Capy Research

► How To Read The Hashrate? Do you see hashrate going up on mining coins and immediately call it bullish? But is it tho? Looking at BTC, PRL, QTC and ZEC lately… there's a lot more to these charts than I thought. 1/ $BTC BTC is at 1,058 EH/s on the 7D average, back in 1 ZH/s territory. Difficulty at 140.58T after a +5.92% adjustment, with blocks coming in around 9m25s instead of 10m. Looks strong but hashprice is only ~$37 per PH/day, which means competition is increasing while revenue per unit of compute stays pretty thin. A 12 J/TH ASIC can theoretically cover electricity costs up to ~$0.126/kWh, while an older 30 J/TH machine only reaches ~$0.051/kWh, before other costs. Same BTC price, completely different miner economics. What people overlook is replacing 30 J/TH machines with 12 J/TH hardware gives ~2.5x more hashrate using the same electricity. So rising hashrate doesn't always mean fresh capital flooding in. Sometimes it's just better machines replacing old ones. Top 3 pools also account for ~63.3% of recent blocks, top 5 around 79.3%. So BTC hashrate can rip while miners get squeezed. Both can be true. 2/ $ZEC Network hashrate is ~30.1 GSol/s, up ~28.6% over 90D, with estimates pointing to ~4x growth YoY. That's equivalent to roughly 33,000 Z15 Pro machines based on Hashrate Index estimates. But mid-October estimates also suggested ZEC ASICs were generating substantially more revenue per MW than contemporary BTC ASICs. When the economics suddenly look attractive, miners pay attention. Machines come online, old hardware gets reactivated, and more capital starts chasing rewards. Doesn't necessarily mean everyone suddenly became a ZEC maxi. Sometimes miners just follow the money lol. Another interesting bit is ~57.4% of recent blocks being classified as unknown. Some might call that decentralized, but unknown doesn't mean distributed. We simply don't know who's behind that capacity. And for ASIC-heavy networks, knowing who controls compatible hardware matters a lot more than people think. 3/ $PRL / $Pearl @prlnet uses INT8 matrix multiplication, a different kind of computational work. Network hashrate ~75.34 EH/s, up ~63% in a month, difficulty 42.2M. PearlTrack also reported ~7.46K TMAC of verified AI compute in 24h. But verified compute doesn't automatically mean useful AI demand. You can prove a miner performed matrix calculations without proving those calculations served an actual paying AI customer. That's the gap between Proof-of-Useful-Work as a narrative and actual economic utility. And with Kryptex accounting for ~49.6% of recent blocks, PearlHash another ~19.1%, nearly 69% comes through just two pools. Compute is growing fast, but I wanna see whether that growth translates into real AI workloads, external revenue, and miners staying when profitability cools off. 4/ $QTC @QuantusNetwork uses Poseidon2-based QPoW, targets 12-second blocks, and adjusts difficulty every single block instead of BTC's 2,016-block cycle. – Sep 30: 24 TH/s with difficulty above 300T – Oct 22: 30.6 TH/s and difficulty around 368T Some growth, but remember mainnet only launched Sep 9. We're barely six weeks in. What caught my eye is how quickly difficulty reacts to incoming miners. On BTC, fresh hashrate can speed up blocks until the next adjustment. QTC adjusts to that competition much faster. But since it's GPU-oriented, miners can potentially rotate elsewhere when another opportunity pays better. The question is whether those GPUs are still there after the early mining opportunity gets crowded. Anyone can show up for juicy rewards. Staying when margins compress is where conviction gets tested. 5/ Smaller PoW coins Imagine a coin suddenly becomes more profitable to mine. Existing miners rotate their machines in, hashrate pumps, difficulty catches up, profitability drops, then some miners leave. On the chart, it looks like massive adoption. In reality, it might just be the same hardware chasing better rewards elsewhere. That's why percentage hashrate growth alone can be such a trap. I think we're measuring the wrong thing when we only compare raw hashrate. I'd rather look at 7D/30D/90D trends, difficulty, block times, hashprice, hardware efficiency, pool concentration, and how easily miners can switch networks. Also worth remembering that BTC price and mining profitability often lead hashrate, not the other way around. So when hashrate finally starts ripping, it might just be miners deploying hardware based on decisions made weeks or months ago. Hashrate staying up when profitability gets ugly tells you something much more interesting.