There's also another kid on the block: Julia with a syntax similar to Matlab but with speeds orders of magnitude faster than Matlab ie near C/C++ speeds. Julia is especially useful for organizations that prototype algorithms in MATLAB and then convert them to C/C++ for production. With Julia, they don't have to spend developer time recoding and optimizing the Matlab code.
Julia Interoperates well with both Python and Fortran which is useful for handling legacy code and it's free.
The one thing Matlab has over Julia is its developer environment and its large base of tutorials, packages and tool kits.