CITHORUM — APRIL 2026

Jam proof now.
Cithorum infrastructure next.

Cithorum is already in market. ₹1.25 Cr funds the one missing artefact — a 1 PB proof environment — that converts existing demand into revenue.

M2M proves Jam works in production. The 1 PB proof environment is the only thing standing between us and 5–10 more M2M-shaped customers.

02 — PROOF WALL

The point is not belief. It is evidence.

Nine proof points, tiered by evidence weight. Every tile is one click from its source archive.

LIVE REVENUE

$7k / month MRR

Formal per-TB / month contract — ~50% below commercial cloud

Jam in production on unattended M2M workloads. Linux delivered, MacOS access given, heartbeat dashboard live, first invoice sent.

PROGRAMME

Inception programme member

“I think I just witnessed accelerated compute”

Accepted Inception member. Ecosystem validation, not NVIDIA product certification.

QUOTE + INTEGRATION ROUTE

“What JAM has done is changed the game”

50 GB → 7 MB; reconstructed on a MacBook in < 5 min

Production deployment partner. Spark Java/Scala adapter required before licence execution.

LIVE BENCHMARK

135 GB → 17.2 GB · 7.85×

281 MB/s encode · 1.13 GB/s decode · NVMe-bound

April 2026, unedited. Watch the cut →

VETTED

Vetted by National Authority

BRAVE1 portal live status confirmed; eligible for Ukrainian defence procurement matching.

ECOSYSTEM APPROVED

Approved Raytheon supplier

Profile reviewed, approved, added to the RTX ecosystem via Supplier.io.

GRANT WON

CAD $30,000 non-dilutive

“JAM is way beyond MVP”

Award letter on file. Public quote on the partner site.

BENCHMARK

12% smaller FASTQ; 22-min alignment

Versus 2.5 hours for the next-closest aligner

Genome sequencing benchmark documented in the Cithorum Jam evidence archive.

SUPPLIER APPROVED

Corporate-buyer database

Profile approved. Visible to thousands of active corporate buyers.

03 — LIVE TEST · APRIL 2026

Unedited footage. Real workload. Real numbers.

135 GB of VM snapshots. JAM compressed 3.8×, JAM+ZSTD 7.85×. Pinned at NVMe ceilings end-to-end.

~58 second summary. Title cards → live encode → result panel.Watch the full 6:37 clip →
Original
135 GB
JAM only
35 GB · 3.8×
JAM + ZSTD
17.2 GB · 7.85×
Encode
281 MB/s · 481 s
Decode
1.13 GB/s · 118 s
RAM
< 400 MB encode · 45 MB decode
Bottleneck
NVMe-bound throughout
Operator
Lucas Marsh · standard NVMe rig

JAM is the only codec that keeps modern enterprise NVMe pinned at hardware ceilings end-to-end. Compression ratio scales with dataset size — the 1 PB proof environment will produce stronger numbers than this 135 GB test, not weaker.

04 — TODAY VS END OF ROUND

Same company. Six months apart.

Today is verified evidence already in market. End of round is the artefact set this raise produces.

TODAY · APRIL 2026

  • M2M live — $7k/month MRR, formal invoice trail.
  • NVIDIA Inception — programme member, on-record quote.
  • NATO BRAVE1 / UNITE — vetted by National Authority.
  • RTX — approved in the Raytheon partner ecosystem.
  • ZenLaunchpad — CAD $30k non-dilutive grant won.
  • SwissVault — genome alignment benchmark validated.
  • Atreides IO — 50 GB → 7 MB demo + integration roadmap.
  • JAM+ZSTD live test — 135 GB → 17.2 GB, 7.85×.

END OF ROUND · M6

  • 1 PB proof environment live, ingesting customer data under MSA.
  • 5–10 customer datasets benchmarked with restore proof.
  • 2+ paid Jam Engine customers signed beyond M2M.
  • Atreides Spark integration funded as a named workstream.
  • Expansion conversation in motion with infrastructure investors.
  • KG Ops layer in internal use on real telemetry.
  • Investor update with real pipeline + MRR + conversion data.

05 — FUNDING NOW

Raise ₹1.25 Cr. Anchor with Suraj’s ₹50L.

Six months. One proof environment. Three paths to revenue.

IMMEDIATE RAISE

₹1.25 Cr

Closes within month 2

SURAJ ANCHOR

₹50L

Confirmed

REMAINING

₹75L

Warm angel checks

USE WINDOW

6 months

Break-even target M12–18

If you anchor ₹50L, here is what changes.

Each card is a 30 / 90 / 180-day commitment we own back to Suraj in writing.

30 DAYS

  • Investment instrument signed.
  • ₹50L deployed against the Phase 0/1 plan.
  • 1 PB hardware facts confirmed: usable capacity, power, network.
  • Pilot MSA template signed and ready to push to first prospects.

90 DAYS

  • 3–5 pilots actively ingesting data on the proof environment.
  • M2M revenue reporting formal and visible monthly.
  • India / Karnataka channel active through Suraj’s network.
  • First benchmark report drafted and circulated under NDA.

180 DAYS

  • 5–10 customer datasets benchmarked end-to-end.
  • 2 paid pilots converted into paying agreements.
  • Expansion path defined: raise / convert pilots / both.
  • Investor update with real conversion data, not projections.

06 — PHASED PLAN

Seven phases. Cost. Build. Commercial. Gate. Risks.

Phase 0 is alignment. Phase 1 is Suraj’s anchor. Phase 2 is the proof. Phases 3–6 are conditionally funded.

PHASE 0 · DECISION CLARITY

Phase 0 — Alignment and control

Time
Now to 2 weeks
Focus
Decision clarity
Cost
INR 0L

Confirm route, funding structure, 1 PB access, pilot criteria, IP / customer-data rules, and Suraj's role before any serious spend.

Founder time only. Legal, structuring, IP licence drafting, diligence. No round drawdown.

Build

  • Document Suraj's role as investor, partner, data-centre channel, and scientific/commercial lead where applicable.
  • Agree entity and IP structure: where Jam IP sits and how the India route licenses it.
  • Confirm 1 PB server facts: location, hardware, network, redundancy, power, remote access, operating costs.
  • Freeze near-term scope: Jam Proof Cloud, Jam Engine, metering dashboard.

Gate

  • Suraj INR 50L confirmed; instrument structure finalised in writing.
  • Investor ask set at INR 1.25 Cr.
  • No large infrastructure commitment before anchor demand.

Risks

  • Instrument debate stretches past 2 weeks. Mitigation: timebox legal to 14 days; choose simplest workable instrument.
  • 1 PB facts come back weaker on usable capacity. Mitigation: gate Phase 2 capex on confirmed usable spec, not raw.

07 — GANTT · 24 MONTHS

The first six months prove it. The next six monetise it.

Proportional time axis. Today line. M6 decision as the visual climax. Burn-vs-revenue overlay crossing at month 12–18.

NowM1M3M6M9M12M18M24
  1. Fund + alignINR 1.25 Cr + 1 PB access
  2. 1 PB Proof CloudBuild, ingest, benchmark, verify
  3. Customer pilots5–10 workloads + proof reports
  4. Jam EnginePaid deployments + integrations
  5. KG platformOps KG → first vertical
  6. Vertical KG (1st)One paid design partner
  7. Data-centre expansionAnchor demand / project finance
  8. Target break-evenTarget window 12–18 mo
  9. Conservative bufferConservative buffer 18–24 mo

Burn line crosses revenue line in the 12–18 month target window. The conservative buffer sits beyond.

08 — UNIT ECONOMICS

The 12–18 month break-even target, derived.

Operating burn ~₹9L / mo. M2M baseline + 3–4 paid Jam Engine + 1 hosted KG design partner closes the gap.

MONTHLY BURN

Engineering (3 FTE)
₹6L
Infrastructure operating
₹1.5L
GTM + ops + contingency
₹1.5L
Total
₹9L / mo

REVENUE STREAMS

M2M baseline (live)
$7k / mo
Jam Engine mid-tier
$3–4k / mo each
Hosted KG design partner
Prepaid, amortised
Funded integrations
Atreides Spark + others

BREAK-EVEN LOGIC

Monthly gross margin ≥ ₹9L. Implied: M2M baseline + 3–4 paid Jam Engine at $3–4k/mo + 1 hosted KG design partner closes the gap. Achievable at the Phase 3 commercial conversion gate.

  • Optimistic — break-even at 12 months. Two Jam Engine by M9; third + KG design partner by M12.
  • Base — break-even at 15–18 months. Conversion lags one to two quarters.
  • Conservative — break-even at 18–24 months. Bridge angel cash extends runway.

09 — STACK + KG

Jam is the economic layer. The KG is the compounding software layer.

One proof engine. Two revenue tracks. One conditional infrastructure expansion.

INPUT

Customer datasets

Logs · VM images · FASTQ · M2M / IoT · Game assets · AI datasets

LAYER 1 · LIVE

Jam Engine

Compression, indexing, keyless envelope, reconstruction. Customer hardware stays intact.

LAYER 2 · LIVE

Telemetry + Metering

Original size, Jam size, restore status, CPU, throughput, invoice basis. The seed of every KG product.

LAYER 3 · INTERNAL FIRST

Knowledge Graph Core

Customers, data, workloads, evidence, incidents, workflows, outcomes. Internally consumed first.

LAYER 4 · GATED ON BUYER PULL

Cithorum Platform + Verticals

Accounts, dashboards, billing export. Workspace KG (beta) · Law KG (design-partner gated) · regulated sequencing KG (deferred).

10 — GO TO MARKET

Three engines turn proof into revenue.

Jam Engine sells the proof. Hosted KG compounds it. India tender route finances scale. Each engine has a buyer, a unit price, and a channel.

ENGINE 01 · LIVE

Jam Engine

Mid-market data-heavy enterprises (M2M-shaped).

Unit
$3–4k / mo per customer · tier-based by TB
Channel
Outbound proof-led + partner co-sell (Atreides, NVIDIA Inception)
Lead indicator
M2M live at $7k/mo MRR · target 3–4 paid by month 9

ENGINE 02 · NEXT

Hosted KG

Vertical software buyers — Workspace ops, Law firms, Genomics labs.

Unit
Design-partner pricing — one-time + retainer · SaaS expansion thereafter
Channel
Design-partner-led, one vertical at a time
Lead indicator
Ops KG used internally by month 12 · first external design partner by month 15

ENGINE 03 · IN FLIGHT

India tender route

State, central, public-sector procurement — Karnataka, Keonics, defence-adjacent infrastructure.

Unit
Project-scale work orders · bankable for project finance
Channel
Suraj’s network + Chirag Labs partnership
Lead indicator
Karnataka mapping under way · infrastructure and regulated-data opportunities in qualification

CUSTOMER JOURNEY · WEEK 1 → WEEK 8

From qualify to signed MRR in eight weeks.

  1. 01

    Qualify workload

    Week 1

    Cithorum + customer

    Fit assessment + dataset characterisation

  2. 02

    Sign pilot terms

    Week 1–2

    Legal

    MSA + benchmark-rights frame

  3. 03

    Provision access

    Week 2

    Ops

    Bucket, credentials, quota

  4. 04

    Run Jam

    Week 2–4

    Cithorum

    Ingest, compress, restore, meter

  5. 05

    Report proof

    Week 4

    Cithorum + customer

    Benchmark report — anonymised public + private detailed

  6. 06

    Convert to paid

    Week 4–8

    Sales + customer

    Production contract · MRR live

ACQUISITION CHANNELS

Six channels feed three engines.

FOUNDER CONTENT

Cadence · weekly

LinkedIn, blog, Twitter/X. Topic: data-centre waste economics, Jam vs cloud cold storage, real-world benchmarks. Drives inbound + builds reputation.

PARTNER CO-SELL

Cadence · monthly

NVIDIA Inception network, NATO BRAVE1 partner directory, Atreides production deployment, RTX ecosystem. Warm intros via existing programmes.

INDIA TENDER

Cadence · per opportunity

Suraj-led. Karnataka, Keonics, central-government and defence-adjacent pipeline. Project-financed and bankable.

DIRECT OUTBOUND

Cadence · rolling

Founder-to-founder + sales-led. Mid-market data-heavy ICP (M2M-shaped). Proof-led, short cycles.

INBOUND FROM PROOF ARCHIVE

Cadence · passive

Public benchmarks, GitHub presence, search referrals. Compounds with each new proof point published.

VERTICAL DESIGN PARTNERS

Cadence · sequential

One vertical at a time (Ops → Workspace → Law → Medical). Paid design-partner per vertical anchors the next SaaS launch.

VERTICAL EXPANSION

Each vertical compounds revenue. ARR ×2 per vertical.

JAM DEPLOY · TODAYLive revenue engineOPS KG · M12 · INTERNALCompounding seedWORKSPACE KG · M15+Beta · ops-adjacentLAW KG · M18+Design-partner gatedMEDICAL SEQ KG · M21+Regulated-gated×2 ARR×2 ARR×2 ARR

11 — CAPITAL ROUTES

Each capital source unlocks a different stage of the same company.

Anchor and angels now. Customers and tenders next. Project finance and platform investors after the proof clears.

NOW · M0–M1

Suraj + India route

₹50L anchor confirmed. Brings 1 PB access, partner credibility, Karnataka / state channel, and clinical-science network.

NOW · M0–M3

Warm angels

₹10–50L checks to complete the round.

NEXT · M3–M9

Paid pilots + prepaids

Customers fund proof through MRR, setup fees, paid pilots, annual prepayments. Reduces dilution.

NEXT · M6–M12

Infrastructure partners

Storage OEMs, data-centre operators, cloud companies funding integrations (e.g., Atreides Spark).

LATER · M9–M18+

Tender finance

Indian government / state work orders unlock project finance for larger infrastructure — outside this round.

LATER · M12+

SaaS / KG investors

Invest once Jam proof shows infrastructure moat and KG verticals show software upside.

DEFERRED — ON PURPOSE

What we will not build yet.

  • Customer-managed VMs
  • Public Docker hosting
  • Terraform provider
  • Serverless functions
  • Kubernetes-as-a-service
  • Global edge network
  • Public developer cloud
  • Regulated medical workflows before controls
  • 10–100 PB capex from common equity in this round

OPEN QUESTIONS — OWED TO SURAJ

Decisions still on the table.

12 — POTENTIAL OUTCOMES

Four credible paths. Each one is a different shape of win.

The Cithorum Group is a tender-led commercial business with real revenue. It also has four asymmetric pathways to a >$100M outcome over the 12–18 months that follow this round. They are not mutually exclusive. They are not promises. They are the upside that justifies the build.

PATH 1 · INDIAN TENDER PIPELINE

Karnataka 1,000 MW & state IT contracts

Karnataka's sustainable data-centre programme, NIC framework agreements, and the broader eprocure.gov.in pipeline. Single tender awards range from low-crore to multi-hundred-crore over 3-year horizons.

Lead
Suraj (COO)
Horizon
M3–M18
Shape
Recurring revenue · 3-year contracts
Ticket
₹1–100 Cr per tender

PATH 2 · NON-DILUTIVE GRANTS

BIRAC & biotech grant pipeline

BIRAC, India Innovation Fund, Karnataka biotech grants, and equivalent non-dilutive instruments targeting the Sequencing KG and broader genomics-archive workloads. Pre-vetted regulatory and IP credibility.

Lead
Suraj (COO)
Horizon
M3–M12
Shape
Non-dilutive · zero equity
Ticket
₹50L – ₹5 Cr

PATH 3 · CANADIAN ITB

Prime-contractor route via CANSEC 2026

Canada's Industrial and Technological Benefits programme requires foreign primes to invest in Canadian-domiciled technology. Tarig and team are positioned at CANSEC (May 2026). A successful prime engagement on Cithorum Canada's Jam Engine surface produces non-dilutive multi-year contracts.

Lead
Lucas + Tarig (Canada)
Horizon
M1–M12
Shape
Non-dilutive · multi-year
Ticket
CAD 100M+ over contract life

PATH 4 · STRATEGIC ACQUISITION

DDN — and a credible second bidder

M2M TechConnect's live deployment is the peer-reviewed validation gate. If the digital-twin workload validates at scale, DDN becomes a credible acquirer for the Jam Engine. A prime engagement in Canada makes NVIDIA a plausible second bidder; the bidding-war scenario follows from there.

Lead
Lucas (Jam Engine) + Nikos
Horizon
M6–M18
Shape
Liquidity event
Ticket
USD 100M+

None of these four outcomes is contractually promised. They are the realistic upside paths the Cithorum Group is actively building toward. Path 1 pays the bills and funds growth; Paths 2 and 3 are non-dilutive accelerants that don't consume runway; Path 4 is the asymmetric exit. The build plan in sections 04–11 holds whether or not any of these four outcomes lands — but each one materially compounds the others if it does.

DETAIL · HARDWARE

1 PB Proof Environment — Hardware Specification

1 PB usable

A software-controlled storage rig sized to demonstrate Jam on real customer datasets under MSA. S3-compatible endpoint via MinIO or Ceph gateway, isolated buckets, Postgres-backed metering for every byte ingested and restored.

Capacity is the single most contested number on the page. Usable capacity must be confirmed at 1 PB after RAID, parity, dedupe headroom, and Jam working space. Raw exceeds 1 PB; the usable figure gates capex release.

The reference 1 MW-equivalent comparator: with Jam, a greenfield site delivers the same workload at ~720 kW IT draw, ~14 racks, ~$8–10M rack-level CapEx, and a 180 kWh VRLA bridge — versus 1,000 kW, 28 racks, $14–17M, and 250 kWh LFP without Jam. Annual energy avoided: ~3.87 GWh.

Usable capacity
1 PB (confirmed after RAID/parity/Jam working space)
Raw capacity
Exceeds 1 PB; specific figure pending Phase 0 hardware discovery
Storage topology
ZFS (single-server) or Ceph (clustered) depending on hardware shape
S3 endpoint
MinIO or Ceph S3 gateway, isolated buckets, per-customer credentials
Network
100 GbE class internal; static IP, DNS, TLS
Power
Tier III-like N+1 intent; UPS + diesel genset; PUE ~1.33–1.40
Cooling
Mechanical plant sized to IT heat rejection; ~33–40% facility overhead
Redundancy
Quorum writes 3-of-5 default; 2-of-3 supported for edge
Durability target
RAMOPS-64 enterprise tier: 99.999999999% annual
Availability target
99.9% monthly control plane (managed deployments)
Restore latency
≤ 500 ms p95 standard tier; ≤ 4 hours p95 cold tier
Reference rig
2× AMD EPYC 7543, 512 GB RAM, 4× 7.68 TB NVMe, 100 GbE

DETAIL · JAM TECHNICAL PLAN

Jam Technical Implementation Plan

7.85×

Jam is a software data plane for large, changing datasets — a layer below applications and storage services that compresses, deduplicates, indexes, envelopes, stores, transfers, and reconstructs data while keeping customer hardware and applications largely intact.

RAMOPS uses content-addressed immutable objects keyed by SHA-256, an append-only deterministic layout, AEAD authenticated transport (AES-256-GCM or ChaCha20-Poly1305), and quorum replication. Session keys derive on demand via HKDF-SHA-256 from BIP39 seed material plus the handshake transcript, then discarded — no static keys at rest. Forward secrecy per session, default 15-minute rekey. PQC migration mapped to NIST FIPS 203 ML-KEM and FIPS 204 ML-DSA.

Performance envelope: up to 100× on sparse payloads, 35–75% typical on enterprise workloads, 5× average on the Scale-AI MVP, up to 400× faster than rsync on the same MVP, up to 1000× faster AI/ML loading on compatible pipelines. The April 2026 live test confirms 3.8× JAM only and 7.85× JAM+ZSTD, NVMe-bound throughout.

Components

Jam worker / gateway
Ingest, compress, index, envelope, write to RAMOPS; large-file retry queue.
RAMOPS networked storage
Content-addressed object store, append-only log, quorum replication.
JSL primitives
DeriveSessionKey, PutObject, GetObject, SignRelease, VerifyRelease, SelfTest.
jamd daemon
Host-side managed service; systemd; peer discovery via VLAN/VPN.
jam CLI
Operator surface — verify, seed init, upgrade, offline verify.
Postgres ledger
Account, bucket, quota, TB ingested, TB saved, restore status, invoice basis.
TechConnect transport
Authenticated M2M for OEM, embedded, unattended workloads.
SaaS Archiver mode
Managed immutable archive, WORM retention, signed audit records.
S3 gateway
MinIO or Ceph S3-compatible endpoint, per-customer isolated buckets.
Spark Java/Scala adapter
Roadmap — required for Atreides Spark integration; not yet shipped.

Supported OS

Ubuntu 22.04/24.04 LTS x86_64/ARM64; RHEL/Rocky 9; Debian 12; Windows Server 2022 via WSL2; hardened Yocto/DM&V ARM64 by request.

DETAIL · BENCHMARK + SOURCES

Benchmark + Source-Truth Register

19 claims

Every public claim on the v2 site traces to one canonical entry below. Anything labelled ‘historical’ or ‘roadmap’ carries an explicit qualifier in copy.

Methodology: 64 TB aggregate corpus across five workload classes (sparse telemetry, mixed log, document, binary-heavy, pre-compressed), 12.8 TB per class, on a 2× EPYC 7543 / 512 GB RAM / 4× 7.68 TB NVMe / 100 GbE rig running Ubuntu 22.04 with io_uring. Median of five runs reported.

Claims to sources

M2M $7k/month MRR (formal)
Live invoice trail + Lucas WhatsApp 2026-04-29
M2M per-TB / month rate (~50% below commercial cloud)
Cithorum-M2M-TechConnect-SLA
NVIDIA Inception membership
Reception invitation + badge
NATO BRAVE1 / UNITE vetting
NATO Communications and Information Agency email
RTX / Raytheon ecosystem approval
Supplier.io profile acceptance email
ZenLaunchpad CAD $30,000 grant
ZenLaunchpad award letter
Atreides 50 GB → 7 MB demo
Atreides demo recording + public partner quote
SwissVault 12% smaller FASTQ
Cithorum Jam evidence archive + user note 2026-04-25
T-Mobile corporate-buyer DB
T-Mobile approval email
JAM+ZSTD 7.85× live test
April 2026 footage; raw run output on Zoho
Up to 100× on sparse payloads
MASTER-REFERENCE; Jam-Networked-Storage v2.1
35–75% typical reduction
MASTER-REFERENCE
5× average on Scale-AI MVP
MASTER-REFERENCE; Scale-AI scope
Up to 400× faster than rsync
MASTER-REFERENCE; Scale-AI scope
Up to 1000× AI/ML loading
MASTER-REFERENCE; commercial licensing note
~50% storage footprint reduction
MASTER-REFERENCE
~40% compute + bandwidth reduction
MASTER-REFERENCE
Ubuntu VM 45% smaller via Jam
User benchmark note 2026-04-25

Retired or qualified

  • 48,900× claim — retired; not used externally.
  • 140× RAM reduction — retired; not used externally.
  • Universal 100× compression — qualified to ‘sparse / highly compressible payloads’ only.
  • Apache Spark integration — roadmap, not shipped; presented as funded workstream not finished feature.

DETAIL · KNOWLEDGE GRAPH

Knowledge Graph — Schema and Ontology

4 layers

The KG is the central nervous system of Cithorum’s product suite. Every entity is a node; every relationship is an edge. The graph encodes meaning — not just storage.

Storage: Neo4j for entity nodes and traversal, Postgres for structured ACID data per product, pgvector for embedding-based semantic search across entity boundaries. Products communicate via a namespaced event bus; events are the real-time pulse, the KG is the persistent memory.

The first KG product is the Data-Centre Ops KG — internally consumed by Cithorum to operate the proof cloud day-to-day. Vertical KGs (Workspace, Law, Medical Sequencing) are gated on buyer pull and reuse the same core.

Core entities

  • Customer / Org / Account
  • Dataset / Bucket
  • Workload / Job
  • Node / Rack / Site
  • Incident / Restore / Repair event
  • Capacity / Energy / Cost record
  • Person / Operator
  • Benchmark / Evidence record

Relationship types

  • ran_on (Workload → Node)
  • stored_in (Dataset → Bucket → Node)
  • restored_from (Restore → Object → Node)
  • incident_affected (Incident → Node | Customer | Dataset)
  • metered_by (Workload → Usage Ledger)
  • owned_by (Bucket → Customer)
  • evidence_for (Benchmark → Claim)
  • cited_in (Evidence → Investor Update | Pitch | Quote)

Provenance model

Every node carries a source-evidence pointer (canonical doc + run ID + timestamp). Every edge carries the event ID that created it. The provenance chain is queryable end-to-end so a diligence reviewer can trace ‘this claim → this benchmark → this dataset → this customer → this MSA’ without leaving the graph.

DETAIL · UNIT ECONOMICS

Unit Economics — Detailed Model

₹9L / mo

The 12–18 month break-even target is derived, not asserted. Operating burn at steady state is approximately ₹9L per month (~$11k USD).

Revenue stacks in three streams. M2M baseline: $7k/month MRR live today, formal per-TB / month contract at approximately 50% below commercial cloud list rates. Jam Engine mid-tier: ~$3–4k/month per customer, sold from the Phase 2 benchmark book. Hosted KG / pilot prepaid: design-partner cash up-front, amortised.

Scope separation is critical. ₹1.25 Cr funds the operating company through the proof window. The 10–100 PB infrastructure expansion is project-financed against bankable demand and is deliberately not in this round.

Burn table

Engineering (3 FTE)
₹6L / mo
Infrastructure operating
₹1.5L / mo
GTM + ops + contingency
₹1.5L / mo
Total monthly burn
₹9L / mo

Revenue table

M2M baseline (live)
$7k / mo · live now
Jam Engine mid-tier
$3–4k / mo / customer · Phase 3+
Hosted KG design-partner prepaid
Up-front cash · Phase 4–5
Customer prepaids + funded integrations
Atreides Spark + others · Phase 3–4

Sensitivity

  • Optimistic. Break-even at 12 months. Two Jam Engine customers signed by M9; third + KG design partner by M12.
  • Base. Break-even at 15–18 months. Conversion lags one to two quarters.
  • Conservative. Break-even at 18–24 months. Bridge angel cash extends runway.

Follow-on sizing logic. Confirmed monthly recurring revenue × 18-month runway target − customer prepaids − non-dilutive grants − project finance attached to tender. Derivable from real data, not from a market-size deck.

DETAIL · CUSTOMER ONBOARDING

Customer Onboarding Playbook

6 steps · 6 weeks

The pilot must feel simple even when the system underneath is hard. Six steps, six weeks, with a clear conversion-to-paid moment at the end.

Data terms are the silent killer of enterprise pilots. Customer retains all data rights; Cithorum retains rights to anonymised benchmark output; named publication only with explicit customer approval. Regulated workloads default to Jam Engine into customer infrastructure — Cithorum never holds the data.

  1. 1 · Qualify workload. Week 1. Cithorum + customer. Workload-fit memo: logs, M2M/IoT, VM images, FASTQ, game assets, backups, AI datasets. Confirm size, sensitivity, restore-time tolerance.
  2. 2 · Sign pilot terms. Week 1–2. Cithorum legal + customer procurement. Signed pilot MSA — quota, timebox, support level, benchmark rights, conversion path. Pre-approved template used for ~80% of pilots.
  3. 3 · Provision access. Week 2. Cithorum. Customer bucket on the 1 PB environment, API keys, S3 endpoint, upload guide, baseline measurement.
  4. 4 · Run Jam. Week 3–4. Cithorum + customer ops. Compress, index, envelope, record metadata, verify checksum, test reconstruction. 100% restore verification on sample set is gate-blocking.
  5. 5 · Report proof. Week 5. Cithorum. Per-dataset benchmark report — original size, Jam size, ratio, restore time, CPU, throughput, estimated cost avoided.
  6. 6 · Convert. Week 6. Cithorum commercial + customer decision-maker. Hosted storage, Jam Engine into customer stack, KG layer attached, annual prepaid, or infrastructure partnership. Exit options explicit; pricing on the table.