The clean four-year cycle — accumulate quietly for two years, send for one, capitulate for one, repeat — has been the most-cited mental model in Bitcoin for over a decade. As I write this in May 2026, twelve months past the April 2024 halving, that model has failed its first real out-of-sample test. Bitcoin sits around $80K, ETF cumulative inflows have rebounded to $58.7 billion, and almost every macro newsletter I read has declared the four-year cycle officially dead.
I want to walk through what the four-year cycle actually claimed, what changed in 2024–2026 that makes the test interesting, and what the on-chain and flow data we can verify ourselves are saying. I have no thesis to defend here — I run a node, I DCA, and I sleep fine in either world. But I do want a clean answer for my own understanding, and I think the honest answer is more nuanced than “dead” or “alive”.
TL;DR. The strongest version of the four-year cycle thesis — that the halving’s reduction in new supply mechanically drives a bull market 12–18 months later — never had a particularly clean causal story, and it survives the 2024 halving with weak signal at best. The weaker version — that Bitcoin moves in extended boom-bust cycles roughly four years long, driven by reflexive flows and macro liquidity — is harder to falsify and may still be intact, but the dominant variable has shifted from miner sell-pressure to ETF flows + global liquidity. The honest answer is “we have one post-ETF data point — wait for two more halvings before declaring anything.”
What the four-year cycle actually claimed
Two distinct claims tend to get blurred when people say “the four-year cycle”:
The first is the supply-shock claim: every four years, the protocol halves the block subsidy. Miners now produce half as much new BTC. Assuming demand is roughly constant, halved new supply means rising price. The 2012, 2016, and 2020 halvings each preceded a major bull run that peaked 12–18 months later, which is the textbook evidence for this thesis.
The second is the reflexive-flow claim: rising prices attract media attention, attract new buyers, attract more attention, attract more buyers, in a positive feedback loop until the marginal buyer is leveraged and the unwind triggers a 70–85% drawdown — a winter that lasts 12–18 months until the cycle restarts. The halving might be the trigger, but the driver is reflexivity in flows, not the protocol mechanic.
These are very different claims with very different testable predictions. The supply-shock claim is mechanical and quantitative — you can compute miner sell pressure as a fraction of price, and watch it shrink. The reflexive-flow claim is behavioral and survives almost any specific failure as long as you allow that “this cycle the trigger is X instead of halving.”
Why 2024 was the first real out-of-sample test
By the 2024 halving, miner-issued new supply had fallen to about 0.83% of total supply per year (450 BTC/day from a 19.7M BTC float). After the April 2024 halving cut subsidy from 6.25 to 3.125 BTC, that fell to roughly 0.42% — already a small fraction of the float that gets traded daily.
For comparison, the 2012 halving cut new issuance from a base where roughly 4% of supply was being mined annually. The marginal mechanical impact of a halving on supply has been falling exponentially every cycle. By the 2028 halving, daily new-supply issuance will be roughly 225 BTC against a daily traded volume that is regularly above $30 billion — a rounding error.
Then in January 2024, three months before the halving, US spot Bitcoin ETFs launched and pulled in $25 billion in their first six months. That single fact swamps any plausible halving-supply-shock by an order of magnitude. The 2024 halving removed about 450 BTC/day of new supply (~$36M/day at $80K). The first six months of ETF flows added about $138M/day of new buy pressure on average. The mechanical supply-shock claim was always going to face this for the first time in 2024, and it has.
What the post-2024 data actually shows
Let me list the things I can verify from public sources, treating each as a discrete data point rather than fitting a narrative around them.
Halving date (April 19, 2024) to peak. Bitcoin reached an interim local high around $73K in March 2024 — before the halving, driven entirely by the ETF launch. It then chopped sideways and pushed to roughly $109K in January 2025, about nine months after the halving. The 12–18 month post-halving peak window from prior cycles would have placed a peak somewhere between April and October 2025. The actual peak (so far) was earlier than that window, in January, suggesting the catalyst was ETF flows + rate-cut anticipation, not the halving itself.
Drawdown depth. From the January 2025 peak around $109K, the drawdown to the November 2025 low around $74K was about 32%. Prior post-halving cycles produced 70–85% drawdowns from peak to bottom over 12–18 months. A 32% drawdown over 11 months is a much shallower correction with a different shape — suggesting the marginal seller is structurally different (institutional rebalancing rather than retail capitulation + leverage unwinds).
ETF flow correlation. Farside’s daily flow data shows the strongest single explanatory variable for daily price action since launch is net ETF inflows. The two-month outflow stretch from November 2025 to February 2026 corresponded almost exactly with the price weakness in that window; the resumption of inflows from February tracks the rebound to the current $80K range. The traditional cycle thesis would have you watch hashrate, miner reserves, and on-chain MVRV. Those still work directionally, but ETF flows are now the dominant proxy for marginal buyer behavior.
Hash rate. Network hash rate has continued its monotonic climb through the entire post-halving period and currently sits near 950 EH/s. There is no “miner capitulation” signal in the hash data — miners aren’t selling reserves to cover costs at any abnormal rate, because energy efficiency improvements (ASIC generations, energy contracts) have absorbed the subsidy halving more cleanly than any prior cycle. That is itself evidence against the supply-shock thesis: the cycle’s primary mechanical lever (forcing inefficient miners offline, reducing sell pressure) has been quietly removed by industry maturation.
Realized volatility. Annualized 90-day realized vol has been sitting in the 35–50% band since the ETF launch, roughly half the 70–110% range characteristic of prior cycles. This is consistent with ETF rebalancing rather than retail flows: pension funds and RIAs do not panic-sell at 5pm Tokyo time the way 2017 retail did.
The honest reading
I find all of those data points individually credible and collectively pointing in one direction: the mechanical supply-shock cycle is dead, and the reflexive-flow cycle is not yet falsified but operates on different inputs now.
Bitcoin’s marginal buyer is no longer a Korean retail trader checking the price every 30 minutes; it is a pension fund’s quarterly rebalance, an Argentine saver buying $50/month through a Strike-style app, a Black Rock/Fidelity model-portfolio inflow. Those flows have macro and liquidity drivers that have nothing to do with the halving. Whether the rolling 12-month return pattern will still rhyme with prior cycles — which is what most people actually mean when they say “the four-year cycle is alive” — depends on whether reflexivity in those flows produces similar boom-bust dynamics. We will not know until the next halving cycle (2028 → 2030) is half-played-out.
What we can say cleanly:
- The 2024 halving itself had no detectable price impact in the 30-day window around it. Prior halvings also didn’t have a sharp price reaction at the halving date (price moves were always lagged 6–18 months), but the shape of the post-halving trajectory in 2024 has been substantially different from 2012, 2016, and 2020.
- ETF flows are now the dominant proxy for marginal demand, which was true of no prior cycle. If you only watched on-chain metrics and ignored Farside flows over the past 18 months, you would have been confused by the price action.
- The drawdown shape (shallower, slower, more correlated with macro) is more consistent with a maturing institutional asset class than with a retail-driven reflexive cycle.
How I verify this myself, and how you can
Three free data sources are enough to reach your own conclusion without any newsletter:
- mempool.space — current hash rate, difficulty adjustments, and miner activity. If the cycle thesis predicts miner sell-pressure spikes, hash rate should reflect it. It hasn’t.
- Farside Investors — daily ETF flows by issuer. This is the single highest-signal series for marginal buyer behavior since January 2024. Free CSV.
- CoinGecko market data API — daily price data going back to 2010. Plug into our DCA calculator to see what a constant monthly buy from any prior date would look like today, or our HODL calculator for lump-sum returns. Run those over different “cycle” definitions and see whether the post-halving 12–18 month hold beats a simple monthly DCA — for the 2024 halving, the answer so far is no.
You can also read our Power Law Projection tool, which models Bitcoin’s long-run trajectory as a power-law fit through cycle peaks and troughs rather than as a discrete four-year wave. The power-law model has held up arguably better than the cycle model post-2024 — it predicts the price band the asset will trade in over multi-year horizons, which is useful for DCA planning and irrelevant to short-term timing.
What I do given this
Practically, the cycle-vs-no-cycle debate doesn’t change much for a long-horizon Bitcoiner. If you DCA on a fixed schedule, the cycle is irrelevant — your average buy price dollar-cost-averages through whatever shape the trajectory takes. If you were trying to time the cycle exit, the post-2024 data has been a much harder market to time than 2017 or 2021, and most cycle-based exit signals have either fired prematurely or not fired at all.
The thing I find actually useful is that the risk profile of holding Bitcoin has changed. A 70-85% drawdown is no longer the base case for the next cycle, but neither is a 100x bull market — both tails have compressed. For someone allocating from a 60/40 portfolio or treating Bitcoin as a long-duration savings vehicle, that’s probably more important than whether the four-year pattern survives one more cycle.
If I’m wrong about all of this — if 2026–2027 produces a textbook reflexive blow-off followed by a 75% bear and the cycle thesis lives — I will be neither surprised nor poorly positioned. The DCA absorbs it. Don’t trust the narrative — verify it against your own data.