What was changed
By 2015 AMD was down to roughly two dollars a share after years of losses. It had handed its fabs to GlobalFoundries, abandoned a mobile-chip bet, and was being priced for a terminal decline in x86. New CEO Lisa Su's answer was to stop competing for scraps and make one clean bet: return to high-performance CPUs and GPUs, where the margin was.
The plan was a multi-year roadmap. AMD slashed costs to curb losses, got its Radeon into Apple's MacBook Pro, and launched Ryzen in early 2017 as its first credible rival to Intel in years. By the third quarter of 2017, revenue had climbed 26% to $1.64bn and net income hit $71m, against a $400m-plus loss a year earlier. Su also said AMD would skip intermediate nodes and jump straight to 7nm.
The market's reward was uneven at first — a later guidance warning on cooling crypto-GPU demand knocked the stock nearly 14%. But the direction was set. AMD had turned from a company being written off to one dragging Intel into a real fight, and the share price had already risen more than fivefold from the prior February.
What it achieved
Q3 2017 revenue climbed 26% to $1.64B and net income hit $71M after a $400M-plus loss a year earlier.
Why it worked
AMD stopped spreading effort across low-value products and concentrated on high-performance compute.
Cutting costs limited losses while the new roadmap was still going to market.
Ryzen gave AMD real x86 performance at a lower price, so it finally had a product to lead with.
Committing to jump to 7nm showed it would deliver the roadmap rather than conserve cash.
What can be applied
When a chipmaker is outspent and outmuscled, the way back is to stop competing everywhere and aim a smaller portfolio at the highest-value customer.
Aftermath
AMD carried the momentum into 2019 with the first 7nm Ryzen and EPYC server chips, taking server share from Intel and pushing its market value to record highs.
FOLLOW THE EVIDENCE