Intel’s memory business
Intel (INTC) is expanding its data-centric business beyond data centers into the IoT (Internet of Things), 5G, FPGAs (field-programmable gate arrays), and memory. The company has resumed its memory chip business after 20 years. Even after such a long time, it still has the technical skills to make its chips competitive with those of other major memory chipmakers. It plans to spend ~$5.5 billion in 2017 and 2018 to convert its facility in Dalian, China, to manufacture non-volatile 3D NAND and 3D XPoint memory chips.
In January 2018, Intel ended IMFT (Intel Micron Flash Technologies), its 12-year partnership with the world’s fourth-largest NAND chipmaker, Micron Technology (MU). Intel’s memory chips were largely for self-consumption. Now, it is opening up to external customers as there seems to be strong interest in long-term supply agreements. There are also rumors that it may supply NAND wafers to China (FXI).
Earnings from the memory business
Intel’s revenue from the memory business rose 37% YoY (year-over-year) to $3.5 billion in 2017. Even though the business witnessed strong double-digit revenue growth, it continued to post losses. Its operating loss narrowed from $544 million in 2016 to $260 million in 2017, as the company did some major spending on research and development. It expects to make its memory business profitable in fiscal 2018.
In fiscal 4Q17, its memory revenue rose 9% YoY as demand outpaced supply. The revenue growth was largely driven by strong demand for data center FFT (fast Fourier transform) solutions. On a sequential basis, revenue growth was flat due to the delayed qualification of data center SSD (solid-state drive) volumes. The sales from this delay could appear in 1Q18. The signs of profits were visible in 4Q17, with the memory business reporting an operating profit of $31 billion, its first quarterly profit in two years. Next, we’ll see what Intel has planned for its memory business in 2018.