Have a look at the chart of Sandisk.
This company was once a sleepy maker of solid-state drives, memory cards and flash drives. You’re likely to have an old stick drive in a cupboard somewhere or one of their insertable memory chips in an old phone or digital camera.
At this time last year, it was trading at $38 per share after a spinoff from Western Digital. That gave it a market cap of $7.3 billion.
Jump ahead to today and that’s jumped to $265 billion on an incredible scramble to buy memory. In the first quarter, the earnings per share were $23.41 compared to $14.66 expected. Those quarterly earnings were 60% of the entire market cap just a year ago.
Shares are up another 4% today and are up nearly 50x from a year ago.
It’s one of the all-time great stock market moves and is just breathtaking to see. The monthly RSI is now at 99, which is something I don’t think I’ve ever seen before. On the daily RSI shown above, you can see it’s frequently jumped above 70 but the pullbacks have been moderate throughout the run.
It’s not like the company is soaking up all the AI capex boom either. Its revenue in Q1 was about $6 billion and compare that to the trillion dollars being spent on AI this year. That spending may continue until the end of the decade.
The big shift was the realization last year that inference was the latest AI bottleneck and memory was critical for it. In order to perform inference, it requires ultra-fast, low-latency storage retrieval.
At a conference last week, CEO David Goeckeler said:
And I was reading reports as late as, let’s say, late summer that said, oh, SanDisk is going to miss their numbers in December because pricing is going to be down. So it didn’t quite work out that way. I don’t think anybody could have anticipated the real intensity at which data center has come on.
On valuation, it’s not bad if you take Q1 numbers and extrapolate. That annualizes at $93/share so it’s trading at roughly 19x. The consensus though is $64 so that pushes it to a more-lofty 28x.
The thing is, the consensus has undersold this company for awhile and next year’s consensus is $173, which is roughly 10x today’s share price.
The problem is that the memory business has been around for a long time and it’s extremely cyclical with booms and busts. The sense is that we will get an over-build of memory like we always do and prices will come crashing down. If it doesn’t this party could continue from here
