ARM announces next-generation CPUs and GPUs focused on efficient machine learning; ARM Cortex-A75 is new flagship mobile chip, now with 22% better performance
ML plus AI, AR, and VR — ARM is taking on the full set of trendy initialisms — On the eve of Computex …
Context & Ripple Effects
ARM's Computex-eve flagship reveal has become an annual ritual: last year's Cortex-A73 and Mali-G71 were pitched at powering mobile VR in 2017, and this cycle the marketing frame rotates to efficient machine learning, with the Cortex-A75 claiming 22% better performance over its predecessor.
The pattern holds across the corpus — the following year brings the Cortex-A76 with 35% better performance reframed around VR and AI, then the Cortex-A77 and Mali-G77 with a dedicated ML processor, and eventually the Cortex-A78 and Mali-G78 designs destined for 2021 flagship Android phones. The A75 is one link in a steady per-generation performance ladder where the AI theme deepens each year.
First-order effects
- Phone makers building 2018 flagships on ARM's big-core designs get a 22% faster CPU core whose GPUs and system IP are explicitly tuned for machine-learning workloads.
- ARM shifts its pitch from last year's VR framing to efficient on-device ML, signaling to licensees that neural workloads are now a first-class design target alongside graphics.
Second-order effects
- Rival silicon vendors must match both the performance step-up and the ML story in their own flagship cores, since ARM's annual Computex cadence resets the mobile benchmark conversation every spring.
- As ML moves into the CPU/GPU package itself, the differentiation battle among Android chipmakers migrates from raw clock speeds toward per-watt efficiency on AI workloads — exactly the trade-off ARM's A72 messaging framed as triple the performance or matching it at 75% less power.
Third-order effects
- If the cadence holds — as the A76, A77, and A78 announcements suggest it does — on-device machine learning becomes a baseline expectation of every flagship mobile SoC rather than a differentiator, with ARM's licensing model spreading that capability across the whole Android ecosystem at once.
- Mobile silicon consolidates into an annual platform rhythm owned by ARM's reference designs, narrowing the space in which individual chipmakers can claim architectural independence.
The trend: ARM's yearly Computex flagship reveals are turning on-device machine learning from a marketing theme into a standard feature of every successive mobile CPU generation, compounding per-watt gains along the way.