Tape Stocks' AI Upside

🌾 Welcome to StableBread’s Newsletter!

How do companies store AI data that has to be retained for years but rarely accessed?

The only real answer is magnetic tape. Tape costs roughly 1/5 of what hard disk drives (HDDs) cost to acquire and 1/20 of what they cost to run over a decade once you factor in media, power, cooling, and floor space.

The modern enterprise tape standard is Linear Tape-Open (LTO), an open format jointly developed by IBM, HPE, and Quantum. The latest generation (LTO-10) holds 30 TB native per cartridge, with a 40 TB version slated for early 2026.

We discussed the magnetic tape industry in a previous write-up, where we walked through the macro setup, the LTO roadmap, the alternative archive technologies, and the broader public-market universe.

That work narrowed the universe of investable pure-play tape OEMs in the U.S. to two names: Quantum (QMCO) and Qualstar (QBAK).

The reason both names matter right now is that demand for tape inflected in 2025-2026 across three fronts:

When flash and disk get more expensive and constrained, customers offload cold data to tape (the tier below) and free up the upper tiers for the workloads that actually need them. That's the cycle.

This write-up focuses on the best pure-play stock that is shipping into that demand curve at scale: Quantum (QMCO).

Let's begin by ruling out Qualstar (QBAK) despite it being one of the other pure-play magnetic tape stocks.

QBAK is an OTC-listed micro-cap founded in 1984 that assembles tape libraries (Q-Series, plus the new Q1000+) using drives bought from HPE, IBM, or Quantum.

The company also has a separate dying power-supply business (N2Power) that fell 64% y/y in FY2025 on the loss of a single large customer.

FY2025 results were mixed. Revenue was $6.6M, down 18% y/y, but the headline drop came almost entirely from the N2Power collapse. The data-storage segment, which is the slice that should benefit from the tape macro, grew only 9% to $5.5M.

The company finished the year with $70K of net income, $2.8M of cash plus securities, no debt, and ~$32M of federal NOLs sitting unused.

QBAK's Q3 2025 (Sep 2025) quarter was strong (+49% y/y, $0.15 EPS, adjusted EBITDA flipped positive to +$69K). Q4 2025 (Dec 2025) reversed it—revenue down 27% y/y, gross margin compressed from 33.6% to 20.3%, negative EPS ($0.39).

One up quarter, one bad quarter, no clear trend yet on tape demand actually showing up in QBAK's reported numbers at scale.

Beyond the financials, QBAK is a less direct tape-macro play than QMCO:

To QBAK's credit, there are positive signals worth noting:

None of this changes the core scale gap vs. QMCO, but it means things are happening underneath the headline numbers.

At $3.78 with 4.14M post-split shares, QBAK is a ~$16M market cap on $6.6M of revenue. That's ~2.4x sales. For a debt-free micro-cap whose data-storage segment ran a small operating profit in FY2025, that's a reasonable valuation.

The issue isn't that QBAK is expensive. It's that QBAK is small, sub-scale, and not yet visibly capturing the AI-driven tape demand the way QMCO is.

So I view QBAK as a smaller, lower-leverage version of the tape thesis with optional corporate-action upside. The stock has paths to upside if a real M&A transaction closes, if a Nasdaq uplisting completes, or if magnetic tape sales climb materially through the Q1000+ ramp.

But QMCO is the play actually shipping into the macro right now, with the operating inflection already visible in the financials. That's what the rest of this write-up walks through.

Quantum (QMCO) sells software and hardware that helps companies store, manage, and protect "unstructured data," the kind of data that doesn't fit cleanly into a database.

The data has to be kept for years (sometimes decades), and most of it's rarely accessed once it's archived.

Quantum handles the entire lifecycle of that data, from high-speed ingest (when video is being edited or sensor data is being processed) all the way to long-term archive (where it sits cheaply for decades).

Four product lines do that work:

Revenue splits three ways across those four product lines (FY2025):

A more useful breakdown comes from the 10-K's solution-level revenue table:

Primary storage (Myriad and StorNext) grew 9% in FY2025 (ending March 2025) as the AI-driven storage tailwind started showing up.

It then softened. Through the nine months ending December 2025, primary storage was down ~23% y/y. CEO Hugues Meyrath confirmed on the Q3 call that the year started "very, very slow" before Q3 turned.

Secondary storage, where tape libraries live, was the big drag in FY2025, down 27% as hyperscale customers cut their orders by $20M y/y. That line flipped in Q3 FY2026 (more on this below).

Geographically, FY2025 revenue split Americas 52%, EMEA 35%, APAC 10%.

APAC moved to new exclusive distributors after Quantum terminated Quantum Storage Asia (QSA) in Q2 FY2026. APAC revenue more than doubled q/q under the new model.

Top five customers were 21% of revenue, with no single customer above 10%. So no single customer can break the company by leaving, but hyperscaler concentration still moves quarterly numbers meaningfully.

Quantum has ~635 employees globally, headquartered in Centennial, Colorado, with customers in over 100 countries.

The current management team is essentially new since mid-2025:

The company also moved its headquarters from San Jose, California to Centennial, Colorado effective August 1, 2025 as part of the broader cost-out program.

The Q3 FY2026 inflection happened under this newly assembled team.

The quickest way to understand QMCO today is that the business has been contracting for almost three years and just inflected, which is the core reason for this write-up.

Through late 2025, the business was deteriorating and the credibility issues were piling up.

Revenue declined 12% to $274.1M. GAAP gross margin held flat at 40.1%, but the operating loss widened ~45% to -$41.7M.

GAAP net loss was -$115.1M, with ~$48M of that non-cash from warrant revaluations and debt-extinguishment marks.

The auditor also issued a going-concern qualification on the FY2025 10-K and flagged material weaknesses in internal controls around revenue recognition, sales-order entry, and inventory at third-party locations.

On top of that, the company restated Q3 FY2025 (the December 2024 quarter) in this same 10-K, following an earlier restatement of FY2022/FY2023.

Q3 FY2026 product revenue is when the inflection becomes visible:

Three points stand out:

The Q3 FY2026 inflection isn't just about revenue.

The operating loss improved in tandem: ($13.2M) Q4 FY2025 → ($12.6M) Q1 FY2026 → ($8.2M) Q2 FY2026 → ($1.2M) Q3 FY2026, a ~$12M improvement in three quarters.

Adjusted EBITDA also flipped to +$2.9M in Q3 FY2026, up from +$0.5M in Q2 FY2026.

That improvement is partly revenue (tape inflection) and partly cost-out. Total operating expenses ran ~$41M per quarter in early FY2024 and dropped to ~$30M in Q3 FY2026, roughly $11M per quarter, or ~$44M of annualized run-rate savings.

The savings came across all three operating-expense lines: sales and marketing (-$3.4M from peak), G&A (-$4.7M from peak), and R&D (-$3.0M from peak).

Restructuring charges totaled $4.1M in FY2025 alone, reflecting the active workforce reductions, and SBC also dropped from $4.7M (FY2024) to $2.8M (FY2025).

This matters because the operating-loss inflection isn't just demand-driven. Even if Q4 FY2026 revenue comes in flat to Q3, the loss should stay near breakeven on the new cost base.

The one offset is gross margin, which is compressing slightly rather than holding flat or growing above 40%.

Q3 FY2026 GAAP gross margin was 38.8%, down from 40.6% in Q3 FY2025, with Q1 FY2026 at 35.3% and Q2 FY2026 at 37.6%.

On the Q3 call, Hugues would not even guide gross margin for Q4, saying pricing is "changing on the docks" of suppliers. He still affirms 40% as the long-term target post-component-normalization but explicitly framed it as a 2-3 quarter problem to work through.

The OpEx-out story is real, but the gross margin story is going the other way for now, which delays the EBITDA scaling.

What's driving the demand side ties back to the broader storage market. CEO Hugues Meyrath, on the Q3 FY2026 call, framed the inflection as a supply-chain spillover from the AI buildout:

Marc Staimer of Dragon Slayer Consulting made the macro point cleanly in Quantum's April 16, 2026 NAB-2026 release:

Two specific customer wins drive most of the Q3 visibility and the backlog: