TL;DR: BSV Association‘s Teranode team has mined a 500GB test block containing 2.8 billion transactions, processing an average of over 1.1 million transactions per second. CTO Siggi Óskarsson told CoinGeek the test cluster costs around EUR 10,000 per node, down from roughly EUR 250,000 per node in the 2024 proof of concept, and that the block’s 42-minute mining time came from small CPU miners on the test network, not from any limit in Teranode itself.
Key Takeaways:
- Teranode mined a 500GB block with 2.8 billion transactions in internal scale testing, averaging 1,112,901 transactions per second.
- The 42-minute block time was hash variance from small CPU miners on the testnet, not Teranode processing time.
- Scale testing has surfaced real issues: O(n²) blowups in recovery code paths triggered by exceptions like double spends and forks. All recovered correctly and are now being optimized.
- The test cluster runs on OVH at around EUR 10,000 per node, versus roughly EUR 250,000 per node in the 2024 proof of concept.
- A public Teranode testnet is coming and “will be announced in good time”; indexers will get the block data to test their own setups.
- Óskarsson says equivalent capabilities already exist on the BSV mainnet, which has processed multi-gigabyte blocks in production.
The team behind BSV blockchain‘s new horizontally scalable node software, Teranode, has mined a single 500GB+ Bitcoin block, containing more transactions than the BTC network has processed in its entire history. The massive block was processed this week on its test network.
Siggi Óskarsson, CTO and Teranode Director at BSV Association, posted the results on November 4, 2026: a 504.73GB block containing 2.8 billion transactions, mined “without issues” as part of ongoing large-scale testing.
The block took 42 minutes and 23 seconds to mine, with an average throughput of 1,112,901 transactions per second (TPS), according to the team’s stats. Block explorer data the team shared with CoinGeek puts the exact figures at 504.73GB and 2,830,106,624 transactions, and the math checks out: that transaction count over 42 minutes and 23 seconds averages to almost exactly the TPS figure Oskarsson quoted. For comparison, BTC processes somewhere between three and seven transactions per second, and its entire blockchain history, accumulated over more than 17 years, is not much larger than this single test block. As one reply to Oskarsson’s post noted, BTC’s roughly 1.5 billion historical transactions would fit inside it with room to spare.
Speaking to CoinGeek after the announcement, Oskarsson said the lengthy 42-minute block time was not Teranode working hard, but the test network’s miners working lucky.
“This was just hashing distribution. It took that long for the hasher to find a solution. Nothing to do with Teranode really, we are just using small CPU miners on this test network and therefore it is expected to have large variance in block times. (Which is good BTW to test edge cases like this).”
In other words, the software kept pace with over a million TPS flowing in, and then waited for a CPU miner to stumble on a valid hash. On mainnet, with real hashpower, that wait would average the more-standard ten minutes.
Finding problems is the point of testing
So did the test actually surface any of those issues Óskarsson mentioned that only appear at scale? Yes, he said, though mostly in earlier runs over the past couple of months.
“Our tests over the last couple of months have surfaced lots of issues, mostly related to O(n²) problems when things go wrong. What that means is that on exceptions, like double spends, forks etc., some of Teranode’s recovery algorithms were not tested properly at scale and caused issues. Everything worked OK, and it recovered, but we are optimizing those to be much faster to allow Teranode to function properly.”
That’s an important detail. The code paths that handle things going wrong, rather than the easy path, are where big systems often hide their scaling shortcomings, and they’re exactly what a test program like this exists to flush out before mainnet traffic starts to depend on them.
The hardware bill may actually fall
Perhaps the most striking number Óskarsson shared wasn’t about the block itself at all, but the machines that processed it.
“Currently we are running on OVH on a cluster that costs around EUR10,000 per node,” he said. “It is difficult to estimate exactly, but compared to the POC in 2024 where each node cost around EUR250,000, this is a HUGE improvement.”
A 25-fold drop in per-node cost in two years might matter more than it appears at first. The standard criticism of huge-block scaling is that only data-center-sized operators can afford to participate, which concentrates (and thus “centralizes”) the network. If a node capable of keeping up with million-TPS traffic costs roughly what a small business server costs, that argument gets much weaker.
Public testnet ‘in good time’
Óskarsson confirmed the test ran on an internal network rather than the public Teranode testnet, and that the team intends to open one up.
“Yes, this is coming, but we don’t want to talk about it soon. Will be announced in good time,” he said. He added that the block data would be shared so that indexer services can test their own setups against it: “We will give the indexers a chance to test their setups on this.”
Asked whether the test network would be listed on exchanges, he was direct: “This network is just a testnet, so won’t be listed, but these capabilities are already on mainnet. Build apps and blast transactions!”
This was a controlled test, but the team has shared block explorer data with CoinGeek backing it up: the block sits at height 1155 on the test chain, timestamped 11:50:31 UTC on 4th November, about 18 minutes before Oskarsson’s post went up. The explorer record shows a coinbase tag of “/dev-ovh-1.scale-1/,” matching the OVH cluster he described, and a difficulty of 1 with a 50 BSV block reward.
Note: Both are what you’d expect from a young test network—the subsidy schedule is based on block height, starting at 50 coins and halving every 210,000 blocks, so a chain only 1,155 blocks old is still in its first era, and difficulty 1 is plenty when the miners are small CPUs. And BSV’s mainnet has already mined blocks of 4GB and above in production, with a design target of one million TPS public since Teranode was announced. The test shows the software is indeed hitting that target in practice.
Why scale is the whole ballgame
Teranode replaces the original monolithic SV Node software with a microservices architecture that spreads validation, block assembly, and propagation across multiple machines, scaling horizontally rather than demanding ever-bigger single servers. Transactions are streamed into batches called “subtrees” almost as soon as they arrive, so by the time a block is mined, most of the validation work is already done.
The economic argument matches the technical one. Block subsidies halve roughly every four years, which means miners everywhere must eventually live off transaction fees. A network that caps its capacity forces fees up and prices out ordinary transactions for all users. A network that scales can keep fees at fractions of a cent and make up the difference in volume. One reply to Óskarsson’s post put it in BSV terms: a recent 1.4GB mainnet block carrying 5 million transactions paid around 1 BSV in fees, the Amazon (NASDAQ: AMZN) high-volume model rather than the convenience-store model.
At 2.8 billion transactions per block, even minuscule fees would become serious revenue, and entirely new categories of application become practical: micropayments, on-chain data storage, machine-to-machine payments between AI agents, and high-frequency logging or ticketing systems that would be economically impossible on a capacity-capped chain.
Kurt Wuckert Jr., whose GorillaPool operation mines BSV blocks (including Teranode), summed up the community mood in a reply: “It’s an honor to be on this journey with you and the team.”
The real test comes when the public Teranode testnet opens, and outside parties can hammer the software themselves, and when indexers, exchanges, and application builders confirm their own systems can keep up. Blockchains are only as useful as the ecosystem built around them. But as demonstrations of raw capacity go, 2.8 billion transactions in a single block on hardware a small business could afford is a number the rest of the digital asset industry cannot currently match or come close to.
FAQs
What is Teranode?
Teranode is the BSV blockchain’s next-generation node software, developed under the BSV Association. It replaces the original monolithic node design with a horizontally scalable microservices architecture, targeting sustained throughput of over one million transactions per second with no fixed block size limit.
How big was the Teranode test block?
The block mined in the November 2026 internal test was 504.73GB and contained 2,830,106,624 transactions, at block height 1155 on the test chain. It took 42 minutes and 23 seconds to mine, averaging 1,112,901 transactions per second.
Why did the 500GB block take 42 minutes to mine?
The test network uses small CPU miners, so block times vary widely while the miner searches for a valid hash. The BSV Association says the 42 minutes was hash variance, not processing time: Teranode itself sustained over 1.1 million transactions per second during the test.
Was the 500GB block mined on the BSV mainnet?
No. It was mined on an internal test network. The BSV Association plans to open a public Teranode testnet and says the same capabilities are already available on mainnet, where 4GB blocks have been mined in production.
How much does it cost to run Teranode at scale?
The BSV Association’s current test cluster runs on OVH at around EUR 10,000 per node, down from roughly EUR 250,000 per node in the 2024 proof of concept.
How does Teranode compare to BTC’s capacity?
BTC processes roughly three to seven transactions per second and caps blocks at around 1MB (4MB weight units). The Teranode test block alone contained more transactions than BTC has processed in its entire history.
Why does blockchain scalability matter?
As block subsidies halve over time, miners depend increasingly on transaction fee revenue. A scalable network can process billions of low-fee transactions profitably, enabling micropayments, on-chain data applications, and machine-to-machine commerce that are uneconomical on capacity-limited networks.
Related articles
Watch | Teranode Explained: How Teranode changes blockchain scaling
Jon Southurst
Jon Southurst is CoinGeek’s Associate Editor for Asia-Pacific, based in Tokyo. He has reported on Bitcoin and blockchain since 2013, including the Mt. Gox collapse, and previously wrote for CoinDesk, Bitcoin.com, and Bitsonline, which he co-founded. His coverage focuses on BSV development, artificial intelligence and agentic AI, stablecoins, CBDCs, and asset tokenization. He has interviewed figures including Tim Draper, Roger Ver, Craig Wright, and Steve Keen, and moderated panels at the London Blockchain Conference.
