Samsung’s genius chip pivot could help reverse its worst-in-a-decade slump

Article by Jacob Bourne | Apr 10, 2023

The news: Samsung is getting walloped by the economic downturn.

On Friday, the company said its Q1 profits fell 95.8% YoY, the biggest dropoff since 2009, per The Wall Street Journal.
Most of the tech giant’s income comes from its semiconductor division, including memory chips, which have seen plummeting sales.
Topline performance was slightly buoyed by strong Galaxy S23 Ultra sales. However, faster depreciation compared with other smartphone brands might affect the S23’s staying power.

A big picture problem: Persistent inflation continues to wreak havoc on the tech industry, with Samsung the latest casualty.

Higher prices have severely curtailed consumer interest in devices that use memory chips, like PCs and smartphones.
Samsung isn’t alone—other memory chip makers, like Micron, SK Hynix, Western Digital, and Kioxia Holdings, are also dealing with a supply glut.
Samsung, an Apple component supplier and the world’s biggest smartphone and TV seller, is an example of the tech industry’s broader economic troubles.
The plight makes timing for domestic semiconductor expansions risky. In addition, high inflation might force the company to spend an extra $8 billion on material and labor costs to get its Texas factory built.

On the right track: Samsung’s profit nosedive can largely be attributed to external factors. Despite that, it’s making some sensible moves.

Chip demand for consumer electronics has tanked, but advanced chips for AI model training are a hot commodity.
Samsung has started mass production for its first-generation neural network processing unit (NPU) called Warboy, which could give it a competitive edge if the chip can rival Nvidia’s A100 and H100 on performance, cost, and efficiency.
On the device front, Samsung could double-down on efforts to bolster on-device AI capabilities for consumer PCs.
Samsung employees recently leaking confidential company data to ChatGPT underscores a growing need for enterprise hardware capable of running niche or proprietary generative AI models—demand that Samsung could help meet.