Both the M1 and this upcoming processor leverage AVX 512 based vector operands to heavily skew their results in Geek Bench and other applications. In truth, neither are a leap of performance in workflows because very few applications leverage AVX 512 based vector operands on either the ARM or X86 architectures.
What's going to be a game changer is Xilinx merger with AMD for now the Zen 4 is fully incorporating its own Neural Engine, ML Engine, FPGAs for encode/decode large raw image/video formats for scrubbing and processing in real-time.
Chiplets with unified memory access via Infinity Fabric 3.0 are coming to Zen 4 Fall 2021. Anything either Intel and Apple are producing will take a back seat to these designs.
Overall, three companies are pushing the envelope forward for targeted compute and general process computing.
ARM will never be the champion of general process computing and thus the market king. It's just reality. Apple has no interest in dominating the market in computers. Nor does it have an inkling of interest in the HPC/Data Center markets.
AMD is the new leader in those and will expand over Intel leaving these Intel processors with a shrinking market space. Intel has a long way to go before its Jim Keller led designs make a dent, if at all, in the x86 markets.