Alright degenerates, strap in. I think I found an algorithmic blindspot you could drive a semi through.
Most retail screeners, quant algos, and small shops all operate the same way: negative TTM (trailing twelve-month) net income = immediate trash bin. No exceptions.
If you run a basic screener right now, SanDisk ($SNDK) looks like a complete dumpster fire. TTM Net Income sits around -$1 Billion. Slapped with a "C-" profitability rating across most retail platforms. The bots look at those trailing twelve months and immediately pass.
But the algorithms are completely blind to what's actually happening under the hood. Those trailing metrics are mathematically poisoned by a single line item.
Here's the play.
---
### The $1.8 Billion Phantom Loss
Remember when Western Digital ($WDC) spun off SanDisk in February 2025? $SNDK started trading on Nasdaq February 24th and then absolutely cratered — bottomed out at $27.89 in April. Because of that technical drop in market cap immediately following the separation, accounting rules forced $SNDK to book a massive **$1.83 Billion non-cash goodwill impairment charge** in Q3 FY25 (quarter ended March 28, 2025).
That single entry resulted in a **$1.93 Billion GAAP Net Loss** for the quarter.
It's a non-cash paper loss. A literal accounting ghost. But because it sits directly in the trailing-twelve-month window, it mathematically wrecks the company's P/E ratio and net income metrics. To the bots, $SNDK is violently unprofitable.
In reality? They just posted **$3.03B in revenue** for Q2 FY26 (January 2026), absolutely smashing the $2.62B estimate with a **GAAP EPS of $5.15**. Back out the phantom loss, and they are currently printing real money at scale.
---
### The Inflection
So why is revenue up ~78% in just three quarters (climbing from $1.70B to $3.03B)?
Everyone's obsessed with the GPUs, but AI data centers require a completely insane amount of storage. This is the "storage supercycle." SanDisk is currently dominating with their **BiCS8 218-layer 3D NAND** technology — jointly developed with Kioxia.
They're making 2Tb QLC enterprise SSDs that are significantly cheaper to produce than the 176-layer or 232-layer alternatives from competitors. When hyperscalers are building billion-dollar data centers, density and cost-per-bit are everything.
Want the smoking gun? Look at the gross margin trajectory since the spin-off. This is what a structural shift looks like:
* **Q3 FY25 (Mar '25):** ~22.5%
* **Q4 FY25 (Jun '25):** 26.4%
* **Q1 FY26 (Sep '25):** 29.9%
* **Q2 FY26 (Jan '26):** **50.9%**
They nearly doubled their gross margins in 9 months. And management just guided for **65%–67% gross margins** next quarter. Seeing margin expansion like this in hardware is basically unheard of unless your ticker is NVDA.
---
### The Numbers Don't Lie
Here's what matters: free cash flow doesn't give a shit about goodwill impairments.
* **Free Cash Flow:** Went from deeply negative post-spinoff to approximately **$843 Million** in the most recent quarter.
* **Debt:** They've aggressively paid down the $2 Billion term loan from the spinoff to roughly **$603 Million** by Q2 FY26 end. Management said on the call they went "from $2 billion to $650 million." That's deleveraging at warp speed.
* **Capex:** Direct capex runs lean at ~$39–50M/quarter for back-end operations. The front-end fab spend goes through their Kioxia joint venture, which means they don't need to blow billions on new fabs to ramp BiCS8 production. Smart capital structure.
---
### Valuation
Because TTM P/E is mathematically meaningless right now thanks to the goodwill write-down, we have to look forward. Analyst consensus for FY2026 EPS is approximately **$35**, which puts the stock at roughly **~16.8x forward earnings** at current prices around $590.
Sixteen times forward earnings. For an AI hardware play growing revenue at 78% YoY with margins expanding toward 65%+.
For context, Micron ($MU) just posted a 74.4% gross margin in their February quarter — they still hold the gross margin crown. But SanDisk is closing the margin gap fast, and you're not paying anywhere near an incumbent premium for $SNDK at these multiples. The market hasn't fully repriced the convergence.
---
### The Bear Case
I try to be objective here, so let's talk risks. The stock has gone from $27.89 to ~$590, peaking at $777.60 in March. That's a 2,000%+ move from the lows. This is not a sleepy value play.
**1. "It's already priced in."**
Institutional and retail flows hit extreme levels around the January earnings beat. It got added to the S&P 500 in November 2025 (replacing Interpublic Group), forcing a wave of institutional buying. So yeah, smart money knows this name.
**2. The Citron Short Report.**
Andrew Left's outfit put out a short report in late February '26 arguing NAND is cyclical and $SNDK "sells a commodity." Stock dipped ~5%. Honestly, Citron betting against an AI hardware trend is basically a buy signal at this point.
**3. Cyclicality.**
Not a macro person, but NAND is notoriously more cyclical than DRAM. An oversupply in late 2027 could definitely crush these margins. This is the real structural risk.
But here's why the bears are early: There is a documented global NAND shortage happening right now. NAND prices have surged 80–90%+ and Phison's CEO confirmed all 2026 production is already sold out. On top of that, SanDisk is in a joint workstream with SK Hynix under the Open Compute Project to standardize **"High-Bandwidth Flash" (HBF)** for AI inference — a new memory tier sitting between HBM and SSDs targeting 1.6 TB/s read bandwidth. Samples expected H2 2026, products early 2027. This isn't a flash-in-the-pan demand spike; they are literally building the standard for the next hardware generation.
Until those trailing metrics roll off the calendar and the screeners update to show a wildly profitable 65%+ margin AI powerhouse, there is still an arbitrage on automated ignorance. Once TTM EPS turns violently positive in the next quarter or two, the quants are going to be forced to pile in.
---
**Position:** Loading up on shares because I sleep better at night, but also grabbed some June $700 Calls to play the next earnings print where that 65% margin guidance hits reality.
---
**TL;DR:** ALGO SCREENERS THINK SANDISK IS UNPROFITABLE BECAUSE OF A $1.8B ACCOUNTING GHOST FROM LAST YEAR. REALITY: REVENUE UP 78%, MARGINS JUST DOUBLED TO 51%, THEY'RE PRINTING $843M IN FREE CASH FLOW A QUARTER, AND ANALYST CONSENSUS HAS THEM AT $35 EPS — TRADING AT JUST 17x FORWARD. THE QUANTS HAVEN'T CAUGHT UP YET.
*This is not financial advice. I'm not a financial advisor. I just hate these guys.*