Yes, Junger v. Daley held that source code is speech for the purpose of First Amendment analysis in the Sixth Circuit. (The question has not yet been presented to the Supreme Court.)

But the inquiry doesn't stop there. First Amendment jurisprudence doesn't stop at asking the question "is this speech?". That's just a threshold question. If the answer is "no," then the case is analyzed one way. If the answer is "yes," then the case analyzed a different way. Think of it like a decision step in a flowchart.

Code being treated as speech doesn't automatically mean that software developers can do whatever they want, just like it doesn't mean that people who write, draw, make music, etc. can do whatever they want. It's not the magic unassailable defense you claim it to be. Courts will apply a balancing test to determine whether the Government's interest in prohibiting the speech outweighs the Constitutional interest in protecting it. The answer comes out different ways depending on the facts and the law, which is complicated.

For example, it's unlawful to make and distribute CSAM. It's unlawful to defame someone. It's unlawful to defraud someone. It's unlawful to violate others' copyrights. You have to label food and medicine a certain way. These are all forms of speech, but the courts have held that none of these restrictions (or requirements) violate the First Amendment. And I don't think we'd all be cheering someone on and claiming "code is speech" for someone who writes ransomware or software whose sole purpose is to drain your bank accounts or shut down power plants.