BLZE · Backblaze, Inc. · Stock research

Current Thesis

Backblaze sells the cold and warm object-storage tier that sits underneath the neocloud GPU buildout, and on 2026-06-23 that thesis got its anchor customer. The company signed a five-year, roughly $335M multi-exabyte agreement with CoreWeave in which B2 Neo — the white-label build of B2 Cloud Storage launched 2026-02-23 at 1 Tb/s — supplies the dense HDD tiers inside CoreWeave AI Object Storage, holding training datasets off the expensive NVMe pools that run active training and inference. The tape re-rated it: a +14.1% gap on the announcement carried the stock from the high-$7s in early June to a fresh 52-week high near $18 ($17.98) and a ~$1.07B cap by mid-July. Sell-side is still catching up in real time — Craig-Hallum to Buy $16 (from $6.50), Needham to $14 (from $8.50), B. Riley to $16 (from $7), and now Citizens raising to $16 (from $8, Market Outperform) on 2026-07-23. The narrative is ACCELERATING and maturing at the same time: at ~$18 the stock trades above the whole $14–16 target band, and the ~2026-08-06 Q2 print is the first hard read on whether CoreWeave is a repeatable template or a single whale.

Bullish and bearish views on Backblaze, Inc.

The model's bull view on Backblaze, Inc. (BLZE), in brief: Five-year, ~$335M CoreWeave contract announced 2026-06-23 — a multi-exabyte, HDD-tier storage deal that is the first at-scale proof a Tier-1 neocloud will white-label B2 rather than build its own object store; Craig-Hallum flagged it as "very material" to 2027 estimates. The bear view: Price ran roughly 130% in about five weeks (high-$7s in early June to ~$18 by 2026-07-10) and sits above the entire $14–16 target band and near 2x InvestingPro's ~$9.90 fair value; the CoreWeave catalyst is announced and largely in the quote. Both cases follow in full.

Bull Case

• Five-year, ~$335M CoreWeave contract announced 2026-06-23 — a multi-exabyte, HDD-tier storage deal that is the first at-scale proof a Tier-1 neocloud will white-label B2 rather than build its own object store; Craig-Hallum flagged it as "very material" to 2027 estimates.
• The deal validates B2 Neo (launched 2026-02-23, 1 Tb/s throughput), purpose-built for GPU-centric operators; one signed anchor lowers the go-to-market friction for the next Crusoe / Nebius / Lambda-class prospect.
• Q1 2026 (reported 2026-05-05) showed the pivot working before CoreWeave landed: B2 Cloud Storage +24% YoY to $22.4M, AI customers +76% YoY and more than a third of new bookings, gross margin 61% (from 56%), adjusted EPS $0.04 against a breakeven expectation; FY26 guide lifted to $161.5–163.5M revenue and 23–25% EBITDA margin.
• Sell-side re-rated hard against the pre-deal $8–11 band — Craig-Hallum $16, Needham $14, B. Riley $16, and Citizens to $16 on 2026-07-23 — a cluster still moving up rather than a single stale call.
• Eight consecutive EPS beats into the 2026-08-06 print, with CoreWeave revenue incremental to the standing FY26 guide because the contract was signed after that guide was set.
• Third parties size the neocloud-storage TAM near ~$35B in 2026 scaling toward ~$236B by 2031 (~46% CAGR), and BLZE is now the reference white-label vendor named in that lane.

Bear Case

• Price ran roughly 130% in about five weeks (high-$7s in early June to ~$18 by 2026-07-10) and sits above the entire $14–16 target band and near 2x InvestingPro's ~$9.90 fair value; the CoreWeave catalyst is announced and largely in the quote.
• The ~2026-08-06 Q2 print is a binary into a vertical move — consensus sits around EPS −$0.01 on ~$39.9M revenue, inside the company's own $39.8–40.2M guide, so the beat has to be large and the forward commentary strong to defend the multiple.
• $335M is a five-year figure (~$67M/yr at a flat run-rate) layered onto a business guiding ~$162M for FY26 and recognized over time; it is material but not a next-quarter step-change, so the tape is paying today for cash flows that arrive slowly.
• Single-customer concentration runs both ways: one whale de-risks the story, yet the next several quarters lean on landing a second neocloud to prove B2 Neo is a platform rather than a bespoke CoreWeave build.
• The blended top line (+12% YoY) is diluted by the melting Computer Backup legacy segment, so a casual read of "total revenue" understates the AI line and a skeptical read questions the durability of the rest.
• Jim Cramer publicly flagged the valuation cohort on 2026-06-04 — a marker that mainstream coverage has already noticed the move, which raises saturation risk on any further headline-driven leg.

Setup & Price Structure

The structure is a post-catalyst breakout that has extended past its confirmation. The 2026-06-23 gap took the stock from the high-$7s to a 52-week high of $17.98 on heavy volume, and it held near ~$18 into mid-July rather than filling the gap — constructive continuation, but with no pullback to work against. Support steps down in tiers: the first shelf sits in the $14–16 zone that now overlaps the analyst-PT cluster, then the pre-breakout $8.32 shelf / early-May highs far below. Trading above every published target with fair-value estimates near half the quote means the risk/reward on a fresh entry at ~$18 is asymmetric to the downside unless a second neocloud win or a large Q2 beat resets expectations higher. Short interest is thin (~2.4% of float, −23.8% MoM), so this is not a squeeze mechanic — the moves are fundamental- and flow-driven, and a disappointment has no forced-cover bid to cushion it. The cleaner entries are a pullback that holds the $14–16 breakout base on volume, or a post-print continuation once the binary clears. Buying the vertical leg into the earnings window is the beginner trap here: peak coverage, price above the band, and a binary event inside two weeks.

Catalyst Calendar (next 30 days)

• Unscheduled, any day: a second neocloud customer win (Crusoe / Nebius / Lambda-class) would be the single most bullish off-calendar surprise and the cleanest proof of a repeatable platform.
• Mid-to-late August (est.): post-earnings analyst revisions; the $14–16 cluster either re-rates higher on a beat or compresses toward fair-value estimates on a miss.

Elapsed catalysts

• ~2026-08-06 (confirmed): Q2 2026 earnings. The dominant near-term catalyst and the first data point on whether CoreWeave-style white-label deals repeat. Guide is $39.8–40.2M revenue; watch the B2 Cloud Storage YoY line and any FY26 revision, not the blended top line. (passed 3d ago)
• ~2026-08-01 to 08-03: binary-risk blackout for fresh entries opens (three trading days ahead of the print); avoid initiating into the print. (passed 8d ago)

What Would Change Our Mind

The bullish read breaks on a weekly close below $13, which forfeits the post-CoreWeave breakout leg and drops the stock back under the lower edge of the $14–16 analyst-PT cluster. A secondary invalidation fires if the ~2026-08-06 Q2 print shows B2 Cloud Storage growth decelerating under 20% YoY or an FY26 guide cut — either would undercut the AI-storage acceleration that justifies a premium to fair value. The theme itself would flip to saturated if mainstream coverage peaks with no fresh neocloud win to extend the story, at which point the setup becomes a pass rather than a pullback buy. Conversely, a second at-scale neocloud contract or a large clean Q2 beat with a raised guide would re-arm the narrative and reset the entry zone higher.

Correlation Notes

BLZE now trades as a derivative of the neocloud capex cycle: its fate is tethered to CoreWeave's buildout continuing and to the broader GPU-cloud complex (CRWV, NBIS) sustaining object-storage demand. It also carries beta to the small-cap AI momentum cohort — the Russell 2000 AI names that ran alongside the Micron memory rally — so a risk-off rotation out of speculative small caps would pressure it independent of company news. Within storage it is not a proxy for the NVMe-tier vendors (PSTG, NTAP); B2 Neo competes on cold/warm HDD economics, and the bear scenario for the whole white-label thesis is hyperscalers and neoclouds choosing in-house builds over an outside vendor. The single-customer concentration gives it idiosyncratic decoupling potential in both directions — a second-customer win detaches it upward from the cohort, while a CoreWeave-specific stumble would hit it harder than peers.