$1,024 billed. $355 left, in liquid USDC.

That is the reference month for a healthy decdn node at the launch-fleet target: 100 TB served at a market rate of $0.01/GB. The point is not to claim every node always earns that number. The point is to show the cash path, line by line, without mixing operator revenue, token alignment, and protocol treasury flow into one blurry headline.

The setup

The reference node serves 100 TB in a month. It runs on a low-cost 1 Gbps-class setup: a Hetzner CPX22 with bandwidth overage, or an equivalent unmetered dedicated tier once traffic makes metered bandwidth a bad trade. Around 100 TB/month, those curves converge. Below that, a cheap VPS can work. Above that, a flat bandwidth port starts to matter.

The rate is $0.01/GB. That is the public target rate, not a protocol-enforced price. Nodes advertise their own rates; clients choose among price, latency, and reputation. The model uses $0.01/GB because it is the launch reference price: cheap enough to be a real CDN alternative, high enough for efficient operators to clear costs.

The traffic mix is:

Stream	Monthly volume	What it means
Client delivery	70 TB	End users pulling content from this node
Peer delivery	30 TB	Other nodes paying this node for cache-miss pulls
Total served	100 TB	Every served byte is paid delivery

Both streams are revenue. When another node needs a blob this node already has, it opens or extends a payment channel and pays for the pull. From this node's point of view, that peer request is another paid delivery.

This is why cache quality matters. A useful cache is not a favor to the network. It is inventory.

Gross billing

100 TB at $0.01/GB is $1,024 billed.

The arithmetic is:

100,000 GB x 1,024 MB/GB x $0.00001/MB = $1,024

That $1,024 is gross settlement volume, not operator take-home. Every settlement goes through FeeRouter:

Bucket	Share	Cash path
Operator base	60%	USDC to the serving operator
Buyback-and-burn	30%	USDC buys TOKEN and burns it
Treasury	10%	USDC to protocol treasury

Node-to-node cache-miss pulls do not bypass this split. A peer settlement routes the same way a client settlement does. That keeps the accounting simple and makes fake self-volume uneconomic.

So the reference month becomes:

Stream	Gross billing	Operator cash share
Client delivery, about 70 TB	$717	$430
Peer delivery, about 30 TB	$307	$184
Total operator inflow	$1,024	$614

The other $410 is real value flow, but it is not cash the operator can use to pay the server bill. About $307 funds buyback-and-burn, and about $102 funds treasury.

Cash costs

The largest variable cost is paid cache misses.

The 70/30 traffic split describes outgoing served bytes: 70 TB to clients, 30 TB to peer nodes. Separately, the 85% cache-hit assumption says this node has to buy 15 TB from peers to cover its own misses. Peer delivery is revenue; cache-miss pulling is cost.

When this node does not have a requested blob, it can redirect the client to another node, or it can pull the blob through, cache it locally, and stream it to the client. Pull-through is the better default: it keeps the client on this node's channel, keeps the request fast, and warms the cache for the next request.

At an 85% cache hit rate, 15% of served traffic has to be pulled from peers:

15% x 100 TB = 15 TB paid wholesale
15 TB x $0.01/GB = about $154 out

Infrastructure is the next line. At this volume, the model uses $95/month as the conservative hosting anchor. That reflects the point where a metered VPS with overage and a 1 Gbps unmetered setup are effectively competing choices.

L2 gas is small, but not zero. Channel opens, withdrawals, cooperative closes, and settlement all need budget. The model uses $10/month for channel operations at normal L2 fee levels.

Total cash costs:

Cost	Amount
Cache-miss pulls	$154
Infrastructure	$95
L2 gas	$10
Total	$259

Operator P&L

Put the revenue and costs together:

Line	Amount
Client delivery, operator cash share	+$430
Peer delivery, operator cash share	+$184
Cache-miss pulls	-$154
Infrastructure	-$95
L2 gas	-$10
Net liquid USDC	+$355

That is $355/month of liquid USDC, about 35% of gross billed delivery, before any TOKEN-side upside.

This number moves with operating quality. A warmer cache improves it. Better bandwidth improves it. The reference case is deliberately ordinary: a 1 Gbps-class node, 100 TB/month, 85% cache hit rate, $0.01/GB market pricing.

What not to count as cash

The operator-aligned return is larger than the cash line, but it should not be mixed into OpEx math.

FeeRouter sends 30% of every settlement to buyback-and-burn. For bonded operators, that burn is aligned with the operator side of the network: capacity providers hold TOKEN bonds, and network usage mechanically removes TOKEN supply. Counting direct USDC plus burn, 90% of every settled byte is operator-aligned.

But invoices are paid in liquid currency. The operator's cash P&L is the operator base share minus cache-miss pulls, infrastructure, and gas. The burn belongs in the upside story, not the server bill.

Why the economics work

The economics work because three pieces line up.

First, operators earn in USDC. They are not asked to cover dollar-denominated bandwidth bills with volatile emissions.

Second, payment channels make per-MB billing cheap enough to matter. Most byte-level payments happen off-chain; the L2 only sees channel lifecycle transactions and settlement.

Third, cache misses are paid. Popular content spreads because the next node can earn from holding it, not because someone asked nicely.

That is the operator thesis in one sentence: serve useful bytes, keep the right cache warm, and settle in stable currency. The launch-ramp subsidy covers the cold-start period while traffic fills in. After that, break-even sits around 5 TB/month per node in the reference setup. At 100 TB/month, the reference node clears $355/month in liquid USDC.