Storj

What Storj actually is

Storj is decentralized, S3-compatible object storage. Your files are encrypted client-side, split with erasure coding into fragments, and spread across a network of independently operated storage nodes in over 100 countries; when you read an object back, the pieces are reassembled and decrypted. The pitch is a storage cloud without a hyperscaler's data centers or prices, with no single point of failure and no operator able to read your data. Storj emerged in 2014 from founder Shawn Wilkinson's idea of monetizing idle hard drive space, is run by Storj Labs out of Atlanta, and has been a production object store under the Tardigrade brand since 2019. The company has raised $35.4 million to date and employs a remote team of 60-plus people across 15 countries.

Two pieces of corporate history matter for buyers. In October 2024, Colby Winegar moved from CRO to CEO, with Ben Golub becoming executive chairman. In October 2025, Storj was acquired by Inveniam Capital Partners, a fintech focused on private-market asset data; Storj continues as a subsidiary with its team and customer relationships intact, Golub joined Inveniam's board, and the STORJ token — the cryptocurrency that pays node operators — gets integrated into Inveniam's ecosystem. Acquisition prices were not disclosed. We flag the ownership change because token-driven infrastructure businesses carry strategy risk that ordinary SaaS vendors do not.

Technically the product is straightforward to use. An S3-compatible gateway means existing S3 tooling works against Storj with a changed endpoint and keys. Object Mount, a desktop client shipped with every account, streams cloud files as if they were local and also mounts third-party S3 backends (AWS, GCP, Azure, Backblaze). The codebase is open source (the storj/storj repository), which matters for a storage vendor: you can inspect the encryption and erasure-coding implementation instead of trusting a white paper.

Pricing: what you really pay

Pricing changed substantially in 2026, so older articles — including the "$4/TB" figure that still circulates in comparisons — are out of date. Since July 1, 2026 Storj runs a two-tier simplified model. Standard is $7/TB per month for storage plus $7/TB egress, distributed across the global node network. Advanced is $10/TB per month for storage with $7/TB egress, with data held in U.S. SOC 2 Type 2 compliant data centers for compliance and residency needs. Storj states all tiers are designed for 11 nines of durability (99.999999999%) through erasure coding. Egress on Standard actually got cheaper in the transition — from $20/TB to $7/TB.

The small print is where budgets get surprised. There is a $5 minimum monthly fee on active accounts, and it is a floor, not a credit. Objects are billed at a 50KB minimum size. Data deleted before 30 days is still billed for the full month as an early deletion fee (Object Mount users excepted). Storage is measured continuously and rounded up per gigabyte per tier, storage and egress counted separately. The only free offering is a 30-day trial with 25GB of storage, two Object Mount licenses, and no credit card required — there is no permanent free tier, and our database's old "free 25GB" framing was wrong in exactly that way. Object Mount licenses beyond the included two run $29/license per month (each including a 500GB storage credit, with a $50 monthly storage minimum attached to that credit) or $39/license per month for mounting any third-party S3 storage.

The transition terms deserve their own warning. Existing customers had to accept the new pricing or opt out by July 30, 2026 — and opting out meant your account would be frozen and your data deleted within 45 days. EU customers had to opt in explicitly or face the same. We understand pricing migrations happen, but a storage vendor putting a 45-day data-deletion clock on a pricing decision is a strong-arm move, and it is part of why our trust rating here is cautious.

Performance & infrastructure

There is no fixed infrastructure inventory to compare, which is the point and the caveat. Storj's own description: tens of thousands of storage locations across more than 100 countries, orchestrated as one network, with node operators paid in STORJ token for storage (recent published rates: $1.35 per TB-month) and egress ($1.00 per TB). Uploads are erasure-coded and parallelized across nodes, which makes throughput scale with the number of fragments in flight; reads reconstruct from any sufficient subset, so node churn does not lose data. The Advanced tier keeps data in vetted U.S. data centers rather than the open network, which is the option we would pick for regulated workloads.

In practice, performance is workload-shaped. Large sequential objects — backups, archives, video, dataset storage — are where the architecture performs well and where the economics are strongest. High volumes of tiny objects, heavy metadata churn, or latency-sensitive hot paths are where a centralized store in one region will beat a globally distributed network most days, and the 50KB billing minimum quietly taxes small-file workloads. We have not run our own benchmark suite against Storj, so we will not quote throughput numbers; for the providers we have instrumented, see benchmarks.

Reliability design is the strongest engineering claim here: 11 nines of durability target, no single point of failure, end-to-end encryption by default, and a network whose operators are economically incentivized to stay online. The honest counterweight is that durability design is not the same as availability SLA — Storj publishes no uptime SLA that we could find for the storage service, and a decentralized network introduces more moving parts than one region of S3.

Where it falls short

Pricing churn, twice in one year: a November 2025 change was announced, then superseded by the July 2026 simplified model, with hard migration deadlines and data deletion as the stick. Small print stacks up — $5 monthly floor, 50KB object minimum, 30-day minimum retention, per-GB rounding — and while each rule is documented (Storj's docs at storj.dev are genuinely good), the combination means your invoice rarely matches back-of-envelope math. Compare that with the flat, boring billing of a Backblaze B2 or an S3 lifecycle policy.

Then there is the decentralization trade-off itself. Variable performance, and a dependency on a community of node operators you do not vet, is not what everyone wants from their backup target. The STORJ token economics are the second-order version of that risk: operator payouts, token value, and now Inveniam's strategic priorities all shape whether the network stays large and cheap. The new owner says all customer relationships remain intact and the roadmap accelerates; corporate acquirers always say that, and the next two years will show it.

Finally, maturity gaps: no permanent free tier, an ecosystem still thinner than AWS's (fewer third-party integrations, smaller community), and support tiers aimed at sales-led accounts rather than self-serve help. For a category where the incumbent is Amazon, "less mature ecosystem" is not a small thing.

Who should pick this

Our pick for bulk, write-mostly data where per-terabyte cost dominates: backups and archives, media and film workflows (a segment Storj actively targets), dataset storage for AI and science, and any workload where you want S3 semantics without hyperscaler egress bills. If you care about data sovereignty, the Advanced tier's U.S.-pinned placement is a clean answer, and privacy-minded teams get real value from client-side encryption being the default, not an option.

We would not pick it for hot, latency-critical application storage, heavy small-object workloads, or organizations that need a contractual uptime SLA and a support desk with escalation paths. In those cases a regional S3-class store or a plain-vanilla provider in our object storage category is the calmer choice. And before committing a large archive to Storj under the current ownership, we would get the retention and exit terms in writing.

Key takeaways

• Storj (Atlanta, emerged 2014) is decentralized, S3-compatible object storage: erasure-coded fragments on independent nodes in 100+ countries, client-side encryption by default.
• Current pricing since July 2026: Standard $7/TB/month storage and $7/TB egress; Advanced $10/TB/month with U.S. SOC 2 Type 2 placement; $5 monthly minimum fee, 50KB object minimum, 30-day minimum storage duration.
• No permanent free tier — just a 30-day trial with 25GB and two Object Mount licenses; the commonly quoted "$4/TB" and "free 25GB" figures are both out of date.
• Strong durability design (11 nines target) but no published uptime SLA; performance varies with node geography, best for large sequential objects.
• The 2026 pricing migration's opt-out terms (account frozen, data deleted within 45 days) are the sharpest trust issue we found; get exit terms in writing for large archives.
• Acquired by Inveniam Capital Partners in October 2025; Storj runs as a subsidiary, and the STORJ token economics now sit inside a fintech's roadmap.

Sources

• https://www.storj.io/pricing/
• https://storj.dev/dcs/pricing/simplified
• https://www.storj.io/pricing/change-FAQs/
• https://www.storj.io/about
• https://blocksandfiles.com/2025/10/23/web-3-storage-supplier-storj-sold-to-inveniam/
• https://storj.dev/node/payouts
• https://blokt.com/news/storj-to-disrupt-amazon-google-and-microsoft-says-co-founder-john-quinn?amp=