How much does a Xandeum pNode earn? Real numbers, no hype

The honest answer: today, very little, and anyone telling you otherwise is selling something. But the mechanics matter, because the same math that pays near zero at under 1% storage usage produces very different numbers if usage grows. Here is exactly how pNode earnings work, what a typical node makes right now, and what would have to change.

By Abdennour T Bada, Pulsar Network, operators of 12 pNodes across devnet and mainnet. We monitor the network live and publish what we see.

How pNode rewards actually work

Xandeum is a decentralized storage network (blockchain-agnostic by design, live today on Solana). Apps pay fees to store and access data. Those fees collect into a pool, and once per Yuga (roughly 2.8 days, so about 130 times a year) the pool is settled and paid out in wSOL:

• 90% to pNode operators, split by contribution weight
• 5% to the Xandeum Foundation
• 5% to Xandeum Platform Corporation

This system is called STOINC (storage income). Your slice of the operator pool is your weight divided by the whole network's weight:

Four levers, all multiplied together. That multiplication is the single most important thing to understand: if any factor is zero, your weight is zero. A perfectly-run node with no stake earns nothing. A heavily-staked node that is offline earns nothing.

The four levers you control

1. Storage committed

The disk space you dedicate to the network. More real, provable storage means more weight. A budget VPS can commit a few hundred gigabytes; serious operators commit terabytes.

2. Performance

Uptime, responsiveness, reliability. The network tracks whether your node is actually there when asked. This is the lever most operators lose earnings on, and the cheapest one to fix. (Our pNode tips page covers the common failure modes we have hit running our own fleet.)

3. Stake

XAND staked to your node, your own or delegated by others, scales your weight. There is no published minimum, but zero stake means zero weight. On qualifying licensed nodes the Foundation matches community delegation 1:1 (capped at 10M), which effectively doubles that part of the lever.

4. Boost factor

A multiplier set by your purchase era (the current top era is 16×) plus up to two optional boost NFTs whose multipliers combine (a 1.2× and a 1.5× make 1.8×). Boosts are optional; running the software itself is free.

What the network looks like right now

We self-source these numbers from the network's own pRPC interface and publish them live on our dashboard. As of July 2026, mainnet has roughly 114 pNodes run by 88 operators, about 355 TB of committed storage, and around 75M XAND staked.

And the number that decides everything: stored data is well under 1% of committed capacity. The network's supply side showed up first; the demand side (apps paying storage fees) is still arriving. Small usage means small fee pools, which means small payouts, no matter how good your node is.

Worked example: a typical node, three scenarios

Take a realistic setup: 1 TB committed, 100k XAND staked, 16× era boost, perfect performance. Against today's live network weight, that node holds roughly 0.16% of the operator pool (you can verify this in our live calculator, which computes the real network denominator from per-node data).

What the pool is worth per Yuga is the unknown, so here are three scenarios. To be clear: today's actual pools sit near the bottom of, or below, the low scenario, precisely because usage is under 1%.

To put dollars on it, multiply by the SOL price yourself, or better, open the calculator: it pulls the live SOL price and the live network weight, and lets you change every input to match your actual setup, including delegation, commission and ROI payback on any upfront cost.

What it costs to run one

• Server: the only required cost. A VPS with 4 cores, 4 GB RAM and an SSD runs about $5 to $10 a month ($60 to $120 a year). 80 GB disk is enough to run; 256 GB or more if you want storage weight that matters.
• Software: free and open to anyone. No license is required to run a node.
• Boost NFTs: optional, from the pNode Store. They multiply rewards but earn nothing by themselves.
• Stake: not spent, but capital you commit (or attract as delegation).
• Time: setup is about 30 minutes (our guide), then near-zero if you monitor properly.

So the honest 2026 P&L for a typical small node: roughly $60 to $120 a year of hard cost against near-zero current STOINC income. That is the whole trade. You are not buying income, you are paying a small carry cost to hold a position (with earned uptime history and delegation) in case usage arrives.

When does it become interesting?

The formula never changes; only the pool does. The pool follows paid storage usage, so the question "when do pNodes become profitable" reduces to "when do apps store real data on Xandeum". Things worth watching, all visible on the live dashboard:

• The used % of committed storage. This is the demand signal. Under 1% means pools stay tiny; each order of magnitude changes the picture.
• Node count vs usage. More nodes without more usage dilutes everyone (bigger denominator, same pool). Usage growing faster than nodes concentrates rewards.
• Ecosystem apps. Xandeum's thesis is smart-contract-native storage for any blockchain, starting with Solana. Real sedApps storing real data are what fill the pool.

If you decide to run one, do these three things

1. Protect uptime first. Performance multiplies everything. A cheap VPS that never goes down beats a strong machine that reboots weekly. Back up your node key before anything else (why).
2. Get stake on the node. Own stake or delegation, without it your weight is zero. If you would rather not run hardware at all, you can also just stake XAND to an existing operator and share their rewards.
3. Model it before you spend. Five minutes in the calculator with your real numbers beats any blog's promises, including ours.