अंतरिक्ष / ANTARIKSH Cloud
Get an API key
Built on Indian carrier infrastructure

A full cloud stack, run on Indian soil, billed in rupees.

IAM, VPC, managed databases, object storage, monitoring and billing: the same depth you'd expect from a hyperscaler, sitting on compute and network capacity leased from Tata Communications, Reliance Jio and Yotta Data Centers. No discovery call, no proposal deck, no pricing "on request."

$ curl -s antariksh.cloud/v1/regions | jq .latency_ms
  { "bom-1": 4, "maa-1": 6, "del-1": 9, "blr-1": 5 }
Live backbone, indicativeMulti-carrier
IAX cable SEA-ME-WE del-1 · Delhi NCR bom-1 · Mumbai blr-1 · Bengaluru hyd-1 · Hyderabad maa-1 · Chennai
Compute region Cable landing station Backbone: Tata Communications, Jio, Yotta
What actually changes

The part everyone hates about buying cloud, removed.

Hyperscalers built India regions and kept the US enterprise sales motion. We skipped the motion.

onboarding.log

// typical hyperscaler, India region

01"Submit a request," routed to a partner reseller
02Discovery call scheduled: 6 business days out
03Solutions architect sizes a "custom quote"
04Pricing withheld pending "commitment tier"
05MSA redlines routed through US legal, +3 weeks
06Invoice lands in USD, GST handled separately
07Data residency confirmed only after signature

// antariksh.cloud

01Sign up with a company PAN, verify by OTP
02Card or NEFT, instance provisions in ~90s
03Rate card is the pricing page: no floor, no quote
04Every region is Indian soil by construction
05Standard ToS, no MSA negotiation below ₹10L/mo
06GST invoice, itemised, on the 1st of each month
07Talk to a human only if something's actually broken
The infrastructure underneath

We didn't build a network. We plugged into three.

Antariksh runs on capacity leased across Tata Communications, Reliance Jio and Yotta Data Centers: subsea cable systems, carrier backbones and data-centre floors that already carry a large share of India's internet routes.

6
Compute regions at launch: Delhi NCR, Mumbai, Pune, Bengaluru, Hyderabad, Chennai
3
Carrier partners feeding peering into every region: Tata Communications, Jio, Yotta
2
Subsea systems on-net at launch: India Asia Xpress, SEA-ME-WE 5
Tier III
Minimum data-centre tier, N+1 power and cooling on every floor
Console

The same depth you'd expect from a hyperscaler.

Compute is the entry point, not the whole product. Everything below ships from day one, with the same account and the same bill.

IAM
Users, roles and scoped API keys. Policy-based access down to a single bucket or instance.
VPC
Isolated networks per project, subnets per region, security groups at the instance level.
Database
Managed Postgres and MySQL, automated backups, read replicas across regions.
Monitor
Metrics, logs and alarms per resource. The CloudWatch equivalent, included, not upsold.
Object storage
S3-compatible API, so existing SDKs and tooling point at it without rewrites.
Billing
Line-item usage in real time, budgets and alerts, one GST invoice per account.
What gets built on top

A platform layer, not just a rack of VMs.

Deploy from a git push, get a subdomain and a build pipeline, and resell that experience to your own users under your own brand. If you've used Vercel or Render, the workflow will feel familiar; the difference is what's underneath it.

The vision

Indian cloud today is built in pieces. We're building the whole stack.

Yotta owns the GPU story. Vayu owns the enterprise network story. CtrlS and Sify own colocation. Nobody has assembled a single platform with the breadth of AWS and the sovereignty of a domestic operator. That gap is the plan, worked in three phases.

01
2026, in progress

Anchor on GPU and reserved compute

Ride existing capacity rather than build it: reserved GPU and general compute from Tata Vayu and comparable partners, with our own control plane, IAM, billing and monitoring wrapped around it from day one. Target buyers are AI teams that need training and inference capacity in India, priced and provisioned without a sales cycle.

Compute + GPUIAMVPCBillingMonitor
Liveself-serve console, 6 regions
02
2027

General-purpose IaaS and PaaS depth

Managed databases, object storage, Kubernetes, and the deploy-from-git platform layer that lets other companies build and resell products on top of us. MeitY empanelment and sector-specific compliance (BFSI, government) so regulated workloads have a real reason to move off the hyperscalers.

Managed databasesObject storageKubernetesDeploy-from-gitMeitY empanelment
Plannedcapital + partner-dependent
03
2028 and beyond

Full-depth hyperscale platform

A marketplace of third-party services, edge and 5G integration through telco partners, and dedicated capacity blocks where reserved leasing stops making sense. The measure of success isn't matching AWS's 200-service catalogue line for line, it's being the default answer whenever an Indian company asks where to run.

MarketplaceEdge / 5GDedicated capacityMulti-region failover
Ambitionown the full stack
Rate card

The price is the price.

Per-second billing, no reserved-instance maze. Everything below is what a self-serve account pays today; nothing is "contact sales."

Line itemSpecRegionRate
c1.small
Compute, general purpose
2 vCPU / 4 GBbom-1₹3.10 / hr
c1.large
Compute, general purpose
8 vCPU / 16 GBbom-1₹11.40 / hr
g1.infer
GPU, inference, single A10
1×A10 / 24 GBdel-1₹98.00 / hr
block storage
SSD, replicated 3×
per GB / moall₹2.10
egress, intra-India
NIXI-peered traffic
per GBall₹0.00
egress, international
via IAX / SMW5
per GBall₹1.85
Prices exclude GST. Billed monthly by usage-second; no minimum commitment, no early-termination clause.
Data stays where it's provisioned

Sovereign by geography, not by marketing slide.

Every Antariksh region sits inside a Tier III+ Indian data centre, operated by Tata Communications, Jio or Yotta. Nothing routes through a foreign parent's control plane.

MeitY empanelment in progress CERT-In incident reporting compliant RBI data-localisation aligned for regulated workloads SOC 2 Type II, audit scheduled Q1 ISO 27001, via partner DC certification

Compute is a utility. Buy it like one.

Provision your first instance in the console, or script it against the API. If you'd genuinely rather talk to someone first, that's fine, but it's an option, not a gate.

Create account no card required to read the docs