Apple debuts the M2 chip, a CPU 18% faster than the M1, with 20B transistors, 100Gbps of memory bandwidth, 50% more than the M1, and 24GB of unified memory
Context & Ripple Effects
Apple's base-tier silicon refresh lands seven months after the M1 Ultra fused two M1 Max dies into one chip, completing the top of the first-generation Apple Silicon ladder. The M2 is deliberately incremental where the Ultra was dramatic: an 18% faster CPU and a 10-core GPU over the M1, with the real gains in memory — 100GB/s of bandwidth and support for 24GB of unified memory.
That positioning matters because the mid-tier chips Apple shipped last fall still outrank it: the M1 Pro and M1 Max carry 70% faster CPUs and 32–64GB memory ceilings, so the M2 slots in below them rather than replacing them. AnandTech's teardown confirms the transistor budget grew only from 16B to 20B — this is a refinement node, not a leap.
First-order effects
- Entry-level Mac buyers get a modest CPU bump but a materially higher memory ceiling (24GB vs. 16GB) and 50% more bandwidth, shifting what 'base model' can handle for video and multitasking workloads.
Second-order effects
- The M2 leaves the M1 Pro/Max tier unchallenged, pressuring Apple to fill the gap quickly — which it does with the M2 Pro and M2 Max seven months later, doubling the M2's transistor count and pushing bandwidth to 200GB/s.
Third-order effects
- If the pattern holds, Apple's ladder becomes an annualized cadence where each generation re-climbs from base to Ultra via die scaling — the M4 Max's 2.2x-over-M1-Max speeds show the compounding — and unified memory bandwidth, not raw CPU clocks, becomes the spec competitors are measured against.
The trend: Apple Silicon is settling into a yearly, tiered release rhythm in which memory bandwidth and capacity, rather than CPU speed alone, define each rung of the Mac performance ladder.