Sunday morning, 09:28 UTC, a single Raspberry Pi running Public Pool on Umbrel found Bitcoin block #948,146. Thursday morning, 09:22 UTC, a self-hosted 272 PH/s "small farm" found block #948,440. Reward each: roughly €213,000. Pool fee: €0. Custody: none. In 48 hours, the Bitcoin protocol paid €425,000+ directly to two non-custodial solo miners — without an intermediary holding the funds even briefly.
The two blocks — verifiable on-chain
Block #948,146 · Sunday 6 May 2026 · 09:28 UTC
- Coinbase scriptSig:
"Public Pool on Umbrel" - Setup: Raspberry Pi + Umbrel + Public Pool — Bitaxe-class hashrate
- Reward: 3.154586 BTC (~€213,000)
- 1 single coinbase output — paid directly to the miner's bitcoin-address
Block #948,440 · today, 8 May 2026 · 09:22 UTC
- Coinbase scriptSig:
i://6b831b2b-70d9-4c3d-9973-984a92cadf88(a UUID — telltale signature of self-hosted ckpool without branding) - Setup: ~272.8 PH/s self-hosted "small farm" (per soloblocks.io)
- Reward: 3.1384 BTC (~€213,000)
- 1 single coinbase output — paid directly to the miner's bitcoin-address
Two completely different hashrate classes — a Pi running ~1 TH/s and a small farm running 272,000 TH/s — used the same fundamental architecture: their own pool software on their own hardware, no intermediary. Both received the full block reward via the Bitcoin protocol's coinbase mechanism. Neither had to wait for a payout cycle, neither had to trust a pool wallet, neither lost a percentage to a fee.
This isn't lucky — it's variance behaving normally
It's tempting to read "two solo blocks in 48 hours" as a hot streak. It isn't. Solo block-finding is an exponential distribution: each block-find is statistically independent, the mean and median diverge wildly when probabilities are tiny, and clusters of finds in short windows are expected. April 2026 was quieter than the long-term average. The first week of May is bringing us back.
For context: in May 2026 the Bitcoin network averages 144 blocks per day. With ~30 active solo / non-custodial pools globally absorbing roughly 0.1-0.3% of total hashrate, statistically you'd expect about 0.4 - 1.3 solo finds per week. April had fewer than that. This week we've already had two and it's only Thursday.
The lottery is real. The lottery pays out. The lottery isn't paying out FOR you specifically — but it is paying out somewhere, every week, and a non-zero portion of that payout flows through self-hosted, non-custodial setups exactly like Maenpolder's.
What this proves — for any miner watching
The non-custodial-solo model works at every hashrate scale. A Raspberry Pi pulled it off Sunday. A 272 PH small farm did it Thursday. The architecture is the same — single-output coinbase, no pool wallet, full reward to a single address — and it scales from a $200 home setup to multi-megawatt operations.
Maenpolder offers the same architecture as a hosted endpoint, so you don't have to also be a Bitcoin Core sysadmin or run a Pi 24/7 to participate. Same fee structure (zero), same custody model (none), same protocol-enforced payout. Lower barrier, same economics.
Setup — 30 seconds
Two settings in your Bitaxe's web UI under Settings → Stratum:
Host: maenpolder.com
Port: 3334 (TLS — recommended)
or 3333 (plaintext)
Username: your-bitcoin-address[.optional-worker-name]
Password: xSave and reboot. Within 30 seconds your hashrate appears live on maenpolder.com, with real-time time-to-block math and a next-block-reward estimator.
One last thing
Both Sunday and Thursday's winners had something in common beyond their setup: they happened to be turned on. The miners who weren't running this week didn't win. The miners who were running and pointed at custodial pools won, technically — but had to wait for payouts and lose a percentage. The miners who were running and pointed at non-custodial pools won the cleanest possible version of the lottery.
It's a small thing — turning a miner on and pointing it at the right URL. The two miners who took home €213k each did exactly that.