💾 SanDisk +3,000%: Why Memory Makers Are Buying Back Billions

TigerObserver
09-01 16:05
Reward Tiger-CoinsReward 500 Tiger-coins

👋 Hi, Tigers!

$SanDisk Corp.(SNDK)$ is up more than 3,000% over the past year. $Micron Technology(MU)$ has gained well over 200% YTD. Numbers like that are wild enough to make anyone wonder whether the market has gone too far.

But the more interesting story isn't the stock charts — it's what these companies are doing with the cash they're generating.

Record memory pricing has turned $Micron Technology(MU)$, $SanDisk Corp.(SNDK)$ and $SK hynix(SKHY)$ into cash machines. And their very different capital-allocation choices may offer some clues about how long this memory supercycle can actually last.

📊 By the Numbers

Ticker

Aug. 31 Close

52-Week Range

$Micron Technology(MU)$

$958.73

$114.25–$1,255.00

$SanDisk Corp.(SNDK)$

$1,566.70

$50.07–$2,354.39

$SK hynix(SKHY)$

$164.58

$124.80–$194.80

Market data as of Aug. 31, 2026.

🏗️ The Setup: The Fundamentals Really Did Explode

The memory rally isn't happening in a vacuum. Profitability has gone vertical.

  • $Micron Technology(MU)$: Non-GAAP gross margin hit a record 84.9% in fiscal Q3 2026, up from 74.9% the previous quarter and 39% a year earlier. DRAM revenue jumped 67% sequentially to $31.3B, driven largely by ASP growth in the low-60% range. Micron is guiding for roughly 86% gross margin in fiscal Q4, with results due Sept. 30.

  • $SanDisk Corp.(SNDK)$: Non-GAAP gross margin reached 84.6%, up from 78.4% the previous quarter and just 26.4% a year earlier.

  • $SK hynix(SKHY)$: Q2 revenue climbed 51% sequentially to ₩79.32T, while operating margin expanded to an extraordinary 76%.

That's the key difference between this rally and a purely speculative one: earnings and cash flow are exploding alongside the share prices.

But there's a catch.

At these profitability levels, investors aren't just paying for today's earnings anymore. They're implicitly betting that unusually strong memory pricing can last.

And that's where the companies' cash decisions get interesting.

💰 Where's All That Cash Going?

The three major players aren't following the same playbook.

Company

Capital Allocation

What's Happening

$SK hynix(SKHY)$

₩40T buyback & cancellation

Approved Aug. 19; equivalent to roughly 3.3% of shares outstanding. FCF-return target raised to over 50% for 2025–2027.

$Samsung Electronics Co., Ltd.(SSNLF)$

₩90–110T shareholder returns

Announced Aug. 21 — its largest shareholder-return program ever, including dividends and share repurchases.

$SanDisk Corp.(SNDK)$

$15.5B remaining authorization

Added $14B to its buyback authorization in August after repurchasing $4.5B in fiscal Q4.

$Micron Technology(MU)$

Reinvestment + buybacks

Generated a record $18.3B adjusted FCF in fiscal Q3 while continuing to invest heavily in new capacity.

The contrast is telling.

$Micron Technology(MU)$ is putting enormous amounts of money back into capacity.
$SanDisk Corp.(SNDK)$ is aggressively returning cash to shareholders.
$SK hynix(SKHY)$ is doing both — but is also making a huge bet that its own shares are undervalued.

And $Samsung Electronics Co., Ltd.(SSNLF)$ is combining massive shareholder returns with continued investment in its semiconductor business.

Same memory boom. Very different playbooks.

📈 Why Buybacks Matter — But Don't Mean the Top Is In

$SK hynix(SKHY)$'s ₩40T buyback is particularly interesting because it came after the stock had already fallen more than 50% in two months earlier this year. Management essentially used the selloff as an opportunity to return capital and cancel shares.

That's a meaningful signal: the company itself appears to believe the market was undervaluing its future cash generation. But investors shouldn't mistake buybacks for a guaranteed floor. A company can believe its stock is cheap and still watch it fall further.

The bigger question is whether the underlying earnings power survives.

🧠 The Bull Case: DRAM Still Looks Tight

There's a reason the companies are so confident.

According to TrendForce, DRAM's 2026 supply-demand sufficiency ratio is around -1% to -2%, indicating a genuine supply shortfall. The gap is expected to widen further in 2027. And new capacity isn't arriving overnight. New DRAM fabs require years to construct and ramp, meaning much of the additional output isn't expected to meaningfully hit the market until 2028.

Then there's HBM.

HBM is consuming an increasingly large share of the industry's manufacturing capacity, with TrendForce estimating it could account for roughly 22% of DRAM wafer input by the end of 2026, versus about 18% in 2025.

That matters because every wafer allocated to HBM for AI accelerators is effectively capacity that can't be used for conventional DRAM. So even with memory makers spending aggressively on new fabs, the supply response may lag demand for several years.

That's the bull case: AI demand remains strong, HBM keeps absorbing capacity, and conventional DRAM stays structurally tight.

⚠️ The Bear Case: NAND Could Be Telling a Different Story

Here's where the memory story gets more complicated.

DRAM ≠ NAND.

TrendForce expects the NAND market to move toward a positive supply-demand balance in 2H 2027, meaning the current shortage could eventually turn into surplus. That's a very different setup from DRAM.

And demand outside AI isn't exactly screaming higher: TrendForce expects smartphone production to decline 15–20% YoY and notebook shipments to fall around 10% in 2026.

There is, however, one wildcard: Agentic AI.

If AI workloads increasingly require huge amounts of high-performance storage — including SSDs used for things like KV-cache storage — that could absorb some of the NAND capacity that would otherwise create a surplus.

In other words, the NAND bear case isn't guaranteed. But the risk is becoming harder to ignore.

🔍 So… What Would Actually Signal a Top?

Don't just watch the stock charts. Watch the fundamentals underneath them.

  1. $Micron Technology(MU)$'s Sept. 30 earnings

The company is guiding toward roughly 86% gross margin in fiscal Q4. If it misses that target — particularly if pricing or margins deteriorate — the market could react aggressively because so much optimism is already priced in.

  1. NAND's 2027 supply balance

Does the expected surplus actually arrive?

If NAND supply starts growing faster than demand, pure-NAND exposure could become much less attractive. If AI-driven SSD demand absorbs the extra capacity, the cycle could last longer.

  1. Memory pricing

A sustained rollover in DRAM contract prices would be a much bigger warning sign than a one-week stock selloff.

The real red flag would be falling DRAM prices alongside slowing hyperscaler capex growth — that would suggest demand is finally failing to keep up with the industry's pricing power.

🧭 How to Actually Play This

Buybacks are a signal, not a floor. SK Hynix's massive repurchase followed a 50%+ drawdown. It helped restore confidence, but it didn't magically eliminate volatility.

DRAM and NAND deserve different treatment. The 2027 outlook increasingly favors DRAM-heavy exposure such as Micron and SK Hynix, while NAND faces a clearer potential supply problem.

Cheap forward P/Es aren't the whole story. A stock can trade at a seemingly low multiple because analysts expect extraordinary earnings to continue. The real valuation question isn't "Is it cheap?" — it's "How sustainable are these earnings?"

And finally, respect the volatility. These stocks have already experienced 15–20%+ moves within weeks in both directions. A strong fundamental story doesn't mean the ride will be smooth.

🐯 The Bottom Line

The memory boom is generating an almost absurd amount of cash — but where that cash goes may be just as important as how much is being generated.

Micron is reinvesting heavily in capacity. SanDisk is returning billions to shareholders. SK Hynix is buying back stock after a huge drawdown. Samsung is committing to its largest shareholder-return program ever.

That's not what a broken industry looks like. But it also doesn't mean the cycle can run forever.

DRAM still looks structurally tight. NAND could loosen by 2027. And at ~85% gross margins, there's very little room for earnings expectations to disappoint.

🪙 Discussion: Cash Machine or Cycle Top?

Memory chipmakers are posting ~85% gross margins and shoveling billions into buybacks — but DRAM and NAND are quietly starting to tell different stories for 2027.

Q1: $SK hynix(SKHY)$ bought back ₩40T in stock right after a 50%+ drawdown — confidence signal, or just damage control?

Q2: DRAM looks structurally tight into 2027, NAND doesn't. Are you rotating toward DRAM-heavy names (MU, SKHY), or staying diversified across the sector?

Q3 — Bonus: At today's margins, is there more room for memory stocks to run, or is the easy money already made? One sentence, make your case.

🐯 Drop your take below — sharp comments may catch our eye for Tiger Coins!

Investing is a long journey. Knowing when to move matters — and so does knowing when to recharge.

Markets may not slow down, but you can. Whether it’s a quick break between trading sessions or a long journey, staying comfortable helps you reset for what comes next. The new Night Trading Memory Foam Pillow features 360° ergonomic support and slow-rebound memory foam that gently conforms to your neck, plus a roll-up design that makes it easy to bring wherever you go.

https://laohu8.com/J/redeemGift?goodID=100540&type=delivery

Redeem the new Night Trading Memory Foam Pillow in Tiger Coins Mall, and stay supported for the journey ahead.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
Reward 500 Tiger-coinsDeadline to 09/11 16:05
The originator will select the best recovery and allocate Tiger coins before the reward ends
Reward-post

Comments

  • Jerry Lam
    09-01 18:48
    Jerry Lam
    我更偏向 “超级周期还没结束,但已经进入最考验判断力的阶段”。

    Q1:SK海力士₩40T回购,我更倾向是信心信号,不只是损害控制。 如果管理层只是想稳股价,完全没必要把股东回报目标直接提高到自由现金流的50%以上。更重要的是,它是在HBM订单和现金流都很强的背景下回购,说明公司自己也认为市场把未来盈利压得太低。不过回购不是底部保证,真正决定股价的还是后续价格和利润率。

    Q2:我会明显偏向DRAM/HBM,而不是平均分散整个存储行业。 目前MU和SK海力士的逻辑更清晰:AI服务器持续吃HBM,同时HBM又挤占传统DRAM产能,供给短期很难快速追上。NAND虽然也有企业SSD和AI存储需求,但2027年新增供给压力更值得警惕。

    Q3:我认为还有空间,但“轻松赚钱”的阶段已经过去。 毛利率到了80%以上,本身就意味着市场预期非常高,后面任何价格松动、Capex放缓或者客户削减长约,都可能引发很大回撤。

    一句话:我会继续押MU和SK海力士,但不会把85%的毛利率当成永久状态;真正的顶部不是股价涨太多,而是DRAM价格、长约和AI Capex开始一起掉头。

  • 苏36
    09-01 16:42
    苏36
    My view: the memory supercycle still has room to run, but the easy money is probably behind us. The key is separating DRAM from NAND. DRAM remains structurally tight because AI accelerators and HBM are consuming enormous wafer capacity, while new fabs take years to build and ramp. NAND, however, faces a clearer risk of supply catching up with demand in 2027.

    That makes me more comfortable with MU and SK hynix than pure NAND exposure. SK hynix’s ₩40T buyback is a strong signal that management remains confident in future cash generation, but buybacks are not a guaranteed price floor.

    At ~85% gross margins, expectations are already extremely high. I would watch DRAM contract prices, HBM demand, hyperscaler capex and gross margins rather than stock charts. If DRAM prices roll over while AI capex slows, the cycle could turn quickly. Until then, I remain bullish—but I would buy selectively on pullbacks rather than chase momentum.

    @TigerObserver [暗中观察]

Leave a comment
2