }
+        }
+    };
+
+    const Consumer = struct {
+        fn run(q: *TestQueue, results: []std.atomic.Value(u64)) !void {
+            var count: u64 = 0;
+            while (count < ItemsPerThread) {
+                if (q.pop()) |value| {
+                    const index = value % ThreadCount;
+                    _ = results[index].fetchAdd(1, .acq_rel);
+                    count += 1;
+                } else {
+                    try std.Thread.yield();
+                }
+            }
+        }
+    };
+
+    var threads: [ThreadCount * 2]std.Thread = undefined;
+    var results: [ThreadCount]std.atomic.Value(u64) = undefined;
+
+    for (&results) |*result| {
+        result.* = std.atomic.Value(u64).init(0);
+    }
+
+    // Start producer threads
+    for (0..ThreadCount) |i| {
+        threads[i] = try std.Thread.spawn(.{}, Producer.run, .{ &queue, i });
+    }
+
+    // Start consumer threads
+    for (0..ThreadCount) |i| {
+        threads[ThreadCount + i] = try std.Thread.spawn(.{}, Consumer.run, .{ &queue, &results });
+    }
+
+    // Wait for all threads to complete
+    for (threads) |thread| {
+        thread.join();
+    }
+
+    // Verify results
+    for (results) |result| {
+        try std.testing.expectEqual(ItemsPerThread, result.load(.acquire));
+    }
+}
diff --git a/src/concurrent/concurrent_queues_test.zig b/src/concurrent/concurrent_queues_test.zig
new file mode 100644
index 0000000..74beb14
--- /dev/null
+++ b/src/concurrent/concurrent_queues_test.zig
@@ -0,0 +1,41 @@
+const std = @import("std");
+const testing = std.testing;
+const ConcurrentQueue = @import("concurrent_queues.zig").ConcurrentQueue;
+
+test "ConcurrentQueue.push - basic push operation" {
+    var queue = ConcurrentQueue(i32, .Spsc).init(testing.allocator, 4) catch unreachable;
+    defer queue.deinit();
+
+    try testing.expect(queue.push(1));
+    try testing.expect(queue.push(2));
+    try testing.expect(queue.push(3));
+}
+
+test "ConcurrentQueue.push - queue full" {
+    var queue = ConcurrentQueue(i32, .Spsc).init(testing.allocator, 2) catch unreachable;
+    defer queue.deinit();
+
+    try testing.expect(queue.push(1));
+    try testing.expect(queue.push(2));
+    try testing.expect(!queue.push(3));
+}
+
+test "ConcurrentQueue.push - wrap around" {
+    var queue = ConcurrentQueue(i32, .Spsc).init(testing.allocator, 4) catch unreachable;
+    defer queue.deinit();
+
+    try testing.expect(queue.push(1));
+    try testing.expect(queue.push(2));
+    _ = queue.pop();
+    _ = queue.pop();
+    try testing.expect(queue.push(3));
+    try testing.expect(queue.push(4));
+}
+
+test "ConcurrentQueue.push - different types" {
+    var queue = ConcurrentQueue(f32, .Spsc).init(testing.allocator, 2) catch unreachable;
+    defer queue.deinit();
+
+    try testing.expect(queue.push(1.5));
+    try testing.expect(queue.push(2.5));
+}
diff --git a/src/root.zig b/src/root.zig
index 110175c..c38607a 100644
--- a/src/root.zig
+++ b/src/root.zig
@@ -33,6 +33,7 @@ pub const justification_finalization_helper = @import("consensus/helpers/justifi
 pub const finality = @import("consensus/helpers/finality.zig");
 pub const rewards_penalties_helper = @import("consensus/helpers/rewards_penalties.zig");
 pub const segment_storage = @import("storage/segment_storage.zig");
+pub const queues = @import("concurrent/concurrent_queues.zig");
 
 test {
     @import("std").testing.refAllDeclsRecursive(@This());