Pricing and the math
Unit economics at 100 PB: the storage line, the exit toll, the provider comparison, and the storage-class landscape.
Storage is $5.00 per TB/month, flat. Egress is $0 — free, uncapped, into any cloud, at line rate. Those two numbers are list prices, not promotions; the math below is the whole argument, at list, with assumptions stated.
The exit toll
Hyperscalers meter reads, so the cost that matters at petabyte scale is not the storage line — it is what the meter takes every time the dataset moves.
100 PB × $90/TB egress = ≈ $9M
Hyperscalers meter reads at $87–90 per TB, so moving your own 100 PB dataset off S3 at list costs about $9M — before a single training run. On Ramo the same read-out is $0, every time.
4.6×
cheaper than S3 on storage
$9M → $0
one full read-out
93 days
to ingest 100 PB at 100 Gbps sustained
100 PB, one year, at list
| 100 PB for one year: AWS S3 Standard vs Azure Blob Hot vs Ramo | |||
|---|---|---|---|
| Provider | Storage / year | One full read-out | Bulk ingest |
| AWS S3 Standardbest-effort, ~100 Gbps | $27.6M | $9M | ~93 days |
| Azure Blob Hot (LRS)~25 Gbps default account ingress cap | $21.6M | $8.7M | ~370 days |
| Ramosustained, 100 Gbps | $6M | $0 | 93 days |
Assumes list prices — S3 Standard $23, Azure Blob Hot (LRS) $18, Ramo $5.00 per TB/mo · egress: S3 $0.09/GB, Azure $0.087/GB, Ramo $0 · throughput: Ramo 100 Gbps sustained in production, S3 ~100 Gbps best-effort, Azure ~25 Gbps default account ingress cap · decimal units (1 PB = 1,000 TB = 1,000,000 GB). Ingest durations are transfer times at the stated rates: S3 ~93 days, Azure ~370 days, Ramo 93 days at the 100 Gbps we sustain in production. The S3 and Azure figures are best-effort rates, so those durations are what the transfer takes when the network cooperates, not a rate the provider holds.
Size it to your dataset
Same arithmetic, your numbers:
| Provider | Storage / yr | One read-out | Bulk ingest |
|---|---|---|---|
| AWS S3 Standard | $27.6M | $9M | ~93 days |
| Azure Blob Hot (LRS) | $21.6M | $8.7M | ~370 days |
| Ramo | $6M | $0 | 93 days |
Year one at 100 PB: storing on Ramo instead of S3, then reading the dataset out once, keeps $30.6M in your budget. The transfer lands in 93 days at a rate we sustain; the S3 and Azure figures above are best-effort.
Migration plan sized to 100 PB — endpoint and keys on approval.
Assumes list prices — S3 Standard $23, Azure Blob Hot (LRS) $18, Ramo $5.00 per TB/mo · egress: S3 $0.09/GB, Azure $0.087/GB, Ramo $0 · throughput: Ramo 100 Gbps sustained in production, S3 ~100 Gbps best-effort, Azure ~25 Gbps default account ingress cap · decimal units (1 PB = 1,000 TB = 1,000,000 GB).
The landscape
Built for training data. Priced below the archive tiers.
| Four storage classes compared: hyperscalers, edge object storage, discount storage, Ramo | ||||
|---|---|---|---|---|
| Class | Built for | $/TB/mo | Egress | Throughput |
| HyperscalersS3 · Azure Blob · GCS | General-purpose objects | $18–23 | $87–90 per TB | Best-effort, per-stream caps |
| Edge object storagee.g. R2 | Web objects — KB–MB, global | $10–15 | $0 | Cache-optimized, small objects |
| Discount storageWasabi · B2 | Backup & archive | $6–8 | $0 with fair-use caps | Best-effort, no SLA |
| Ramo | Petabyte datasets, sequential reads | $5.00 flat | $0 · uncapped | 100 Gbps sustained, not best-effort |
Discount-tier detail (list prices, July 2026): Wasabi $7.99/TB/mo (Pay-Go) — free egress under a fair-use cap; reads above your stored volume can limit or suspend service. Backblaze B2 $6.00/TB/mo — free egress up to 3× stored data, then $0.01/GB. Both are best-effort throughput with no SLA. The difference is the workload, not the price: backup & archive vs. TB–PB sequential reads.
One storage class is built for TB–PB sequential reads. This is it.