The Bitcoin network throws off a constant stream of data: block height, mempool size, fee rates, hashrate, difficulty. Each number tells part of the story. Read them together and you can judge the health of the network in real time, avoid overpaying for a transaction, and see the pressure on miners long before it shows up in a headline.

This guide covers the core metrics, what each one means, and where to watch them. It was revised in September 2026, because a lot has moved: fees sit at levels that would have looked like a bug two years ago, hashrate is counted in zettahashes, and Bitcoin Core shipped two of its most argued-about releases in years.

What Is the Bitcoin Mempool?

The mempool (short for memory pool) is Bitcoin's waiting room. A transaction you broadcast is not confirmed right away. It lands in the mempool alongside every other unconfirmed transaction and waits for a miner to include it in a block.

There is no single global mempool. Every full node keeps its own, built from what its peers relay, and caps its size (Bitcoin Core defaults to 300 MB). When a node's mempool fills, it evicts the lowest-paying transactions first. Most nodes see nearly the same pending set, but node settings create differences, and in 2025 those settings became a public fight (more below).

Mempool size used to be a clean congestion signal: big mempool, expensive fees. In 2026 it needs a second look. On the night of September 21, 2026, mempool.space showed about 83,000 unconfirmed transactions, roughly 41 blocks' worth by virtual size, yet the next-block fee was only 2 to 3 sat/vB. Most of that backlog was bidding well under 1 sat/vB and in no hurry. Size tells you how much is queued. The fee distribution tells you what it costs to cut the line.

Think of the mempool as a line at a toll booth where every driver names their own price. When the road is clear, even the lowball offers get through. When traffic spikes, the booth waves the highest bids through first and everyone else waits.

Understanding Transaction Fees

Fees are set by supply and demand. Supply is fixed: each block is limited to 4 million weight units, which works out to 1 million virtual bytes (vB), and a new block arrives roughly every 10 minutes. A full block usually takes about 1.5 to 1.8 MB on disk, because SegWit witness data gets a weight discount.

Miners fill blocks with the transactions that pay the most per virtual byte. When block space is contested, fees spike. When it is not, they sink, and since at least late 2025 the fee market has been the quietest in years. Over the 30 days to September 21, 2026, fees made up about 0.6% of total miner revenue, according to mempool.space's mining statistics. The average block carried roughly 0.02 BTC in fees on top of its 3.125 BTC subsidy.

Blocks are still full. What changed is the floor. Bitcoin Core 29.1 (September 2025) and 30.0 (October 2025) cut the default minimum relay fee from 1 sat/vB to 0.1 sat/vB, and enough miners now accept those transactions that sub-1 sat/vB bids confirm routinely. In a typical September 2026 block, the median transaction paid about 1 sat/vB or less.

Fee estimators such as mempool.space split the current state into tiers: high priority (next block), medium (about half an hour), low (about an hour) and no priority. Fees are quoted in sat/vB (satoshis per virtual byte), and a simple payment runs about 140 to 250 vB, so the total depends on both rate and size. At 1 sat/vB, a 141 vB transaction costs 141 sats. Transactions with many inputs cost more at the same rate, which is why people consolidate small UTXOs during quiet stretches. The satoshi converter shows what a fee costs in BTC and dollars.

What Is Hashrate and Why It Matters

Hashrate is the total computing power pointed at Bitcoin mining, measured in hashes per second. The network moved into a new unit in 2025: its seven-day average first passed 1 zettahash per second (ZH/s) in September 2025. One ZH/s is 1,000 exahashes, or a sextillion hashes, every second. As of September 2026, the network runs at roughly 0.93 to 1.0 ZH/s.

More hashrate means more competition for every block, which makes an attack more expensive, so sharp drops deserve a closer look. The 2021 China mining ban cut hashrate dramatically before miners relocated. In early 2026, a severe winter storm pushed Texas grid operators to ask miners to curtail, just as the price slid from its October 2025 high near $126,000, and some operators were already moving power contracts toward AI data centers. Hashrate fell back below 1 ZH/s and has hovered around that line since.

One warning: nobody measures hashrate directly. It is estimated from how quickly blocks arrive, so daily figures swing on luck alone. Between September 7 and 13, 2026, mempool.space's daily estimate ranged from about 835 to 1,040 EH/s with no real change in the machines online. Use seven-day or longer averages before reading anything into a move.

Difficulty Adjustments

Every 2,016 blocks (roughly two weeks), the network recalculates mining difficulty so that blocks keep arriving about every 10 minutes, no matter how much hashrate is online. If the previous 2,016 blocks came in faster than the target, difficulty rises. If they came in slower, it falls. A single adjustment is capped at a factor of four either way, though real moves are far smaller.

The last twelve months show the mechanism working hard. Difficulty hit a record of about 156 trillion in late October 2025. In early February 2026 it fell 11.2%, the biggest cut since the 2021 China ban, and in mid-June it dropped another 10.1%. The September 19, 2026 adjustment raised it 4.2% to about 132.8 trillion, still roughly 15% below the peak.

A positive adjustment means the same hardware earns less BTC until the price rises or someone unplugs. A negative one means the miners who stay earn more. Analysts track hashprice (revenue per unit of hashrate) for the same reason: when price outruns difficulty, margins widen even if the network barely changes.

The difficulty adjustment is one of Bitcoin's most elegant design features. It keeps issuance on schedule regardless of external conditions. No human intervention is needed; the protocol handles it automatically.

Block Height and Recent Blocks

Block height is the number of blocks added since the genesis block, mined by Satoshi Nakamoto on January 3, 2009. On September 21, 2026, the chain passed block 968,000, and back in March the 20 millionth bitcoin was mined at block 939,999. About 20.09 million BTC now circulate, and the last million or so will trickle out over more than a century.

Since the April 2024 halving, the block subsidy has been 3.125 BTC. The next halving, to 1.5625 BTC, comes at block 1,050,000, which at the current pace lands in 2028, most likely in the spring. With fees this thin, the subsidy is doing nearly all the work of paying for security, which is why the fee market matters so much heading into that halving.

Recent blocks reveal a lot: which pool found each one, how many transactions it held, total fees and the fee range. On TerminalFeed, the BTC Network panel lists recent and projected blocks in the same terminal format.

Policy Changes in 2025 and 2026

Consensus rules decide what makes a block valid. Relay policy decides what a node will forward, and each operator can configure it. That did not stop it from starting a war in 2025.Bitcoin Core 30.0, released October 10, 2025, raised the default -datacarriersize limit on OP_RETURN data from 83 bytes to 100,000, which in effect removes the cap, and allowed more than one OP_RETURN output per transaction. Supporters argued that data was already reaching the chain through worse routes. Critics saw an open door for spam. Operators can restore the old limit with -datacarriersize=83, and a visible share switched to Bitcoin Knots, an alternative client with stricter filters, though estimates of its node share varied widely.

Bitcoin Core 31.0, released April 19, 2026, rebuilt the mempool itself. Cluster mempool groups related transactions and orders them by the feerate a miner would actually earn from mining them together, replacing the old ancestor and descendant limits. A replacement (RBF) is now accepted only if it leaves the mempool's feerate profile strictly better, and the CPFP carve-out is gone. The payoff is more predictable fee bumping and better estimates.

Tools for Watching the Network

mempool.space remains the default explorer. Beyond projected blocks and fee tiers, it offers Goggles filters that highlight transaction types inside each block, block audits that score how closely a pool's block matched the expected template, RBF timelines, a mining dashboard with pool shares, and the Accelerator (launched July 2024), which lets you pay partner pools off-chain to prioritize a stuck transaction. If you run Bitcoin Core, getmempoolinfo and estimatesmartfee answer most questions without any website at all.

How to Use This Data

Regular users watch the mempool to time transactions. Weekends and early mornings in US time zones tend to be quieter, and in a fee market like this one, a no-rush transaction often confirms within a few blocks at well under 1 sat/vB. If you can wait, pick the low tier and signal RBF so you can bump the fee later.

Miners and analysts track hashrate, difficulty and hashprice. A rising difficulty with a flat price squeezes margins; steady hashrate and consistent block times point to a healthy network. Anomalies are prompts to dig, not conclusions.

Developers should pull this data from their own node or a self-hosted mempool instance where they can. Fee estimation needs extra care now that the useful range sits below 1 sat/vB: a wallet that still enforces a hard 1 sat/vB floor pays about three times the going rate on a quiet day.

All of these data points live in the BTC Network panel on the TerminalFeed dashboard, which refreshes every 30 seconds. No account required. For background, see the Bitcoin reference page and the glossary entries for mempool, hashrate and halving.