A merchant doing $10 million a year pays $100,000 to BitPay at 1 percent. The same merchant running PayRam on a $20-a-month VPS pays $240 a year. The gap is $99,760. The sovereignty argument finally has a number attached to it.
This is not a Bitcoin-cypherpunk pitch. This is a SaaS-vs-self-host economics post. The conclusion is straightforward: above $500K of annual processing volume, the math on self-hosted crypto payment gateways becomes structurally unbeatable. Below that volume, SaaS is fine. The crossover happens earlier than most merchants think.
How does the cost stack differ between SaaS and self-hosted gateways?
A SaaS gateway (BitPay, NOWPayments, CoinGate, Coinbase Commerce in its remaining markets) charges a percentage. The percentage scales with your volume. Marginal processing cost on the vendor side is near-zero. The vendor's margin grows linearly with your success.
A self-hosted gateway is software. You deploy it on a VPS, point a domain at it, plug in your wallet addresses, and it runs forever. The cost is the VPS rental and the gas fees on the underlying chain.
PayRam advertises the architecture explicitly. According to PayRam's self-hosted guide: "A merchant processing $10M annually at a 1 percent fee pays $100,000 per year in fees. That same merchant using PayRam pays 0 percent in processing fees. Their cost is limited to the server hosting (about $20/month) and standard network gas fees."
Run the cost stack:
| Annual volume | BitPay at 1% | PayRam self-host | |---|---|---| | $100K | $1,000 | $240 | | $500K | $5,000 | $240 | | $1M | $10,000 | $240 | | $10M | $100,000 | $240 | | $100M | $1,000,000 | $240 |
The crossover is around $24,000 of annual volume. Above that, the self-host is cheaper. By the time you cross $500K, the SaaS gateway is 20x more expensive than the alternative.
Why most merchants stay on SaaS anyway
Three reasons the self-host argument has been slow to catch on.
Setup friction. Deploying a VPS, configuring DNS, generating SSL, connecting wallet addresses, and integrating with your checkout used to take a weekend of engineering time. For a small merchant, the weekend was worth $5,000 of avoided BitPay fees. Not a compelling tradeoff at low volume.
PayRam's recent product changed this. The setup is now reportedly under 10 minutes for someone comfortable with basic VPS deployment. AI agents are deploying it autonomously without human involvement.
Compliance fear. Merchants worry that running their own payment processor exposes them to money-transmitter regulation. The legal reality is the opposite. PayRam's guide makes the legal point: "Self-hosted gateways operate as technology infrastructure rather than financial intermediaries, exempting them from direct custody regulations and preserving merchant privacy." A self-hosted gateway is software you run, not a service you provide. The compliance obligation rests on the merchant's jurisdiction, not on a US-registered money services business.
Key management anxiety. The merchant worries that running their own infrastructure means becoming responsible for private keys. The mitigation is the architecture itself. PayRam advertises "No Keys on Server" (NKOS): "The server never holds private keys. Smart contracts handle all fund sweeping from deposit wallets to cold storage. Even if the server is compromised, funds cannot be moved." The merchant operates the software but the private keys live in cold storage. The attack surface is materially smaller than a custodial SaaS.
What you give up
Self-host is not free in every dimension. Real tradeoffs:
Support is your problem. When the gateway breaks at 3 AM on a Saturday, no one is on call but you. SaaS gateways have (limited) support teams. Self-host has community forums and your own engineering capacity.
Brand recognition at checkout. "Powered by Coinbase Commerce" used to carry trust signal weight with consumers. "Powered by PayRam" does not yet. For B2C merchants who optimize on conversion rate, the brand absence costs basis points.
Feature lag. SaaS gateways ship integrations (Shopify, WooCommerce, Magento, Wix) faster than self-hosted alternatives. PayRam, BTCPay Server, Aurpay, and the rest are catching up but lag major SaaS on the long tail of plugins.
Upgrade burden. A SaaS gateway upgrades for you. A self-hosted gateway requires you to pull updates, rotate keys, monitor uptime. For a single-engineer shop this is real overhead. For a 10-person team it is half a Wednesday afternoon every quarter.
When self-host wins decisively
Three signals that you should be running self-hosted infrastructure:
- 1.You process more than $100K annually in crypto. At that volume, the percentage savings exceed an engineer's annualized cost of a half-day-per-month maintenance window.
- 2.You operate in a high-risk vertical. SaaS gateways gate access to high-risk categories with enhanced KYC, fee premiums, or outright rejection. Self-host has no gating; the rail is the rail.
- 3.You serve a non-US/non-Singapore merchant base affected by the Coinbase Commerce shutdown. The structural risk of "your processor decided your geography is no longer worth supporting" is what self-host eliminates.
Where Plaitr fits in the spectrum
Plaitr sits between the SaaS extreme and the full self-host extreme. The merchant connects their existing wallet (non-custodial like self-host) and pays a flat monthly subscription instead of a percentage (priced like a tool, not a tax). Plaitr operates the gateway infrastructure (no VPS for the merchant) but does not custody the funds (no counterparty risk like SaaS).
The structural property is the same as self-host: funds settle directly to the merchant's wallet, no percentage extraction, no KYC gating. The operational property is friendlier than self-host: zero VPS management, automatic upgrades, included support. The cost is a flat $99-$999 per month depending on volume tier.
For most merchants this hybrid is the right answer. For very-high-volume merchants ($10M+ annual) the pure self-host with PayRam may still win by a few thousand dollars a year. For most others, the operational simplicity of a flat-fee managed non-custodial rail beats the marginal savings of running your own VPS.
What to do this week
If you process more than $50K/year in crypto and pay any percentage fee to a SaaS processor, run the math. Pull last 12 months of invoices, sum the fees, project the next 12 months. Compare against $240 (self-host) and $1,188-$11,988 (flat-fee managed). Both alternatives beat the SaaS percentage above $50K of annual volume.
The sovereignty argument is no longer ideological. It is a spreadsheet question with a one-sided answer above a low volume threshold. The merchants who run the numbers move. The merchants who do not, fund the percentage.