Still, as far as is commonly known, mathematically cracking encryption where the algorithm is good and keys are large and unique remains impractical for conventional computers. If they’re secretly way ahead on quantum computing (which seems unlikely), or if they have discovered mathematical vulnerabilities in common algorithms that have not been published, then that’s a different matter. But as far as we know, it must still be difficult for them to attack encryption directly. You suggest you know more than you can say, but if I were them I’d be looking at putting backdoors into phone/computer hardware to get hold of communications before they are E2E encrypted, and/or placing subtle vulnerabilities in open-source code.