A look at AMD's 25x20 project, which it started in 2014 to improve overall “Performance Efficiency” 25x by 2020, and what its success means in the long run
One of the stories bubbling away in the background of the industry is the AMD self-imposed ‘25x20’ goal. Tweets: @underfox3 and @hpc_guru . Thanks: @iancutress Tweets: Underfox / @underfox3 : In an evident death of Moore's law and all related challenges, it would be foolish to make such a bold proposition for the years to come. But I must agree that the next years will be really fun. 🦊 https://www.anandtech.com/... https://twitter.com/... Hpc Guru / @hpc_guru : .@AMD succeeds in its 25x20 self-imposed goal from 2014 25x20: achieving 25x performance efficiency by the year 2020 By @IanCutress in @anandtech https://www.anandtech.com/... https://twitter.com/... Thanks: @iancutress
Context & Ripple Effects
AnandTech's deep dive closes the book on 25x20, the self-imposed target AMD set in 2014 to lift performance-per-watt 25-fold by 2020 — a bet placed when the company was struggling and Moore's-law scaling was already slowing. Hitting it validated the architectural route to efficiency that underpinned Zen, EPYC and Radeon's competitive comeback.
That efficiency base is the throughline to everything in AMD's later arc: the 71% YoY revenue growth in Q1 2022, Lisa Su's forecast of 35%+ average annual growth with 80% AI data-center expansion, and CTO Mark Papermaster positioning AMD as first customer for TSMC's 2nm node. 25x20 is the proof point that AMD can hit multi-year engineering commitments — the credibility its AI-era promises now trade on.
First-order effects
- The confirmed 25x gain gives AMD a quantified efficiency credential at exactly the moment data-center buyers weight perf-per-watt against Nvidia's incumbency — and it validates the multi-year goal-setting discipline behind Su's current three-to-five-year growth targets.
Second-order effects
- Rivals are pushed to compete on efficiency roadmaps rather than node names alone: with Moore's law slowing, as the 25x20 coverage itself frames it, Intel and Nvidia must answer architectural gains with their own, while AMD's process lead — first in line for TSMC 2nm — compounds the advantage.
Third-order effects
- If efficiency targets become the industry's public scoreboard, competition shifts from who ships the smallest transistor to who executes longest-horizon engineering plans — favoring firms with credible roadmaps like AMD's, and making execution risk, not process availability, the key differentiator.
The trend: As Moore's law decelerates, chip competition is re-centering on architectural performance-efficiency gains and long-horizon engineering commitments, with AMD's 25x20 success as an early proof point.