MATLAB is fast in matrix multiplication because of its underlying implementation. Firstly, MATLAB uses optimized linear algebra libraries such as LAPACK, which provides highly optimized and efficient algorithms for matrix operations. Secondly, MATLAB also uses advanced techniques such as multithreading and GPU acceleration to perform matrix multiplication quickly. Multithreading allows MATLAB to use multiple cores of a computer's processor to perform the calculation in parallel, resulting in a significant speed-up. GPU acceleration allows MATLAB to offload the calculation to the GPU, which is highly optimized for numerical operations and can perform the calculation much faster than the CPU.

In addition to these implementation details, the design of the MATLAB language itself also plays a role in its fast matrix multiplication performance. MATLAB was specifically designed for numerical computing and has a number of built-in functions that are optimized for matrix operations. This means that when you perform a matrix multiplication in MATLAB, you can be sure that it is using the most efficient algorithms available.

Finally, MATLAB also allows you to perform low-level optimization, such as using specific data types or algorithms, if you need even more performance. This level of control is not possible with other programming languages and makes MATLAB a powerful tool for matrix computation.

In conclusion, MATLAB's fast matrix multiplication performance is due to a combination of optimized linear algebra libraries, multithreading, GPU acceleration, and its design specifically for numerical computing.