People ask us how a small business in Bihar runs operations across India, USA, and Singapore. The answer is more boring than the question implies — but the boring details are exactly what nobody documents publicly. Here's the real story of how we structured Amaze Internet across three countries, what worked, what we'd do differently, and what it actually costs.
This is for founders considering similar moves. If you're already running a multi-country operation, you'll recognize the shape. If you're thinking about it, this might save you a year of trial and error.
Why we did it
We started in 2016 as a single Indian Pvt. Ltd. — Amaze Internet Services Private Limited, registered in Muzaffarpur, Bihar. For the first six years that was enough. We served Indian clients, we billed in INR, we paid Indian taxes, life was simple.
Two pressures changed that:
- US clients couldn't easily pay us. Our PayPal account took 5–7 business days to clear, the FX hit was 4-6%, and many US enterprise clients won't sign contracts with foreign entities at all.
- Asian datacenter operations needed local presence. Singapore datacenter colocation contracts, in particular, required a Singapore-registered entity for some carriers.
The choice wasn't "should we expand internationally" — it was "do we want to keep losing these specific deals." We didn't.
The structure we ended up with
Three legal entities, each independently registered in its country, with the Indian Pvt. Ltd. as the parent for all three.
| Entity | Type | Purpose |
|---|---|---|
| Amaze Internet Services Pvt. Ltd. (India) | Private Limited Company | Headquarters, parent entity, all engineering, India clients |
| Amaze Internet Services Limited (USA) | Delaware C-Corporation | US client contracts, US billing, NA partnerships |
| QNET Datacenter Pte. Ltd. (Singapore) | Private Limited Company (Pte. Ltd.) | Asia-Pacific datacenter operations, Singapore-based services |
The Indian entity owns the brand, the IP, and the codebase. The US and Singapore entities pay licensing/service fees back to India for using the brand and software. This keeps the IP in one jurisdiction (cleaner for valuation) while letting each region's entity transact locally.
What it actually cost to set up
USA — Delaware C-Corp
- Initial filing & registered agent: ~$400
- EIN application: free (but takes 4-6 weeks for non-US founders)
- Annual franchise tax & registered agent: ~$450/year
- Bank account opening (Mercury): free, online, no US visit needed
- US accountant for first year setup: ~$1,800
Total year 1: roughly $2,650 USD. Recurring after that: ~$1,800/year including basic accounting.
Singapore — Pte. Ltd.
- Incorporation through a corporate services firm: ~SGD 1,500 (~$1,100 USD)
- Mandatory local director (we used a nominee director service): ~SGD 2,000/year (~$1,500 USD)
- Registered office address: ~SGD 600/year (~$450 USD)
- Annual filing & accounting: ~SGD 2,400/year (~$1,800 USD)
Total year 1: roughly $4,850 USD. Singapore is the more expensive jurisdiction, primarily because of the nominee director requirement.
Combined first-year cost to set up both US and Singapore entities, including legal, accounting, banking, and ongoing requirements. Recurring cost: ~$3.6K/year.
What worked
Mercury for US banking
If you're a non-US founder opening a Delaware C-Corp, Mercury is the answer. They specifically support international founders. The account opens in 1-2 weeks, fully online, no US trip required. Their UI is also genuinely good. We tried Brex first — they were a nightmare to onboard.
Stripe Atlas for the actual incorporation
Stripe Atlas does the Delaware filing, EIN application, registered agent, bank account intro, and tax templates for $500. For a non-US founder, this saved us probably 40 hours of paperwork. Worth every cent.
Singapore corporate services firms over DIY
Singapore incorporation has more rules than Delaware. The mandatory local director, ACRA filings, the GST registration thresholds — there's a lot to get wrong. Pay a Singaporean corporate services firm. Don't try to DIY this from another country.
Independent accounting in each jurisdiction
We tried using one global accountant (a London firm with offices everywhere). They were expensive and slow. Now we use:
- A Bihar-based CA for the Indian entity
- A US CPA for the Delaware C-Corp (~$200/month)
- A Singapore corporate services firm for QNET
Total cost: lower. Speed: dramatically faster. Each accountant knows their jurisdiction's quirks.
What we would do differently
Open the US entity first, Singapore second
We did them in the wrong order. The US entity unlocked North American clients, which directly affected revenue. Singapore unlocked Asian datacenter ops, which affected costs. Always prioritize revenue-generating moves first.
Negotiate transfer pricing earlier
The amount the US/Singapore entities pay to the Indian parent for using brand and software is the "transfer price." Both jurisdictions' tax authorities scrutinize this. We used informal numbers for the first year and had to revise them with a transfer pricing study later. Get the study done in year 1, not year 2.
Use a single brand identity, not three
QNET Datacenter Pte. Ltd. is the legal name in Singapore, but we initially marketed it as a separate brand. That was a mistake. Customers got confused about whether we were one company or three. Now we always say "Amaze Internet" externally regardless of which entity is actually billing them.
What it does NOT solve
Some misconceptions, in case you're considering this:
- It does not reduce your tax burden meaningfully. Modern transfer pricing rules and tax treaties make tax optimization through entity structure complicated and aggressive — not worth it for a small business.
- It does not give you visa benefits. Owning a US company doesn't grant you US work authorization. Same for Singapore.
- It does not simplify accounting. You now have three sets of books, three audits, three filing seasons. It's more complexity, not less.
The honest summary
The three-country structure exists for one reason: so we can sign contracts and accept payments in the way each region's clients expect. US enterprise clients want to pay a US entity. Singapore datacenters want a Singapore counterparty. Indian clients want to pay GST-compliant invoices from an Indian entity.
That's it. There's no tax magic, no visa magic, no growth-hacking secret. It's just plumbing — boring, expensive plumbing — that lets the business work the way our customers expect it to work.
If you're a small business considering this, ask yourself: are you actually losing deals because you don't have local entity presence? If yes, the cost is justified. If no, you're paying for complexity you don't need yet.
And if you do go ahead — get a good accountant in each jurisdiction. The single biggest determinant of whether this structure is a tool or a burden is the quality of your local accountants. Splurge there.