3 September 2026

SanDisk was the best-performing stock in the S&P 500 in the first half of 2026. Up 726%.

Then it lost more than half its value in about a month.

It’s now trying to find its feet. It’s on my watchlist. I’m not in it yet. This post is about what I need to see before I am.

What SanDisk Does

  • Makes NAND flash memory. The chips that store data.
  • The big buyer now is AI data centres. They need huge amounts of fast storage.
  • Spun out of Western Digital in February 2025.

The Run and the Crash

  • Peaked around 2,354 in late June.
  • By the end of July it closed near 1,096. Down more than 50% in barely a month.
  • Nothing went wrong at the company. It was profit-taking and a sector-wide memory sell-off after a massive run.
Lesson: a great business can still drop 50% when the trade gets crowded. The chart and the company are two different things.

What’s Happened Since

Results, 5 August:

  • Quarter revenue $8.97bn. Full year $20.2bn, up from $7.4bn the year before.
  • Data centre revenue $2.98bn, up 103% on the previous quarter.
  • Five new long-term deals, with $94bn in minimum revenue.
  • The catch: next quarter’s guidance came in slightly under forecasts. The stock fell 11% the next day.

Investor Day, 13 August:

  • Long-term targets of around 80% gross margin and 50% free cash flow margin.
  • JPMorgan upgraded to Overweight with a $2,250 target.

The bounce and the rejection:

  • From the July low, price bounced to the 1,828.40 area in mid-August. That’s around +67%.
  • Sellers stepped in. Rising Treasury yields hit it on 18 August. On 24 August a report that Apple may buy memory chips from China knocked it 9%, down to about 1,458.
  • Closed on 3 September at 1,554.99.

The Levels I’m Watching

All prices are actual prices. Distances are from the 3 September close.

Level Price From 1,554.99 Why it matters
All-time high 2,353.58 +51.4% The June peak. Where a lot of people bought at the top and will want out.
Resistance 1,828.40 +17.6% Where the August bounce stalled. A break and hold above here says the trend is back on.
Trigger zone ~1,600 +2.9% I want price back above here, on volume, before I put a starter on.
Support ~1,458 -6.2% The 24 August low. Lose this and the base is broken.
Crash low ~1,096 -29.5% The end-of-July low. The floor for the whole move.

What I’d Do

  • Starter: small position if price pushes back above ~1,600 with volume.
  • First target: take some off into 1,828.40. That’s where it failed last time.
  • Add: only if 1,828.40 breaks and holds as support on a retest.
  • Risk: stop set so a loss is 1% of the account. No more.
  • No trade: if it loses ~1,458, I stay out and wait.

Good Points

  • Revenue is still growing fast. The business hasn’t cracked. The share price has.
  • Long-term deals give years of visibility on sales.
  • Memory is a key part of the AI build-out, and demand is still rising.
  • Price has already had a 50% clear-out. A lot of the froth is gone.

Bad Points

  • Memory is a boom-and-bust business. Profits swing with chip prices.
  • Chinese competition. CXMT and YMTC are growing, and the Apple report shows how quickly that hits the price.
  • Rising bond yields hurt stocks priced on future growth.
  • It moves hard. 9% to 14% days have been common. Size small.
Warning: this stock can drop 10% in a day on a headline. If you can’t sit through that, it’s not for you.