How to Survive - without reducing SNO compensation

A 36-month restructuring and business plan for Storj — without reducing SNO compensation

I want Storj to survive, and I am prepared to contribute real work toward that outcome.

But survival cannot mean transferring the consequences of failed sales expansion, failed product bets, inadequate customer pricing and an unsustainable capital structure to the Storage Node Operators who still provide the infrastructure on which the product depends.

This plan therefore starts with one non-negotiable premise:

The current SNO payout schedule is the floor. There must be no reduction — direct, indirect, temporary, technical or disguised.

That includes no lower storage, egress, audit or repair rate; no unpaid traffic class; no delayed payout; no longer held amount; no lower expansion factor without compensation neutrality; and no replacement of earned payouts with an IOU, equity or a future discretionary bonus.

The numbers reach the same conclusion: a small SNO cut is financially insufficient, a large cut is operationally destructive, and abolishing payouts abolishes the service.

1. The 12-week budget is a bridge, not the plan

The filed cash budget for August 4 through October 25 shows:

Filed 12-week cash budget Amount
Customer billings / cash inflows $930,634
Cost of Goods Sold $390,854
Total cash outflows $1,309,759
Cash deficit before DIP financing $379,125
Budgeted DIP financing $388,000
Cash remaining at the end $8,875

Docket 21 is the entered interim order, not merely a proposed order. It authorizes borrowing of up to $100,000 at the interim stage. The motion contemplates $388,000 in total, with the remaining $288,000 dependent on final relief. The DIP agreement carries 18% interest and matures twelve months after entry of the financing order.1

Even if all $388,000 becomes available, the budget consumes almost all of it. It does not fund the period after October 25, repay the DIP, deliver products or resolve prepetition liabilities.

The filings themselves identify the causes: the enterprise-sales expansion failed to generate revenue fast enough; Object Mount consumed considerable resources without gaining sufficient traction; Petagene and Valdi produced disappointing growth and minimal recurring revenue; and payroll and technology costs greatly exceeded income.1

What the budget implies after October 25

A mechanical annualization is not a forecast, but shows the scale:

Derived annualized view Cash inflows Cash outflows Gap
Full filed run-rate $4.03M $5.68M $1.64M
After removing obvious one-offs and the filed safety margin* $4.03M $5.15M $1.12M

Removed only: final paychecks, professional fees, trustee fee and the 5% safety margin. This is still not a normalized P&L because COGS payment timing, EOR payroll and restructuring effects remain.

Storj therefore needs more than a one-time $379,125 bridge:

  1. enough capital to survive the transition and Chapter 11 process;
  2. recurring collected revenue above recurring cost;
  3. a product plan that creates new profitable demand;
  4. a confirmed treatment of debt and claims;
  5. and enough liquidity to repay or refinance the DIP before maturity.

Within fourteen days Storj should publish an 18-month normalized monthly model and weekly rolling 13-week cash forecast, separating collected revenue, network economics, SNO payouts, gateway/satellite cost, partner and payment fees, payroll, contractors, fixed infrastructure, Chapter 11 one-offs and legacy claims.

2. Why an SNO payout cut does not solve the problem

SNO payouts are not a dividend or an optional community benefit. They are the variable supplier cost for the storage and bandwidth Storj sells. Abolishing them would be the equivalent of a cloud provider “saving” its data-center costs by turning off the data centers.

Current public network statistics show approximately 55.274 PB of logical customer data and 76.038 PB after expansion.2 At the current $1.50 per physical TB-month storage rate, the storage component of SNO payouts is approximately $114,000 per month or $1.369 million per year. This excludes egress, audit and repair.

Storage-rate action Gross monthly saving Gross annual saving Result
10% cut ~$11,400 ~$137,000 Does not close even one eighth of the conservative $1.12M planning gap
20% cut ~$22,800 ~$274,000 Leaves roughly $846,000 of that gap and damages operator economics
50% cut ~$57,000 ~$684,000 Still leaves a major gap and makes many nodes uneconomic
100% abolition ~$114,000 ~$1.369M Removes the storage supplier and therefore the product

Even under the impossible upper-bound assumption that every dollar of the filed $390,854 COGS line were reducible SNO cost, a 20% cut saves only $78,171 and leaves a $300,954 deficit; a 50% cut still leaves $183,698. Eliminating all COGS leaves only $11,729 while also eliminating the direct costs required to provide the service.

The petition separately lists approximately $4.038M owed to Inveniam, $3.742M claimed by the U.S. Treasury and $296,317 as an SNO liability: $8.076M across those three claims alone, before the DIP and its interest.1 An operator cut does not restructure those claims, create customers, finance integrations, repay the DIP or establish a cash reserve.

There is no economically useful SNO-cut range: a small cut is immaterial, a large cut destroys supply, and a 100% cut destroys the business.

The SNO Stability Covenant

Storj should incorporate the following into the restructuring plan:

Resource Permanent minimum
Storage $1.50/TB-month
Customer egress $2.00/TB
Audit and repair $2.00/TB

Lock the rates for at least 24 months with no automatic reduction. Any later change requires published impact and operator-concentration analysis plus six months’ notice. A lower expansion factor is acceptable only with compensation neutrality per logical customer TB.

The forum discussion is not network-wide consent. It is a voluntary sample affected by response and survivorship bias. Littleskunk expressly described his setup as deliberately efficient and warned that it should not be treated as a target.3 Operators outside the forum, those who already left and those who would quietly shut down are absent.

The limited visible reaction to the 2023 cuts does not prove another cut safe in 2026. Those measurements counted node IDs under different economics and placement. Node count does not reveal independent operators, used-data concentration, geographic value or repair load. Storj also stated that the present rates were believed sustainable and that further changes were not anticipated for the foreseeable future.4

In my operation, electricity, servers, network and IP costs alone consume about 58–60% of average payout, excluding hardware replacement, capital recovery, maintenance, labour, taxes and risk. Several contracts cannot be cancelled after a cut.

Large operators can leave with large amounts of used data and large absolute costs. Small operators may stop replacing failed drives because the remaining return no longer justifies their time. Both can happen together. A broad cut cannot selectively remove only unused capacity.

Storj should therefore publish anonymized concentration by the largest 10, 25 and 50 independent operators, country, ASN and /24, together with simulated repair volume and repair cost if the largest operators leave simultaneously.

3. Correct the customer price, not the infrastructure price

The filed budget needs 40.74% more cash inflow merely to cover twelve-week outflows with zero churn, leaving no reserve or product budget. The immediate target should therefore be a 70% increase in weighted collected unit revenue with 24-month price stability.

Proposed prices

Plan Storage Egress Commitment
Standard Self-Service $12/TB-month $12/TB $5 minimum, fully credited against use
Advanced / Select $17/TB-month $12/TB $250 monthly committed spend, fully credited
Enterprise Committed floor $10/TB-month floor $8/TB 12-month take-or-pay, minimum $5,000/month and minimum 60% gross margin

New customers move immediately; month-to-month customers receive 30 days’ notice; fixed contracts move at the earliest lawful repricing or renewal. Standard gets direct credit-card checkout, and no contract is approved below documented fully loaded direct cost and margin floors.

The current public pricing is $7/$7 for Standard and $10/$7 for Advanced. The public page also presents a $5 Object Storage minimum while stale Object Mount text still refers to a $50 minimum, and Standard is still routed through “Contact sales.” These contradictions must be corrected immediately.5

The failed $50-minimum rollout is also a warning: Storj later reverted it to $5 and acknowledged that customers had already left.6 The answer is not another improvised price change. It is one comprehensible structure, proper notice and a 24-month lock.

Optional simple bundles can remove the fear of metered billing. Storage is listed first; every bundle includes automatic provisioning, the relevant Storj Connect integration and Standard overage:

  • Developer Connect — $9/month: 250 GB + 100 GB egress.
  • WordPress Connect — $15/month: 500 GB + 250 GB egress; media and backups.
  • Personal Cloud Connect — $19/month: 1 TB + 250 GB egress; Nextcloud/ownCloud.
  • Media Connect — $29/month: 1 TB + 1 TB egress; Plex/Jellyfin/Emby.
  • NAS Connect — $39/month: 2 TB + 500 GB egress; Synology/QNAP/TrueNAS/Unraid.
  • AI Data Workspace — $49/month: 2 TB + 1 TB egress; datasets, checkpoints and models.
  • Server Backup Connect — $79/month: 5 TB + 500 GB egress; Proxmox/Restic/Borg/Velero.
  • Hosting & Agency Connect — $109/month: 5 TB + 2 TB egress; cPanel/Plesk/multi-site WordPress.
  • MSP & Business Continuity — $219/month: 10 TB + 5 TB egress; multi-customer backup and disaster recovery.

Churn stress test

Exact churn cannot be predicted without customer-level contract, usage and margin data. The defensible approach is a sensitivity test assuming a 70% weighted price increase, no new sales and the filed $1,309,759 outflows.

Weighted usage loss New 12-week billings Result with no cost relief Result if 50% of COGS varies with usage
0% $1,582,078 +$272,319 +$272,319
10% $1,423,870 +$114,111 +$133,654
15% $1,344,766 +$35,007 +$64,321
20% $1,265,662 -$44,097 -$5,011
25% $1,186,558 -$123,201 -$74,344

The reset covers filed outflows with 17.21% weighted usage loss even if no cost falls, 19.64% if half of COGS varies, and 22.86% if all COGS varies. Plan for 10% base, 15% downside and 20% stress; above 20% triggers corrective action and capital—never an SNO cut.

Account churn and usage churn differ: many tiny accounts may leave while representing little data or revenue, whereas large customers face migration and contract friction but greater Chapter 11 trust risk. Within fourteen days Storj must replace these assumptions with a material-customer cohort table covering revenue, usage, term, direct COGS, proposed price, renewal and churn owner.

A price increase can temporarily reduce absolute SNO payouts as customers delete data, but that is not a unit-rate cut. A 10% storage loss reduces payout volume about $11,400/month, while a 70% reset with 10% usage loss raises the filed collection equivalent from about $336,000 to $514,000/month. Released capacity remains available and growth restores payout volume.

Customer deletion releases capacity. Operator departure forces repair and migration of customer data. The first corrects demand; the second damages supply.

Competitive position

Illustrative U.S./European list rates vary by region, requests and contract:

Provider Hot storage/TB-month Internet egress/TB Notes
Proposed Storj Standard $12 $12 Simple price, no general request fee
Proposed Storj Advanced $17 $12 Compliance-controlled placement
AWS S3 Standard ~$26 ~$90 Request and feature charges
Azure Hot Blob ~$20 ~$87 Transaction charges; region/redundancy vary
Backblaze B2 $6.95 free to 3× storage, then $10 Strong low-cost competitor
Cloudflare R2 $15 free Class A/B operation charges

Storj remains substantially cheaper than AWS and Azure.7 It will not beat B2 or R2 on every workload, so it must win on distribution, durability, security, predictable billing, performance, migration and radically easier integration—not a race to the bottom.

4. The 36-month operating plan

The filed billings equal approximately $336,000 per month. A 70% price reset with 10% weighted usage loss produces approximately $514,000 collected monthly revenue before any new customer growth.

The operating model should cap recurring fixed cash cost — payroll, contractors, ordinary administration, subscriptions and non-variable infrastructure, excluding direct COGS and Chapter 11 one-offs — at $300,000/month during Year 1. With direct COGS capped at 40% of collected revenue, cash break-even is:

$300,000 / (1 - 40%) = $500,000 collected MRR

Required targets — not forecasts

Target period Collected MRR / average Direct COGS cap Fixed cash cap Operating cash before restructuring and debt service
Day 30 ≥$500,000 run-rate 40% $300k/month break-even run-rate
Month 3 ≥$550,000 40% $300k/month positive
Month 6 ≥$650,000 40% $300k/month ≥$90k/month
Year 1 $600k average / $800k exit 40% $3.6M/year ≥$720k/year
Year 2 $950k average / $1.1M exit 36% $4.08M/year ≥$3.22M/year
Year 3 $1.30M average / $1.50M exit 33% $4.8M/year ≥$5.65M/year

Management must replace these targets with actual cohort and COGS data; growth spending expands only after collected gross profit pays for it.

Liquidity policy:

  • at emergence: at least six months of fixed cash cost;
  • end of Year 2: at least nine months;
  • end of Year 3: at least twelve months;
  • no distribution, buyback or speculative product investment before those reserves exist;
  • refinance, convert or repay the DIP well before its twelve-month maturity.

Exit financing: minimum $4M, target $5M

The $388,000 DIP is too small and too expensive to fund a turnaround. Storj should raise equity, preferred equity or a deeply subordinated convertible instrument — not another short 18% operating loan.

Proposed $5M use of funds Amount
12-month operating and transition reserve $1.20M
Storj Connect, integrations, security and QA $0.80M
Customer migration, channel launch and paid pilots $0.50M
Chapter 11, legal, tax and plan implementation $0.50M
Reserve for the filed SNO liability pending validation $0.30M
DIP principal plus maximum indicative one-year interest $0.46M
Minimum unrestricted liquidity reserve $0.75M
Contingency $0.49M
Total $5.00M

This capital is not permission to preserve the old burn rate. Release it in milestones tied to price migration, cost caps, product delivery and collected MRR.

The restructuring should seek conversion of the Inveniam prepetition claim into equity, preferred equity or long-term subordinated debt; lawful court-approved treatment of the Treasury claim; full disclosure of the $296,317 SNO liability; on-time payment of all postpetition SNO obligations at current rates; and a funded objective to pay valid prepetition SNO claims in full under the confirmed plan.

Inveniam is listed as a 10%+ equity owner, the largest unsecured creditor and the DIP lender, and the DIP agreement includes credit-bid rights.1 That is not proof of wrongdoing, but it requires an independent restructuring committee, independent valuation and a genuine market check for any asset sale, credit bid, debt conversion or change of control.

5. Build the missing product: Storj Connect

Storj does not primarily have a protocol problem. It has a productization and onboarding problem.

S3 compatibility is valuable because many applications already speak that interface. But S3 is a compatibility layer, not a customer journey. A WordPress operator should not need to understand buckets, access keys, endpoints, regions, path-style access or Signature V4. A Plex user should not need to learn Rclone, VFS cache flags and mount services.

Storj’s own guides currently require exactly those manual steps for WordPress, Plex, QNAP, TrueNAS and other tools.8 Direct integrations are therefore not an alternative to S3. They can use S3 invisibly underneath, or native Uplink when client-side encryption or download performance makes that preferable. The customer should only click Connect Storj.

Architecture

WordPress / Plex / NAS / backup / AI tool
                  |
        Storj login or device code
                  |
       Storj Connect Control Plane
 project + bucket + scoped access + policy
 billing + migration + health + revocation
          /                       \
 S3-compatible adapter       Native Uplink
          \                       /
               Storj network

Control Plane MVP

The first two weeks should define one integration manifest and stable APIs:

POST /v1/integrations       create and bind an app
POST /v1/projects           provision a project
POST /v1/buckets            provision policy and lifecycle
POST /v1/credentials        issue scoped credentials
POST /v1/credentials/rotate rotate or revoke access
POST /v1/migrations         start/resume/rollback migration
GET  /v1/health             integration diagnostics
GET  /v1/billing-estimates  current use and projected bill

The manifest defines permissions, bucket/prefix, cache, retention, migration, health tests and bundle. The platform needs device-code/OAuth login, scoped credentials, rotation, revocation, audit logs, sandbox, webhooks, CLI, Terraform and generated SDKs for TypeScript, PHP, Python, Go, Java and .NET.

Storj Bridge

A small open-source local agent for Windows, macOS, Linux, Docker and common NAS platforms should provide:

  • FUSE/WinFsp filesystem access and optional SMB/WebDAV;
  • read-through/write-back cache, range reads and sequential prefetch;
  • resumable multipart upload, upload watching and offline pinning;
  • local metadata cache so media scanners do not repeatedly list entire buckets;
  • bandwidth schedules, health checks, logs and one-click support bundles;
  • signed automatic updates.

Reuse useful Object Mount technology as a customer-acquisition component, not another speculative standalone product and sales motion.

Named 90-day team

Team: one empowered product owner, two platform/systems engineers, one WordPress/PHP engineer, one frontend/UX engineer, one DevOps/security engineer and one QA/compatibility lead. Roles may be combined, but every release needs one accountable owner.

6. Products that can create demand immediately

Official WordPress plugin — Storj Media & Backup

User flow: install plugin → click Connect with Storj → choose Media, Backups or Both → start migration.

It must automatically provision the account resources, offload new media, bulk-migrate existing media with resumable jobs, rewrite URLs, support a custom domain/CDN, retain a reversible local-copy option, restore to local storage, support Multisite/WooCommerce/page builders, back up files and database, estimate the bill and diagnose failures.

Private alpha in 30 days, public beta in 45 days, production in 90 days. My own earlier customer experience showed the problem clearly: a conventional CDN integrated in an afternoon, while Storj would have required rebuilding the media pipeline over weeks. Price was not the deciding factor; integration friction was.9

Plex, Jellyfin and Emby profile

User flow: install Docker/Unraid/NAS package → enter device code → choose cache folder and media library → point the media server at the mounted folder.

The profile needs metadata caching, range-read optimization, sequential read-ahead, configurable SSD cache, offline pinning, scan throttling, upload watching and migration from local disk or another object store. Alpha by Day 45; public beta by Day 90.

NAS, backup and hosting

Turn existing guides into one-click packages for Synology, QNAP, TrueNAS, Unraid, Proxmox, Nextcloud, cPanel, Plesk, Restic, Rclone, Velero and MSP backup products. Technical compatibility is not a finished customer journey.

Storj AI Data Plane

Do not build another capital-intensive standalone GPU company. Build a Bring Your Own Data / Model / Compute data plane for customer-owned or partner compute.

One click should create a dataset/model workspace, bucket, scoped access, lifecycle policy, local cache and generated Docker Compose or Kubernetes configuration. Initial templates:

  • Hugging Face datasets;
  • MLflow artifacts and model registry;
  • DVC datasets;
  • PyTorch/TensorFlow checkpoints;
  • Ollama, llama.cpp and vLLM model caches;
  • versioned dataset snapshots and experiment outputs.

Storj already documents Hugging Face through S3FS, and MLflow and vLLM support S3-compatible object stores.10 The missing product is automatic provisioning and an idiot-proof workflow.

Any offer advertised as “zero egress to compute” must be funded by compute or partner revenue sufficient to pay the full current SNO egress rate. A customer promotion cannot become unpaid operator traffic.

7. Product-led distribution and strict commercial rules

The court filing says the enterprise-sales expansion failed to generate revenue fast enough. The answer is not another broad sales hiring cycle. Distribution should come through the product:

  • WordPress plugin directory;
  • Docker Hub and GitHub Container Registry;
  • Unraid and NAS app stores;
  • cPanel/Plesk and backup marketplaces;
  • MSPs, WordPress agencies and hosting providers;
  • AI/MLOps templates and compute partners.

Commercial rules:

  • Standard is fully self-service; sales handles Advanced, migrations and committed enterprise contracts.
  • Partner commission is based on collected gross profit or 5% of collected revenue for the first twelve months, not announced ARR.
  • Sales compensation is based on collected gross profit and retention.
  • No substantial custom engineering without a paid pilot, named buyer, conversion deadline and margin model.
  • No new product line without a prepaid design partner.
  • Every customer cohort receives a monthly contribution-margin report.
  • Negative-margin custom contracts are repriced, redesigned or discontinued.

During Chapter 11 Storj should publish weekly aggregated actual-versus-budget collections, categorized COGS, SNO obligations due and paid, unrestricted cash, DIP availability, fixed-cost run-rate, price migration, weighted usage churn, new collected MRR and exit-financing status; monthly reporting continues after emergence.

8. Execution schedule and failure triggers

Days 0–7

Publish the SNO Stability Covenant; confirm postpetition payout dates; disclose COGS and the SNO-liability breakdown; correct Docket 21 status and current DIP availability; announce the new prices and 24-month lock; remove contradictory website pricing; appoint the turnaround owner and Storj Connect product owner; begin the $5M exit-capital process.

Days 8–30

Activate new-customer pricing; issue proper notice to existing self-service customers; finish the contract/cohort audit; launch direct Standard checkout; publish the normalized 18-month model; release Control Plane alpha, public API specification and WordPress private alpha; reach at least $500k collected MRR run-rate.

Days 31–90

Release WordPress public beta, Storj Bridge alpha, media-server beta and initial AI templates; sign at least ten paying design partners; complete eligible price migration; hold COGS at or below 40% and fixed cash at or below $300k/month; reach $550k collected MRR; sign binding exit-financing and restructuring terms.

Months 4–6

Ship production WordPress and Bridge releases; add NAS, hosting and backup packages; launch the integration marketplace; reach $650k collected MRR and positive recurring cash flow; file or become ready to file a confirmable Chapter 11 plan; refinance or convert the DIP.

Months 7–12

Reach $800k collected MRR; emerge with at least six months of fixed-cost liquidity; complete court-approved treatment of valid SNO claims; maintain current SNO rates and on-time payment.

Years 2–3

Reach $1.1M exit MRR in Year 2 and $1.5M in Year 3; expand integrations and partner distribution only from collected gross profit; build nine then twelve months of liquidity; repay or refinance remaining restructuring obligations without using SNO compensation as the balancing item.

Mandatory failure triggers

  • More than 20% weighted usage loss: use the churn reserve, reprice negative-margin contracts, reduce fixed cost and accelerate acquisition — no SNO cut.
  • No $500k run-rate or credible financing path by Day 60: freeze all non-core development and begin a formal strategic-options process.
  • No $550k collected MRR and no binding exit-financing/restructuring terms by Day 90: start an orderly competitive sale or transfer while cash and customer trust remain.
  • No positive recurring cash flow by Month 6: proceed with the strategic transaction rather than another emergency loan cycle.

Any sale must protect customer data, postpetition payments, current SNO rates and independent review. Waiting until cash is exhausted destroys leverage and network value.

9. My concrete contribution — and the conditions required for it

I have founded and built several companies. A serious turnaround requires the real financial, operational, customer and technical data—not only public fragments. If Storj provides it, under an NDA where necessary, I am prepared to help turn this proposal into an executable restructuring and business plan.

I also have several programmers available and will donate development work worth a five-figure amount for Storj Connect, WordPress, Bridge, APIs/SDKs, tests, migration tools or another agreed priority. This requires a named product owner, acceptance criteria, sandbox/API access, reviews and a real release path.

Conditions

  1. Current SNO compensation remains unchanged in nominal and effective terms.
  2. Existing professional multi-node deployments are not retroactively penalized or traffic-suppressed solely for using multiple ISP ranges, routed subnets, VPS endpoints or operator-managed VPNs.

This is not a request to ignore correlation risk or receive preferential treatment. Existing disclosed deployments need a formal grandfathering and approval path, followed by prospective rules based on measured failure domains, uptime, audits, bandwidth, latency, geographic and ASN concentration, and actual repair impact—not labels such as “large farm,” “VPS” or “VPN.” Qualifying deployments should be explicitly permitted to continue; materially different future rules require notice, impact analysis and a practical migration period.

Failure-domain illustration

Public forum data showed approximately 12,844 visible /24 blocks. Assume three professional operators each run 1,000 nodes across 1,000 /24s, evenly split among three independent host and storage failure domains.

If one host domain at each operator failed simultaneously, about 1,000 /24s—or 7.8% of the visible pool—would be affected. With 49 pieces and 29 required for reconstruction, a uniformly distributed segment would lose about 3.8 pieces on average and retain about 45. It must lose at least 21 pieces to become unrecoverable; under this simplified uniform model, the probability in the described outage is approximately 2 × 10⁻¹¹ per segment.

The outage could trigger repair traffic and is not cost-free, but repair is not data loss and does not justify excluding professional farms or VPS/VPN-assisted deployments indiscriminately. A complete operator-wide outage is different and should be tested with Storj’s actual placement-weighted operator, host, subnet and ASN data—not inferred from connection type.

Real correlated failure domains should be managed through targeted controls, disclosure and agreed transitions, not opaque blanket filters based on size or architecture.

My contribution can create customers and reduce development cost. I will not donate it while accepting lower SNO compensation or policies that devalue or exclude the infrastructure on which my contribution and existing investment depend.

Conclusion

Storj’s connected problems are inadequate customer economics, integration friction, unfocused product and sales spending, insufficient liquidity, an unsustainable capital structure and damaged trust. An SNO cut solves none of them.

Price the product sustainably + make Storj radically easier to use + turn integrations into distribution + cap recurring costs + restructure debt + fund sufficient runway + protect the SNO infrastructure.

Storj can remain far cheaper than AWS and Azure, hide S3 complexity behind direct integrations, and turn existing guides into products for WordPress, media, NAS, backup, hosting and self-hosted AI. It can emerge from Chapter 11 with positive recurring cash flow and adequate liquidity.

It cannot repair pricing, product, sales and capital-structure failures by repeatedly lowering payment to the infrastructure that still works.

Not one cent of this plan requires a reduction in the current SNO rates.

Sources

1 Court · 2 Stats · 3 Operators · 4a Rates / 4 2023 · 5 Pricing · 6 Reversal · 7 AWS / 7b Azure / 7c B2 / 7d R2 · 8 WordPress / 8b Plex / 8c QNAP / 8d Rclone · 9 Example · 10 Hugging Face / 10b MLflow / 10c vLLM

That is IMHO a pretty stupid premise.
Customers don’t care about how much it costs YOU to buy gasoline if you are running a petrol station.

STORJ is selling a commodity, just like gas. If you ask a bigger price than the next gas station, they will not come. This is what happened to STORJ over the last decade.

“But my drilling plant is more expensive to operate, we have to ask twice the price” customer does not care.

The USP of STORJ always was, you have a oil spilling in your garden that you can harvest for free. Sure it is a laughable small amount of oil, compared to some big off shore rigs, but hey, the difference is that it is 100% free for you! That is why your small garden rig can undercut the price of a high tech off shore rig.

But I agree with you on that part

That is not enough. We need drastic price cuts for customers, drastic price cuts for nodes and the most important part, drastic cost cuts.

Which means, no more S3!

But more realistically, someone could pick up the code and fork a community shared backup solution. Slow, huge chunks to ease stress on nodes, some community run satellites to keep track of things. New stablecoin token. The idea is not to exchange your token but that I store 1TB for the network and the network will store 1TB for me. Still you can of course exchange it.

But given that I, as a nobody and nocoder, realized that storj devs did not even understand when to use sync writes and when async, I doubt that the code offers much of value.

Please. Have your LLM-of-choice summarize the post: there are too many words for too few ideas.

The current node-count/free-space is the result of the rates that have been paid. The network has higher node-count/free-space than required: by a healthy margin: and it has been that way for years. Of course a reduction should be considered.

Chapter 11 is certainly the time to make some large changes. But it’s also a time to make a large number of small changes: they are cumulative. A reduction of payout could be one such small change. It’s not like if a single change doesn’t solve the entire problem it’s not worth doing.

I struggle to think of a business that commits to pay suppliers at a fixed rate for two years. There may be times contracts dictate those prices… but SNOs are a-dime-a-dozen with no commitment to Storj. Normally a business can choose to try to get better deals at any time. Does Storj gain higher sales, or lower costs, by making such a commitment?

It’s a competive market: there are lots of S3 options out there. Storj must determine what price customers will accept: they can’t dictate it. McDonalds can’t just decide to charge $50 for a hamburger: because there are lots of other places to buy a hamburger…

S3 is a backend technology: prosumer at best. It is not consumer. Storj doesn’t want that Wordpress user… they want IT departments backing up 10TB+. They are not competing with Box, or Dropbox, or Google Drive, or Apple Cloud.

There are lots of interesting ideas in this post. But it was written as if “the market” will go along with whatever Storj decides to do, and buy whatever they’re selling, for whatever price they want, with no support burden. It was written as if customers had no alternatives.

Thanks. Both replies make the same basic move: they treat lower SNO compensation as an obvious consequence of market competition, without showing that it solves the actual financial problem or even reduces the cost they claim it addresses.

So here is the entire argument in one sentence:

Storj pays nothing for unused node capacity. Cutting SNO rates therefore does not reduce the cost of excess free space; it reduces payment for storage and bandwidth that are actually being used by customers, while leaving the larger structural problems unresolved.

@GreenGrinch

If you open with that, you should at least make sure the analogy that follows actually fits the economics.

You are correct about one thing: customers do not care what it costs me, or any other SNO, to provide the infrastructure.

But customers not caring about supplier costs does not make those costs disappear.

A petrol station whose customers refuse to pay more than the wholesale cost of the fuel does not become viable because the customers are indifferent to that wholesale cost. It must reduce its non-fuel expenses, negotiate a genuinely sustainable supply price, create a differentiated offer, increase volume at a positive margin or leave the market.

It cannot solve the problem indefinitely by demanding that the supplier sell below the cost required to keep supplying it.

Storj is also not selling a perfectly identical litre of gasoline. Object-storage providers differ in egress pricing, request costs, geography, durability, compliance, performance, integrations, support and migration friction. Price is important, but it is not the only variable.

That may describe the original marketing story for hobbyist spare capacity. It does not accurately describe the entire production network today.

Some operators use existing hardware and genuinely low marginal-cost capacity. Others operate dedicated or partially dedicated infrastructure with electricity, replacement hardware, bandwidth, servers, network contracts and maintenance costs.

Even where the hardware originally existed for another purpose, storing customer data is not literally free:

  • drives wear and fail;
  • electricity is consumed;
  • bandwidth is used;
  • replacement capacity must be purchased;
  • outages must be diagnosed;
  • and the operator remains responsible for keeping customer pieces available.

A historical narrative about “free garden oil” is not a valid method for determining the current network-wide supplier price.

It is also economically irrelevant to the excess-capacity argument because Storj does not pay for the empty part of those drives in the first place.

You cannot simply put three uses of the word “drastic” next to each other and call it a business model.

If customer prices fall and supplier prices fall simultaneously, Storj’s margin improves only if:

  1. central costs fall by substantially more;
  2. lower customer prices produce enough additional volume;
  3. that additional volume remains contribution-positive;
  4. and the necessary volume arrives before Storj runs out of cash.

None of that has been demonstrated.

Using the figures already sourced in the opening post, the current storage payout is approximately $2.06 per logical customer TB-month after expansion.

A 20% storage-rate reduction therefore saves only about:

$0.41 per logical customer TB-month

If every cent were passed directly to the customer, a $7 storage price could fall to approximately $6.59.

That is not a drastic customer-price reduction. It does not transform Storj’s competitive position.

At the network level, the same 20% storage cut saves approximately $22,800 per month before accounting for operator exits, repair traffic, data migration or the loss of future capacity.

The proposed customer-price reset, even with 10% weighted usage loss, improves the filed monthly cash-collection equivalent by approximately $178,000.

The central cost structure and customer economics are therefore much larger levers.

That is why the proposal includes:

  • a $300,000 monthly fixed-cost ceiling;
  • a direct-COGS ceiling;
  • elimination or repricing of negative-margin contracts;
  • no speculative development without prepaid demand;
  • product-led distribution rather than another expensive sales expansion;
  • debt restructuring;
  • and sufficient exit capital.

This is not an argument against cost reduction.

It is an argument against pretending that SNO compensation is the cost item that caused the bankruptcy or can repair it.

To create a truly drastic customer-price reduction from SNO storage payouts, Storj would have to remove most or all of the payout. At that point the company would no longer have a dependable storage supplier network.

No. That conclusion points in exactly the wrong direction.

S3 is a compatibility interface. It is not supposed to be the user experience.

The proposal explicitly says that a WordPress operator, NAS owner, media user or AI customer should never need to understand buckets, endpoints, access keys, signatures or regions.

The direct integration can provision and configure S3 invisibly underneath. It can use native Uplink where that is more efficient. The customer should only see Connect Storj.

If Storj’s present S3 implementation has excessive central cost, then Storj should:

  • disclose that cost;
  • optimize or self-host the expensive components;
  • move suitable traffic to native integrations;
  • offer customer-operated gateways where appropriate;
  • and charge workloads that create disproportionate gateway expense accordingly.

Removing S3 compatibility entirely would discard access to a very large existing software ecosystem at the exact moment Storj needs more distribution and easier adoption.

“This implementation is too expensive” does not logically mean “remove the interface that most existing storage applications support.”

That may be an interesting separate project.

It is not a restructuring plan for Storj Labs.

It does not resolve:

  • existing customer contracts;
  • the Chapter 11 claims;
  • the DIP financing;
  • enterprise and compliance obligations;
  • current data continuity;
  • support responsibilities;
  • or the need to generate real cash revenue.

A reciprocal community backup network with a new stablecoin would be a different product, company and risk model. It may be worth discussing separately, but it cannot be substituted for a plan addressing Storj’s current obligations.

Then identify the relevant code path, commit, issue or benchmark.

If there is a concrete technical defect, showing it would be useful.

Without a specific example, this is not technical criticism. It is an insinuation that nobody can evaluate or fix.

@Roxor

I am going to be blunt because this pattern has become difficult to ignore.

I provided a quantified 36-month restructuring proposal containing:

  • customer prices;
  • cash-flow thresholds;
  • churn scenarios;
  • fixed-cost ceilings;
  • financing requirements;
  • product architecture;
  • implementation deadlines;
  • and mandatory failure triggers.

Your response was to complain about the number of words and repeat the most obvious sentence available:

Cut node payments.

Whether an LLM helped structure or edit the text is irrelevant. Every calculation and assumption can be examined independently.

You are one of the most visible and prolific people on this forum. Some of your contributions are genuinely useful.

Far too many others are reflexive, underdeveloped and delivered with a level of confidence that the analysis behind them does not justify.

Because you post constantly, familiarity is often mistaken for authority.

It is not.

Volume is not rigor.
Confidence is not evidence.
Being the most familiar voice in the room does not make the quickest answer the correct one.

This is exactly the kind of superficially obvious conclusion I am criticizing.

Storj pays nothing for unused free storage.

If the network has 100 PB of unnecessary empty capacity, the direct storage payout for that unnecessary capacity is:

$0

Reducing the storage rate does not reduce the cost of that excess free capacity.

It reduces compensation for the occupied storage containing customer data.

A broad rate cut also cannot determine which capacity disappears. It does not selectively remove:

  • empty nodes;
  • unreliable nodes;
  • badly connected nodes;
  • or capacity in oversupplied locations.

It may instead cause a mature operator holding substantial customer data to leave while empty capacity elsewhere remains online.

That creates repair and migration traffic rather than solving the capacity imbalance.

If Storj wants to stop capacity growing faster than demand, it has targeted tools:

  • pause or reduce generic node onboarding;
  • slow vetting in oversupplied placements;
  • stop encouraging additional generic capacity;
  • target new capacity only where geography, ASN or performance requires it;
  • and measure concentration by independent operator rather than node ID.

Those measures address excess capacity directly.

Reducing payment on occupied customer storage does not.

That narrow logical point is correct.

A measure does not have to solve the entire problem by itself to be useful.

I accept that describing the saving as “immaterial” can sound too absolute. Approximately $22,800 per month from a 20% storage cut is not literally zero.

The accurate description is:

It is a relatively low-leverage, poorly targeted saving whose operational consequences have not been quantified.

Chapter 11 is indeed the time to make many changes.

But cumulative changes should still be ranked by:

  • financial impact;
  • implementation speed;
  • reversibility;
  • secondary cost;
  • and operational risk.

A customer-price change that can improve monthly collections by approximately $178,000 is a larger lever than a storage-rate cut saving approximately $22,800.

A reduction in central fixed costs is another larger lever.

Debt conversion and exit financing address liabilities that a payout cut does not touch at all.

A company in Chapter 11 should implement the highest-impact, lowest-damage measures first. It should not begin with the measure that saves substantially less while directly weakening the infrastructure needed to deliver the product.

If you believe a payout reduction is still worthwhile, then do the work and provide:

  1. the exact proposed storage, egress and repair rates;
  2. the resulting monthly gross saving;
  3. the expected number of independent operators leaving;
  4. the amount of occupied customer storage affected;
  5. the repair and migration traffic created;
  6. the cost of that repair traffic;
  7. the remaining monthly deficit after the cut;
  8. the customer-price reduction the saving would actually finance;
  9. and the reason another reduction would not be required six months later.

Until those numbers exist, “a lot of free space exists, therefore lower the rate on used space” is not a restructuring analysis.

It is an intuition.

You are conflating abundant prospective capacity with occupied, vetted and functioning production capacity.

An empty new node is not costlessly interchangeable with a mature node already holding customer pieces.

When the latter leaves:

  • its pieces must be reconstructed;
  • bandwidth is consumed;
  • repair payouts are created;
  • redundancy temporarily falls;
  • and operational risk increases.

You also cannot argue both of the following at once:

  1. SNOs have no commitment and may leave at any time.
  2. Storj can repeatedly reduce their rates without creating a continuity risk.

The first statement makes predictable economics more valuable, not less.

Businesses enter fixed-price supplier arrangements constantly when continuity and predictable input costs matter.

Bandwidth, colocation, electricity, logistics, hardware maintenance and cloud commitments are routinely contracted for fixed terms.

More importantly, the proposal does not require Storj to guarantee a purchase volume.

Storj would still pay only for storage and bandwidth actually used.

A two-year unit-rate floor is not a two-year commitment to purchase unused capacity.

Storj gains:

  • predictable infrastructure unit costs;
  • lower risk of operator exits during restructuring;
  • confidence that operators will replace failed hardware;
  • and credibility after repeatedly changing the assumptions under which operators invested.

That stability has real business value during a restructuring in which customers, creditors and investors are already questioning continuity.

Correct.

The market determines what customers will accept.

The proposed $12/$12 price is not a command to the market. It is a testable restructuring proposal derived from the filed cash gap.

That is why the original post includes:

  • churn scenarios from 0% to 25%;
  • a base case, downside case and stress case;
  • calculations with and without variable-cost relief;
  • additional capital requirements;
  • and mandatory failure triggers if the price is rejected.

Nobody proposed a random $50 hamburger.

The proposal also explicitly acknowledges that Storj will not be the cheapest provider for every workload against Backblaze B2 or Cloudflare R2.

Storj must compete through the combination of:

  • pricing;
  • global distribution;
  • durability;
  • predictable billing;
  • migration;
  • security;
  • and easier integration.

If the market rejects the proposed price, Storj must adjust the product, cut central costs further, obtain more capital or pursue a strategic transaction.

What it cannot do indefinitely is preserve a customer price that fails to fund the company and require one supplier group to absorb the difference.

That is not market discovery. It is choosing who finances the loss.

That is your preferred strategy, not an established fact.

The court filing itself describes Storj as serving primarily individual users and only a few enterprise clients. It also says that the expanded enterprise-sales effort failed to generate revenue fast enough and that Object Mount failed to gain sufficient traction.

An enterprise-only strategy has therefore not demonstrated that it can sustain the company.

WordPress Connect is also not intended as a competitor to Dropbox, Google Drive, Box or iCloud.

The relevant customers include:

  • hosting providers;
  • WordPress agencies;
  • managed-service providers;
  • backup platforms;
  • and companies managing hundreds or thousands of sites.

One individual website may be small.

One successful integration with a hosting company or agency can aggregate a commercially significant amount of storage without Storj manually selling to every site owner.

The distinction between a “WordPress user” and an “IT department storing more than 10 TB” is therefore false. A simple WordPress integration can be the distribution mechanism through which a business customer brings hundreds of terabytes.

No, it was not.

The post explicitly compares alternatives, acknowledges stronger competitors for certain workloads, models customer loss and defines what must happen if the assumptions fail.

What it refuses to assume is that the only possible response to competition is repeatedly lowering SNO compensation.

There are other competitive levers:

  • lower onboarding friction;
  • automatic migration;
  • predictable bundles;
  • native integrations;
  • self-service purchasing;
  • lower central overhead;
  • better channel distribution;
  • and a product that does not require customers to understand the backend.

You also mentioned support burden.

That is a legitimate concern, but it is not an argument against integrations. It is a product requirement.

If a self-service connector still creates substantial manual support work, then the connector has failed.

That is why Storj Connect includes:

  • automatic provisioning;
  • scoped credentials;
  • migration and rollback;
  • health checks;
  • billing estimates;
  • and integration-specific diagnostics.

The entire purpose is to turn repeated manual support into software.

The actual disagreement

The disagreement is not whether Storj should reduce costs.

It must.

The disagreement is whether another SNO reduction is an evidence-based cost measure or merely the easiest stakeholder to target.

Anyone proposing a cut should publish:

Required item Required answer
New rates Exact storage, egress and repair compensation
Gross saving Monthly and annual cash effect
Customer effect Actual price reduction the saving enables
Operator response Expected independent-operator and occupied-data loss
Repair consequence Traffic, cost and redundancy effect
Remaining deficit Cash deficit after the cut
Long-term model Why another cut will not follow

Then that proposal can be compared directly with the pricing, fixed-cost, integration, financing and debt-restructuring plan I provided.

Without those numbers, “cut the payments” is not a better plan because it is shorter.

It is simply the easiest sentence in the room.

Finally, I am not arguing that Storj should preserve SNO compensation while everybody else contributes and I contribute nothing.

I have offered development work with a normal commercial value in the five-figure range to build integrations, migration tools and revenue-generating products.

Storj is free to reject that offer.

But it cannot reasonably expect me to donate substantial engineering work while simultaneously reducing compensation for my existing infrastructure or introducing policies that devalue it.

Disagreement is fine.

Challenge the prices.
Challenge the calculations.
Challenge the churn assumptions.
Propose a better cost structure.
Provide a better implementation plan.

But dismissing a quantified restructuring proposal as “too many words” and replacing it with “cut the payments” is not serious analysis.

My position therefore remains unchanged:

Cut central waste, correct customer economics, remove integration friction, restructure the liabilities and finance the transition—but do not use SNO compensation as the balancing item.

Not a replacement. A comment on one part. As an alternative to be quoting all the other parts and saying “sounds reasonable” 20 times :wink:

But you addressed my concern. You started with…

…but have clairified:

I thought you were saying a reduction could not be part of the solution. But I think you just want to be clear it’s not effective alone. And are pushing back against the idea that it could be “the fix”. I agree. It’s just one part.

It’s just not too small to dismiss.

The idea that unused space has no cost also has some nuance. In that if Storj wanted more of it… they would have to pay more to get it. But that’s neither here nor there.

I appreciate the work you’ve put into the plan: hopefully Storj is standing-on-its-own-two-feet again in 6 months and we will no longer worry… :crossed_fingers:

@GfTmbH
First of all thank you for your great post.
Now we are starting to discuss the relevant things it seems.

I cannot tell if I agree to everything you wrote, but the direction is clear: The DIP will take us 12 weeks from now.

Without fundamental changes Storj will be in the same position by the end of October. And the figures prove ones more that the SNO payout is not what is killing the company. It is the lack of revenue.

So basically Storj needs to collect around $1M, maybe through the buyout but maybe through other ideas. And it needs a strict, consistent and cost-effective way to drive demand for its product(s). The latter is where Storj has failed over the years.

Opportunities must be identified and solutions must be offered. I have been saying this basically from day 1, for example here: Has Storj ever looked into the cctv market?

With donated engineering work and strict path to production, this might create focused attention and opportunities to attract new customers.

However I see at least 1 major risk that you did not address: Despite your suggestion to raise prices which might lead to a mass exodus of existing customers (just imagine Vivint leaves) we don’t know the reaction of potential customer to the fact that Storj is bankrupt. What I am saying is, if it was already hard in the past to find new customers, the risk that it is even harder now with a history of bankruptcy needs to be at least mentioned.

And there is still the SOC2 issue. Select might be even dropped if the global network has propper certification.

So it seems the truth is: Storj needs much more money than the loan currently discussed. And Storj needs a path to attract new customers. Anything else will not be sufficient to steer this project into a successful future.

How about the 3.7M tax dept? Chapter 11 is not a tax amnesty. :thinking:

Actually I was referring to

to close the gap.

@Roxor @jammerdan @alpharabbit

Thanks — this is much closer to the discussion I hoped the thread would produce.

@Roxor

To be precise, I have not softened the premise. To put it like a politician: I condemn, in the strongest possible terms, any reduction in SNO compensation as part of the restructuring.

We agree that it cannot be “the fix.” We disagree on whether it is a useful small component.

A 20% storage-rate reduction would generate approximately $22,800 in gross monthly savings. Before calling that useful, Storj would need to calculate the net result after operator departures, repair traffic, lost diversity and the future cost of rebuilding capacity.

Without that analysis, it is a gross saving—not a demonstrated net benefit.

Your nuance about unused capacity is fair: empty capacity costs no payout today, but it has option value and may be expensive to rebuild later. That is why excess capacity should be managed through onboarding, vetting and targeted placement—not by lowering compensation for customer data already stored.

I appreciate the clarification and the constructive response.

@jammerdan

Agreed. Even if the full $388,000 DIP becomes available, the filed budget leaves only $8,875. It is a bridge, not a turnaround.

Your point about bankruptcy affecting customer confidence is important and should be added to the plan. The churn model covers price sensitivity, but Chapter 11 creates a separate risk:

  • existing customers may leave because they fear interruption and data loss;
  • new customers may delay procurement;
  • enterprise legal departments may reject new contracts;
  • and competitors will use the bankruptcy in sales discussions.

Storj should therefore stress-test:

  • zero new customer growth for six months;
  • 20% weighted usage loss;
  • loss of the largest customer;
  • and simultaneous reductions by the three largest accounts.

Material customers should not receive a blind across-the-board increase. They need individual analysis based on contribution margin, contract terms and retention risk.

The initial sales focus should also be on workloads where Storj is not the customer’s only copy: secondary backups, archives, disaster recovery, replicated media and staged migrations. That reduces the trust barrier while still creating revenue.

Your CCTV example fits the product-led approach well—but only with a paying design partner, defined requirements and a strict path to production.

SOC 2 and Select also need a clear decision gate. Select should continue only if it satisfies customer requirements that Global cannot meet and has a credible positive-margin path. If Global can obtain the necessary controls more efficiently, overlapping infrastructure should be consolidated.

@alpharabbit

Correct. The approximately $1.12 million figure is only the derived annual operating gap. It is not the total restructuring requirement.

The figures must remain separate:

  • approximately $1.12 million annual operating gap;
  • fresh turnaround and exit capital;
  • approximately $8.076 million across the listed Inveniam, Treasury and SNO claims;
  • plus DIP principal, interest and other liabilities.

Chapter 11 does not make the Treasury claim disappear. But the petition alone also does not tell us the final allowed amount, priority, payment schedule or court-approved treatment.

That is why the proposed $4–5 million exit-capital target works only if the major claims are restructured rather than paid immediately in cash. If more near-term cash is required, the financing target must increase.

So the revised conclusion is:

Storj needs much more than the current DIP, and probably more than a $1 million investment. It needs positive recurring operations, sufficient exit capital, a credible treatment of legacy claims and a customer-confidence plan for selling while in Chapter 11.

That additional commercial risk makes stable infrastructure more important—not less. Creating another source of uncertainty through lower SNO compensation would move in the wrong direction.

Here I would like to correct myself: even this is still too technical. Storj should provide signed one-click installers—a .exe for Windows, a .dmg for macOS, and native NAS/Unraid packages.

The user flow should be:

Download → install → sign in → select Plex, Jellyfin or Emby and the media library → done.

Bucket creation, credentials, mounting, caching, startup and updates must all happen automatically. Anyone who can install Plex should be able to connect it to Storj without opening a terminal or understanding object storage. Let’s call it a Storj Toolbox with Google Drive Sync etc.

AFAIK tax claims are highest priority and reduction is (almost) impossible.

You may well be right that tax claims are extremely difficult to reduce. My understanding is only that the court-approved DIP financing would normally rank ahead of ordinary prepetition claims, while the tax debt would still need to be dealt with through the restructuring plan rather than simply disappearing.

So my $4–5 million estimate assumes that the tax claim can at least be paid over time and that the other major claims are restructured. If a large part of the tax debt must be paid immediately, the required financing would obviously be much higher.

I may be oversimplifying the exact ranking and treatment here. I can look into it more precisely.

you looks like mixt the product, Storj is not end user product. Storj is backbone product for software solutions like dropbox itself.

@Vadim I am not confusing the product.

I am not proposing that Storj become Dropbox. I am proposing that Storj remain the storage backbone while providing finished integrations that make it easy for software, hosting companies and users to adopt that backbone.

More importantly, the court filings do not support the idea that Storj has successfully operated only as an enterprise backend. They describe Storj as serving primarily individual users and only a few enterprise clients. They also state that the expanded enterprise-sales effort failed to generate revenue fast enough, while Object Mount consumed considerable resources without gaining sufficient traction.

So “Storj only wants large software solutions” is a strategic preference—not a business model that has already worked.

Dropbox itself did not succeed by exposing storage infrastructure and waiting for other developers to build the customer experience. The backend and the user-facing integration are different layers, but Storj currently needs both.

The proposal is not to abandon S3 or the backbone product. It is to make that backbone dramatically easier to integrate and sell.

Fully agree.
Storj needs to re-establish trust so that potential customers can become customers with confidence. This will be difficult I’d imagine.

Agreed. That could be a path to overcome trust issues.
Also free migration services or support with the migration process might reduce cost and friction of implementing an additional risky storage provider for a customer.

In any case it requires to approach potential integration opportunities pro-activeley and assess whether it can be a good fit and bring substantial revenue.

It is complex: Chapter 11 Corporate Bankruptcy Reorganizations and Tax Controversy: A Primer | Tax Executive

Secured Claims—Federal Tax Liens. Claims secured by federal tax liens imposed prior to bankruptcy are considered secured claims and normally must be fully paid to the extent they are allowed and secured, though payment may be deferred under the plan.

This is not a restructuring plan. It is a predetermined demand—“do not reduce my compensation”—buried beneath an industrial quantity of tables, pseudo-precision, product brainstorming, arbitrary deadlines, unsupported growth targets, and imperative sentences.

The core methodology is backwards. You begin with a “non-negotiable premise,” declare one category of expenditure permanently untouchable, and then construct a financial narrative designed to validate that premise. That is advocacy, not analysis. In an actual restructuring, no major cost category gets to declare itself metaphysically exempt before the company’s unit economics, contractual obligations, customer cohorts, marginal costs, and available financing have even been established.

Calling SNO compensation a supplier cost does not prove that the current price is optimal, affordable, or permanently immutable. Supplier costs are renegotiated constantly, especially when the buyer is insolvent. “The network needs suppliers” is not equivalent to “every element of the current supplier pricing schedule must remain unchanged forever.” That is a category error dressed up as an economic law.

The SNO-cut analysis is structurally biased

Your savings table is engineered to produce the desired conclusion.

You compare isolated storage-payout reductions against the entire estimated corporate funding gap and then announce that the reductions are “insufficient.” Of course they are insufficient. Nobody serious would expect one line item to solve every sales, debt, payroll, product, legal, and financing problem simultaneously. A restructuring consists of multiple cumulative measures. A cost reduction does not become invalid merely because it cannot single-handedly rescue the whole company.

By this logic, nearly every individual cost-saving measure should be rejected:

  • A payroll reduction that saves only $300,000 does not close the whole gap.

  • A software-contract reduction that saves only $100,000 does not close the whole gap.

  • A professional-services reduction that saves only $200,000 does not close the whole gap.

  • Therefore, apparently, none of them should happen.

That is obviously absurd. In a distressed business, $137,000, $274,000, or $684,000 per year is not “immaterial.” Those amounts may be insufficient in isolation, but they can still be material components of a viable package.

You also repeatedly slide between “reducing one payout rate,” “eliminating all SNO compensation,” and “abolishing the service,” as though these were economically adjacent proposals. They are not. A modest adjustment, a redesign of incentives, a differentiation between capacity classes, and total non-payment are radically different scenarios. Treating them as points on one inevitable road to network destruction is rhetorical catastrophizing, not modeling.

The claim that there is “no economically useful SNO-cut range” is not demonstrated. To demonstrate it, you would need actual operator-level supply curves, elasticity estimates, geographic redundancy requirements, concentration data, replacement-capacity costs, failure probabilities, churn behavior, and sensitivity by operator type. You explicitly acknowledge that this information is unavailable—and then state a universal conclusion anyway.

That is not evidence-based reasoning. It is epistemic overreach.

The 70% customer price increase is spreadsheet fan fiction

The plan’s supposed alternative to reducing supplier costs is essentially: increase prices by 70%, assume only modest usage loss, and declare the problem largely solved.

This is not a turnaround strategy. It is changing one cell in a spreadsheet.

Your churn table does not model customer behavior in any credible way. It takes the existing billings, multiplies them by 1.7, subtracts an assumed percentage of usage, and calls the result a stress test. That is arithmetic, not market analysis.

It ignores, among other things:

  • Differences between account churn, usage contraction, and complete workload migration.

  • Customer-specific price sensitivity.

  • The possibility that customers remain temporarily but stop expanding.

  • Delayed migration followed by concentrated churn.

  • Contractual repricing restrictions.

  • Discounts, credits, minimum commitments, collection failures, and negotiated exceptions.

  • Competitive responses.

  • Bankruptcy-related procurement restrictions.

  • Damage to customer confidence.

  • The fact that new prices do not instantly become collected cash.

  • The possibility that the least price-sensitive customers already negotiated lower enterprise rates.

The precision of “17.21% weighted usage loss” is particularly misleading. There is no empirical basis in the post capable of supporting two decimal places. It is the output of a simplified equation whose assumptions are doing virtually all the work. Numerical precision is not analytical validity.

The proposal simultaneously says pricing changes must be orderly and lawful, acknowledges that fixed contracts can only move at renewal, and demands a $500,000 collected-MRR run rate by Day 30. Those positions do not cohere. You cannot responsibly reprice a complex customer base, navigate contractual limitations, observe churn, collect the new invoices, and validate a sustainable run rate within thirty days merely because a timeline says so.

This is deadline cosplay.

The operating model is composed almost entirely of arbitrary targets

The fixed-cost cap of $300,000 per month is not derived from an operating design. It is selected because, when combined with a 40% COGS assumption, it makes the break-even equation produce a convenient $500,000 MRR target.

There is no department-level cost model. No proposed headcount. No identification of critical versus removable roles. No transition expenses. No severance assumptions. No explanation of which infrastructure costs are genuinely fixed. No analysis of support requirements under the proposed consumer-facing expansion. No reconciliation between the proposed seven-person product team and the cost cap.

Similarly, the Year 2 and Year 3 targets are merely ascending numbers in a table. Nothing in the post establishes a defensible relationship between the proposed integrations and $1.5 million in exit MRR. The figures are labeled “targets—not forecasts,” which is convenient because it allows them to look like a financial plan without accepting the evidentiary burden of being one.

A target unsupported by a bottom-up acquisition model is just a wish with formatting.

Where are the expected leads per distribution channel? Conversion rates? Activation rates? Average revenue per user? Gross retention? Net retention? Sales cycles? Support costs? Partner economics? Migration costs? Customer-acquisition costs? Payback periods? Cohort margins?

They are absent because the post does not actually model a business. It models a preferred outcome.

“Storj Connect” is uncontrolled scope expansion masquerading as focus

The post correctly criticizes unfocused product spending and then proposes, during Chapter 11, that Storj build:

  • A new integration control plane.

  • New provisioning APIs.

  • Credential-management infrastructure.

  • OAuth and device-code flows.

  • Migration orchestration.

  • Billing estimation.

  • Health monitoring.

  • Webhooks.

  • Terraform support.

  • SDKs in six languages.

  • A cross-platform local agent.

  • FUSE and WinFsp support.

  • SMB and WebDAV access.

  • Read and write caching.

  • Offline pinning.

  • NAS packages.

  • A WordPress media and backup product.

  • Plex, Jellyfin, and Emby integrations.

  • Hosting integrations.

  • Backup integrations.

  • An integration marketplace.

  • An AI data plane.

And much of this is supposed to reach alpha or beta within ninety days.

This is not ruthless prioritization. It is a product fever dream.

Every “one-click” integration hides substantial complexity: authentication, permission scoping, secret rotation, upgrade compatibility, filesystem semantics, cache invalidation, data consistency, corruption recovery, multipart-upload recovery, platform packaging, security review, performance tuning, observability, documentation, user support, and long-term maintenance.

The WordPress proposal alone is not a small plugin. It combines media offloading, URL rewriting, migration, backups, database handling, restoration, custom domains, CDN behavior, Multisite support, WooCommerce compatibility, page-builder compatibility, resumable jobs, diagnostics, and billing estimates. Claiming private alpha in thirty days and production in ninety days does not make the schedule ambitious. It makes it non-credible.

The Plex proposal is equally questionable. Consumer media workloads can be egress-heavy, latency-sensitive, cache-sensitive, support-intensive, and highly price-sensitive. The post provides no evidence that this segment produces attractive contribution margins. It simply notices technical compatibility and leaps directly to commercial opportunity.

Compatibility is not demand. A conceivable integration is not a market. A GitHub repository is not a distribution strategy. An app-store listing is not customer acquisition.

The financing proposal consists of ordering investors to provide $5 million

“Storj should raise $5 million” is not an exit-financing strategy.

Who is expected to invest? At what valuation? With what liquidation preference? Against what security? Based on which credible growth case? What happens to existing equity? What return can the investor plausibly earn? Why would new money fund an insolvent company whose proposed plan protects one supplier group absolutely, imposes a major price increase on customers, and launches an enormous new development roadmap?

The post supplies a detailed use-of-funds table but almost no capital-formation thesis. This is capital allocation without capital availability.

The milestone language does not solve that problem. Investors do not materialize because a forum post has assigned percentages to “transition reserve,” “customer migration,” and “contingency.” The difficult question is not how to spend $5 million. The difficult question is why anyone should provide it.

That question is almost entirely unanswered.

The competitive analysis is selective and self-defeating

The post emphasizes how much cheaper Storj would remain than AWS and Azure while acknowledging that Backblaze B2 and Cloudflare R2 can be cheaper or structurally more attractive for relevant workloads.

That is not a minor footnote. Those are precisely the kinds of competitors a price-sensitive storage customer is likely to evaluate. Beating premium hyperscaler list pricing does not establish product-market fit.

The proposed answer is that Storj will win through distribution, durability, security, predictable billing, performance, migration, and easier integration. That is a list of desirable nouns, not a competitive strategy. Several of those claims would require evidence, investment, operational maturity, and customer validation. The post simply declares them to be the future differentiators and then assigns a ninety-day delivery schedule.

Again: imperative mood is not execution capacity.

The failure-domain calculation is false confidence rendered in scientific notation

The operator-failure illustration may be the weakest part of the entire post.

First, the post correctly argues that public node and subnet statistics cannot reveal actual operator concentration, placement, correlation, or repair impact. It then creates a simplified hypothetical with uniform distribution and derives a probability of approximately (2 \times 10^{-11}) per segment.

You cannot condemn other conclusions for lacking placement-weighted internal data and then produce an astronomically precise risk estimate from an invented uniform model.

The assumptions erase most of the problem:

  • Nodes are treated as though they are evenly represented in placement.

  • Operators are assumed to distribute infrastructure evenly.

  • Host and storage failures are assumed independent in convenient ways.

  • Geographic, ASN, provider, control-plane, billing, legal, and administrative correlations are omitted.

  • Simultaneous operator behavior is reduced to a simplified piece-loss calculation.

  • Repair pressure, repair concurrency, bandwidth availability, and prolonged outage duration are not modeled.

  • A visible /24 is implicitly treated as a meaningful proxy for independent failure-domain diversity.

Once those assumptions are inserted, the tiny probability is largely predetermined. The scientific notation adds aesthetic authority, not reliability.

This is unearned precision.

The conflict of interest is not incidental

The plan’s supposedly universal principle eventually becomes a set of explicit conditions for the author’s own contribution:

  1. SNO compensation may not be reduced.

  2. Existing professional multi-node deployments must receive protection or grandfathering.

That makes the earlier analysis look less like neutral restructuring work and more like an elaborate negotiating document for the author’s own economic position.

There is nothing inherently illegitimate about defending one’s interests. But it should be labeled honestly. The post is not an independent evaluation of all available options. It is a proposal from an interested supplier saying, in effect:

“Do not reduce my rates, do not impair my deployment model, increase customer prices instead, raise outside capital, and build the product portfolio I think should exist. In exchange, I may contribute some development labor.”

That is a stakeholder proposal. It is not a comprehensive turnaround plan.

The donation offer also needs to be kept in perspective. “Five-figure” development work is not remotely commensurate with the proposed product scope, financing need, maintenance burden, support obligation, or restructuring risk. Donated code can even create additional liabilities if it lacks long-term ownership, architectural alignment, security review, documentation, and accountable maintenance.

Software is not free merely because the initial commits are unpaid.

The central contradiction

The post repeatedly demands that management make difficult, evidence-based, commercially disciplined decisions—except regarding the one cost category the author wants permanently protected.

Everything else must be repriced, cut, capped, restructured, converted, frozen, discontinued, subordinated, stress-tested, or sold.

SNO rates alone are declared sacred.

That is not principled financial analysis. It is selective rigidity.

A credible restructuring plan would evaluate SNO compensation alongside every other cost, distinguish between nominal rates and total network economics, model operator elasticity, examine concentration and geographic requirements, identify where incentives are over- or under-paying, and consider targeted rather than indiscriminate changes. It would neither assume that cuts are harmless nor decree that every current rate is untouchable.

This post does neither. It replaces analysis with a covenant.

In summary, the document has the visual grammar of a serious plan—tables, milestones, formulas, scenarios, architecture diagrams, capital allocations, and failure triggers—but very little of the empirical foundation required to justify its conclusions.

It is excessively specific where the author lacks data and strategically vague where real evidence would be inconvenient.

It treats assumptions as findings, targets as forecasts, product ideas as validated demand, integrations as distribution, price increases as collected revenue, desired financing as available financing, simplified arithmetic as risk modeling, and personal commercial conditions as universal principles.

The result is not a 36-month restructuring plan.

It is a sprawling, technically literate, numerically decorated ultimatum.

could you please ask your LLM to be more concise and summarize the main ideas in 1-2 paragraphs. It seems two AIs discussing about Storj survival plan atm

Your message is way to short. The LLM will not be able to understand that. First objective in this thread is burning LLM tokens. So make your message longer!

I’m gonna need to rent some GPU compute from Valdi to run my own model: I can’t afford to participate in the current token wars :slight_smile:

Before I answer the substance, where you see me typing since one hour and yes, it becomes long, one thing about the repeated LLM comments.

I cannot speak for Ambifacient, but I can explain my own process. I write the arguments, calculations and replies myself in German. English is not my native language, so I then use ChatGPT to translate them and clean up the formatting. I read and edit the result before posting it. The position, reasoning, examples and decisions about what to say are mine.

Anyone following the discussion can see that I am responding to specific points, correcting mistakes and refining the proposal as the discussion develops. I have put a considerable amount of time and effort into this thread.

I honestly do not know when we reached the point where a lazy two-line opinion is accepted as obviously human, while a detailed contribution someone has worked hard on is dismissed as “two AIs talking.”

The jokes were funny once. At this point, I find the repeated dismissal unpleasant. Criticize the length, challenge the numbers or dismantle the argument—that is completely fair. But saying that an LLM wrote it is not a rebuttal. Using a tool for translation and formatting does not make that tool the author.