Notice: Editorial analysis with market data from July 30, 2026 (CoinGecko prices at European close). Power law levels are recalculated daily and become obsolete within days. This article does not constitute financial advice. CleanSky does not receive commissions or referral payments from any of the platforms, models, or data providers cited.
Bitcoin has not broken the power law: it is trading 8.1% above it. On July 30, 2026, BTC is worth $63,999 and the model's floor sits at $59,192. The power law (the model describing Bitcoin's price as a straight line when price and time are plotted on a double logarithmic scale, anchored to the genesis block of January 3, 2009) remains intact after seventeen and a half years. What is true is that this floor rises by about $54 every day due to the pure arithmetic of the model, allowing us to date the encounter: if the price stays exactly where it is, the floor will reach it on October 25, 2026. The "mid-December" figure that circulated in February was calculated with Bitcoin at $67,000; the subsequent drop has moved the appointment forward by nearly eight weeks. The rest of the article puts numbers to the calendar: the two published floors, the full band, the crossover date for each flat price, and why the same model crumbles as soon as it is applied to Ethereum.
Has Bitcoin broken the power law?
No, and the distance is not marginal. The two published floors of the model are clearly below the market price on July 30.
The first is from Fidelity. Jurrien Timmer, the firm's Director of Global Macro, has maintained a long-term valuation channel for years whose lower line has captured every cycle floor since 2015, including the 2018 and 2022 lows. On July 12, with Bitcoin at $62,700, Timmer placed that line at a daily reading of $58,237 —7% below the price at the time—. Coverage the following day cited the rounded level around $56,500.
The second is from Giovanni Santostasi, the astrophysicist who formulated the theory in a Reddit post in 2014 and maintains it on the Newhedge data platform. On February 20, 2026, that floor stood at $51,128 with Bitcoin at $67,000. Applying the model's published exponent, the same curve yields $59,192 on July 30. Verification is direct: the curve reproduces the three milestones Newhedge published in February —$62,700 on October 1, $64,400 on October 31, and $68,000 at year-end— with an error of less than 0.2%.
Timmer, however, is not calling a bottom. His reading: the speculative premium that pushed Bitcoin above $120,000 has evaporated, capital rotated first to gold and then to semiconductors, and without a liquidity catalyst, the price may crawl near support "for months before turning." An accumulation zone, in his vocabulary. Not a confirmed low.
What exactly does the power law model say?
The formulation is brief: Bitcoin's price grows proportionally to the time elapsed since the genesis block raised to an exponent that peer-reviewed study adjusts to 5.69 and that public curves, such as Newhedge's, place near 5.8. The horizontal axis is not a calendar date but the days of the network's life —6,417 on July 30, 2026— and both axes are on a logarithmic scale. In this chart, fifteen years of price action align into a straight line, and the bubbles of 2013, 2017, 2021, and 2024 appear as deviations around it.
The less obvious consequence of the exponent is that the model's floor never goes down. Its daily growth today is 0.09% —about $54 a day at current levels— and will decelerate over the years because the curve's derivative is the exponent divided by the age of the network. A floor that rises every day and a price that does not rise have a guaranteed meeting point.
The model entered a new phase in 2026. On June 29, Elsevier published the work of Santostasi and Stephen Perrenod in the journal Nonlinear Science, subjecting the theory to peer review: 5,696 daily prices between July 2010 and February 2026, with the curve explaining around 96% of long-term price variation. The authors do not provide price targets; instead, they list five conditions that would invalidate the model. The first is not "touching the floor": it requires a deviation of more than three standard deviations below the trend sustained for a year —at today's prices, on the order of $10,000 below the floor, maintained for twelve months—. The study's starting data ends in February, so this year's drop falls outside the sample: the first live test of the model as published science is happening right now.
Where is Bitcoin within the band as of July 30, 2026?
With the price at $63,999 and an all-time high of $126,080 set on October 6, 2025, the complete snapshot of the band looks like this.
| Band Level (Jul-30-2026) | Value | Distance to Price |
|---|---|---|
| Model Center Line (Newhedge, extrapolated) | $140,933 | −54.6% |
| All-Time High (Oct-6-2025) | $126,080 | −49.2% |
| BTC Price | $63,999 | — |
| Santostasi / Newhedge Floor | $59,192 | +8.1% |
| Fidelity Floor (Jul-12 reading) | $58,237 | +9.9% |
| Fidelity Floor (Jul-13 rounded reading) | $56,500 | +13.3% |
The center line deserves an asterisk. Newhedge published it at $121,733 on February 20; the $140,933 in the table is that same reading carried forward to July using the model's exponent, assuming the floor and center run parallel on a log scale. Under that assumption, Bitcoin is trading 54.6% below its central value. This depth has few precedents: the power law oscillator tracked by CoinDesk showed Bitcoin on June 3 to be cheaper than 95.6% of all historical readings, a percentile only seen during the March 2020 pandemic crash and the November 2022 FTX collapse.
It is worth separating the two statements the market often conflates. That Bitcoin is in its cheapest valuation percentile since FTX is a measurable fact. That this implies a bottom is an inference, and Timmer, for now, does not endorse it.
When does the floor reach the price if Bitcoin doesn't move?
The February headline ("breakdown in mid-December") omitted the mechanics: the model's "first breakdown in history" does not require a crash: it only requires the price to stay still long enough for the rising floor to catch up. The date of this encounter depends solely on the price at which the market freezes, and it can be calculated exactly.
| If Bitcoin stays flat at… | The floor reaches it on… | Days from Jul-30 |
|---|---|---|
| $59,192 (current floor) | Jul-30-2026 | 0 |
| $60,000 | Aug-15-2026 | 16 |
| $62,000 | Sep-20-2026 | 52 |
| $63,999 (current price) | Oct-25-2026 | 87 |
| $67,000 (Feb-2026 price) | Dec-16-2026 | 139 |
| $70,000 | Feb-4-2027 | 189 |
| $80,000 | Jul-8-2027 | 343 |
| $100,000 | Mar-29-2028 | 608 |
The $67,000 row explains the February headline: with Bitcoin at that level, the crossover fell on December 16, which is where the "mid-December" figure repeated for months came from. That calculation remains correct for its date. What has changed is the price: $3,001 less in market value has moved the appointment from December 16 to October 25. Every $1,000 Bitcoin loses from here brings the encounter forward by about seventeen days.
And it works the same in the other direction. A bounce to $70,000 pushes the expiration to February 2027; a move to $80,000, to the summer of 2027. Reclaiming the all-time high of $126,080 would buy margin until January 2029. The model does not demand that Bitcoin rise significantly: it demands that it rise slightly faster than 0.09% daily, which is the same as saying 3% per month.
There is a third path, the most uncomfortable for the model: that the price pierces the floor for a few weeks and then moves back inside. The formal invalidation threshold of the study is demanding —three standard deviations below the trend, sustained for a year— so a three-day wick below $59,192 invalidates nothing on paper. The market doesn't work that way: a brief piercing with a reentry guarantees months of discussion about whether that counted as a breakdown, with the model formally intact but the headline already written. It is a perfectly plausible scenario for the fourth quarter.
Why doesn't the power law work for Ethereum or the rest?
Think of a subscription service that opened fifteen years ago and hasn't touched its price, plan, or conditions since. If the product is frozen, revenue can only grow one way: new people coming in. And then the number of years it has been open becomes a reasonable proxy for everything else, because nothing else moves.
This is what makes Bitcoin unique in this model. Its monetary policy has been frozen since 2009: 21 million units, halving every 210,000 blocks, and no governance with the realistic capacity to change it. The only live parameter is adoption, and adoption grows regularly enough that time since genesis works as an axis. This property is called scale-invariant growth: the shape of the curve is the same regardless of which segment you look at.
The limit becomes obvious as soon as the product is no longer frozen. Ethereum has rewritten its issuance and fee regime five times in five years: EIP-1559 on August 5, 2021 (base fee burn), the Merge to proof-of-staking on September 15, 2022 (issuance cut by nearly 90%), Dencun on March 13, 2024 (blobs tank Layer 1 revenue and return ETH to net inflation), Pectra on May 7, 2025, and Fusaka on December 3, 2025 (gas limit from 45 to 60 million). A "days since genesis" axis applied to ETH mixes at least five economically distinct assets under the same ticker.
The consequence is seen in the price. ETH trades at $1,904 on July 30, 61.5% below its high of $4,946 from August 24, 2025, compared to Bitcoin's 49.2% drop from its own. No published power band serves as its floor. And there is statistical backing for this asymmetry: the work of Carlos Baquero and Raquel Menezes published on arXiv on May 20, 2026, tested nine series —Bitcoin's price, five on-chain network metrics, the NASDAQ Composite, the S&P 500, and gold— and found that Bitcoin's price is the only one where no single-component growth curve outperforms the power law fit.
What other Bitcoin valuation models remain standing in 2026?
The work of Carlos Baquero and Raquel Menezes (arXiv, May-20-2026) is also the most serious critique the theory has received. Three uncomfortable findings: the adjusted exponent varies by a factor of nearly three depending on which temporal origin is chosen; a stack of sigmoids beats the power law fit in AIC and BIC (standard model selection criteria that penalize adding extra parameters); and scale invariance tests from previous literature do not distinguish a power law from that stack of sigmoids on the same data. A February 2026 preprint also obtains a slope of approximately 4.2 —far from the model's 5.7-5.8— when the temporal axis is warped by network activity instead of calendar days.
| Model | Variable Used | Status as of Jul-2026 | 2029 Projection |
|---|---|---|---|
| Power Law — Center Line | Days since genesis | Peer-reviewed Jun-29-2026 | $395,700 |
| Power Law — Floor | Days since genesis | 8.1% from price | $166,200 |
| Logarithmic Bands (CleanSky) | Diminishing returns per cycle | Active | $178,000-$193,000 |
| Stock-to-Flow (PlanB) | Stock / annual issuance | Broken since 2022 | No usable projection |
| Metcalfe | Active addresses squared | Distorted by Layer 2s | — |
| Rainbow chart | Repainted log regression | Descriptive, not falsifiable | — |
The contrast in the table is what should most unsettle those citing the power law as a bullish argument. Our logarithmic bands model, built on the diminishing returns of each cycle (36x, 16x, 3.5x, 1.8x), projects between $178,000 and $193,000 for 2029. The power law places its floor for year-end 2029 at $166,200 and its center line at $395,700. In other words: what for our model is the central scenario for the end of the decade, for the power law is the worst-case scenario. And the famous "$10 million in 2045" attributed to Santostasi fits: the same curve yields $9.03 million in January 2045.
Of the others, Stock-to-Flow is the reference corpse. PlanB's model predicted price based on issuance scarcity and stopped working after the 2020 halving; ETF flows and corporate treasuries have since proven that the demand side moves the price far more than supply arithmetic —the 2026 structural supply deficit is explained by who is buying, not by how much is mined—. Metcalfe, which values the network by the square of its active addresses, has suffered since activity migrated to Layer 2s and addresses stopped approximating users. And the rainbow chart has been repainted so many times it no longer admits falsification.
What to watch in Bitcoin between August and December 2026?
Four specific markers, with levels and dates, to follow the outcome without relying on headlines.
- $59,192, and rising $54 a day. This is Santostasi's floor today. On September 1, it will be at $60,979; on October 1, at $62,643; on October 31, at $64,344; on December 31, at $67,920. A daily close below the day's reading would ignite public debate about a breakdown —formal invalidation of the study requires much more: three sigmas for a year— and for the first time, that debate would occur with a peer-reviewed article on the table.
- October 25, 2026. This is the crossover date if the price does not move from $63,999. It moves forward about seventeen days for every $1,000 drop and is delayed at a similar rate with every bounce. Update it mentally every time the price moves a relevant segment.
- A change in tone from Timmer. His July 12 position is explicit: accumulation zone, no liquidity catalyst in sight, no bottom call. The day Fidelity moves from "accumulation" to "bottom" without the price having bounced first, it will be the analyst who has moved, not the model.
- The divergence between the two floors. Fidelity and Newhedge publish lines separated by between $950 and $2,700 —$58,237 vs. $59,192, or $56,500 vs. $59,192 depending on the reading—. In a slow decline, this uncertainty band turns "the model broke" into a multi-week discussion about which line counted.
The question remains that no time-based model answers: if an asset's price can be predicted from its age, what role is left for what the network produces? This is the tension that already ran through the price/sales ratio applied to seven chains, where Bitcoin traded at 13,529 times its annual fees. One model looks at the clock; the other at the cash register. The next three months will grade them both at once.
Related articles: Four-year cycles and logarithmic bands: why $200,000 in 2029 is the ceiling, not the floor. ETFs, mining, and treasuries: Bitcoin's structural supply deficit. The price/sales ratio that no one applies to blockchains. Track your real Bitcoin position and the rest of your portfolio with CleanSky — wallet and portfolio tracking in a single dashboard, without connecting private keys.