What evidence do you have to support that claim? Did you see what Apple was able to do with the A14 on the new iPad Air?
My only basis for saying this is Geekbench results for the A12Z, Radeon Pro 5300M, 5500M and 5600M, and reports of A14 GPU increases compared to the A12.
https://browser.geekbench.com/metal-benchmarks
As I understand, the A14 has only 4 GPU cores compared to A12Z's 8 cores. We don't have any real benchmarks yet, but some guesses have been made at:
https://www.cpu-monkey.com/en/compare_cpu-apple_a14_bionic-1693-vs-apple_a12z_bionic-1651
Apple’s stunning new iPhone 12 range has a radical upgrade everyone will love...
www.forbes.com
So I agree than an A14Z with 8 GPU cores would be quite a bit better than the A12Z - about 40% increase is quoted for the A12 vs A14 (non-Z versions) - and if we apply to an A14Z (from which we think the first ASi Mac will be derived) this still would give it a Geekbench Metal score of 14000-15000, vs the c. 25000 for the AMD 5300M, 29000 for the AMD 5500M and 40,000 for the 5600M.
My point was that going from A12Z to AMD 5600M levels is a 400% increase, which sounds like quite a challenge for Apple.
It's by no means impossible considering that a MBP16 SoC would probably be 50-80W TDP compared to the iPad Pros 15W (estimated), but I would be surprised (pleasantly!) if Apple were able to push the improvements in Apple Silicon this quickly. Their own estimate was for a 2-year transition, and they must have given this roadmap for a reason. The high-end iMac and Mac Pro will clearly be at the end of the 2 years, but I would expect the MBP16 to fall in the middle of the transition, i.e. in about a year's time.
I would be pleased if the first ASi Mac gets close to the AMD 5300M in GPU performance - that's pretty decent and very close to the entry-level MBP16. However, I don't think Apple will release a new MBP16 design that performs worse than the nearly one-year-old entry-level model.