Buy a cloud server: self-run, managed or hosting
This is the order page for an AWS server: VPS-class, compute- or memory-optimized, dedicated tenancy, or GPU. Use it to pick the kind of purchase, then ask for a quote. If you are moving an existing site or app, read cloud hosting first.
Last updated 21 September 2026
What you are ordering
A cloud server here is an AWS instance, or a small Lightsail environment for simple sites, provisioned into an account that TiCON manages under the contract. The machine runs in ap-south-1 (Mumbai) or ap-southeast-1 (Singapore), depending on your design and data rules.
- VPS-class and application servers. General-purpose EC2, or Lightsail for small sites. Snapshots, a public address, and SSH or Session Manager access. You can resize later without changing vendor.
- Compute, memory and dedicated. Compute-optimized or memory-optimized families for databases and tight runtimes. Dedicated tenancy when the contract requires that the host is not shared.
- GPU. GPU instance families for training, inference or rendering. They are expensive when idle, so the order includes a start and stop plan. We quote the family that fits the job rather than a public SKU.
| Criterion | Cloud server (TiCON) | Single VPS | Shared hosting |
|---|---|---|---|
| Where it runs | AWS region chosen for the workload | One provider, one facility | Provider's shared data center |
| Growth | Resize, add instances, autoscale | Resize or add another VPS | Plan ceiling, then upgrade request |
| Invoice | Usage-based, quoted in USD | Fixed monthly plan | Usually fixed plan |
The table below compares this option with a single VPS and with shared hosting.
How to order
There is no self-serve cart. You talk to an architect, we quote in USD, then we provision. The calculator is an estimate, not an invoice.
Order a cloud server, from quote to handover: about 1 day
Standard instances are often ready the same working day after you accept the quote. GPU capacity and dedicated tenancy depend on the family and region, and we say so before you accept.
- Step 1:
Say what the machine is for.
Workload (site, API, database, GPU job), expected traffic, operating system, whether you need us to operate it, and any data-residency constraint.
- Step 2:
Pick a shape and a region.
VPS-class, compute or memory, dedicated tenancy, or GPU. The region is usually Mumbai or Singapore. If a residency rule applies, settle it first, because an AWS instance does not place the disk where your rule may need it.
- Step 3:
Get a USD quote.
The quote covers instance hours, storage and an egress assumption. Use the calculator first if you already know the family.
- Step 4:
We provision and hand over.
Access keys or session manager, security group, snapshots, and billing under the agreed arrangement. Self-run means you operate from there. Managed means the managed-cloud scope starts on the same estate.
- Step 5:
Pay the monthly invoice.
Payment methods are set in the contract.
Before the first launch, check the following:
• The AWS account is in your organization's name, with a named owner. Do not buy AWS logins from third parties.
• Budgets and alerts are set.
• Idle GPU instances are stopped on a written schedule.
• Snapshots exist, and a restore test is in scope if you need one.
• Data export and account handover on exit are written into the contract.
Self-run, managed or your own AWS account
Three commercial paths. The first two are orders from TiCON. The third is an account you hold yourself.
| Path | Self-run through TiCON | Managed through TiCON | Your own AWS account |
|---|---|---|---|
| Recommended | ✓ | Not specified | Not specified |
| What you get | An instance we provision, which you operate | The same instance, plus 24x7 operations and a named TAM | Your own AWS account and console, billed by AWS |
| Invoice | Monthly usage invoice from TiCON | Usage plus the managed fee, on one invoice | AWS bills you directly |
| When to pick it | You have an engineer and want one partner invoice | You do not want to run an on-call rota | You are testing, or your procurement prefers a direct AWS relationship |
- Step 1:
Who patches the OS and runtime?
If your team does, choose self-run.
- Step 2:
Who responds outside office hours?
If nobody does, choose managed with 24x7 coverage in the scope.
- Step 3:
Is demand steady or spiky?
Spiky demand favors autoscaling designs. GPU instances should stop when idle.
- Step 4:
Is there a residency rule?
If there is, settle the region and the data path first. See [data residency](/data-residency).
- Step 5:
Do you need a named owner and written response targets?
Then ask for a managed contract. Response targets go in the contract, not on a marketing page.
Five questions decide it:
Price: a quote, not a catalogue
Pricing is quote-based in USD. Other currencies are available only by contract, so ask. A quote covers instance hours, storage and an egress assumption. GPU hours dominate training jobs, and bandwidth dominates video and large downloads.
In our AWS reseller engagements, compute and storage are typically 15 to 35 percent below AWS list, and CloudFront bandwidth can be up to 40 percent lower. These are estimates, not quotes. Ask whether a quote is a reseller arrangement or a consulting engagement, because they are different contracts with different terms. Taxes apply as applicable in your jurisdiction and the contract.
This page does not claim:
- A public price list or a price guarantee
- An uptime percentage. Response minutes live in the contract.
- A savings promise. The percentages above are estimates.
- Legal, tax or procurement advice.
Verify our status yourself. TiCON is an AWS Advanced Tier Partner (Services), listing ID 0010h00001cA3jpAAC, on partners.amazonaws.com. If the directory and this page disagree, the directory wins.
What happens to the machine after launch
Self-run customers get a handover note: how to snapshot, how to stop a GPU instance, and who to contact when an invoice looks wrong. Managed customers get that, plus the scope table on the managed-cloud page.
- You can resize or change instance family later. You are not locked to the first shape.
- Stop GPU instances when the job ends. Leaving them running is how bills spike.
- Backups are snapshots, plus a tested restore if you want one. Say so in the order.
- On exit, the commercial relationship can end and the data is yours to export. Account-ownership terms belong in the contract.
Frequently asked questions
Can we buy a cloud server and pay in USD?
Yes. Quotes are in USD. Other currencies are available only by contract.
Is this the same as buying an AWS account?
No. This page is capacity: an instance provisioned and operated under a contract. We do not sell AWS logins.
Do you offer VPS, dedicated and GPU servers?
Yes. VPS-class (general-purpose EC2 or Lightsail), compute- and memory-optimized families, dedicated tenancy when the contract requires it, and GPU families.
How fast can we get a server?
After you accept a quote, standard instances are often ready the same working day. GPU and dedicated tenancy depend on regional capacity, and we tell you the wait before you accept.
How much does a cloud server cost?
It scales with family, hours, storage and egress. Ask for a quote. Reseller compute and storage are typically 15 to 35 percent below list, and CloudFront up to 40 percent lower.
Which payment methods do you accept?
They are set in the contract. Payment methods are not published on this page; ask during the quote.
Sources
- Amazon EC2 (opens in a new tab). Amazon Web Services. Accessed
- Amazon EC2 instance types (opens in a new tab). Amazon Web Services. Accessed
- Amazon Lightsail (opens in a new tab). Amazon Web Services. Accessed
- TiCON Cloud, AWS Partner Solutions Finder listing 0010h00001cA3jpAAC (opens in a new tab). Amazon Web Services. Accessed
- AWS Customer Agreement (opens in a new tab). Amazon Web Services. Accessed
Talk to an architect to order, or get a free cost review of an existing bill.
This page is not legal advice. Confirm applicable requirements with your counsel and compliance team.