Data check · Own data

Is Bitcoin a bubble?

Bitcoin has fallen more than 50% 8 times and reached a new all-time high after 7 of them. Across 16 years of our own data, this page checks five traits that distinguish Bitcoin from a one-off speculative bubble such as the 1637 tulip mania.

Crashes > 50%
7 of 8 recovered
recovered = new all-time high afterwards · one still open
Loss over 4 years
Never
worst entry +31%
100% of windows in profit
Volatility
3.4× lower
annualised, early vs. now
2010
159%2026
46%
Network security
958 EH/s
hash rate; line = difficulty since 2009 (log)

1 · Crashes and recoveries

The chart shows, for every day, how far Bitcoin sat below its highest price up to that point. There were deep crashes, the largest at 93%. 7 of 8 crashes of more than 50% were followed by a new all-time high.

0%-25%-50%-75%201020122014201620182020202220242026ATH
Biggest drawdown
-93%
below the highest price so far
0%-100%
Below all-time high today
-33%
all-time high October 2025
0%-100%
Longest recovery
1,176 days
about 3.2 years to a new high
04 years

2 · The 4-year rule

The chart shows the worst return after 1, 2, 3 and 4 years across all possible buy dates. Over 4 years no window in the data so far ended in a loss; the worst entry ended at +31%.

0%-83%1 year73% in profit-68%2 years84% in profit-35%3 years99% in profit+31%4 years100% in profit

Bars = worst possible return for that holding period, across all buy dates.

3 · Falling volatility

Annualised volatility per year has fallen from about 159% in the early years to roughly 46%.

0%86%172%'10'12'14'16'18'20'22'24'26

4 · A 16-year trend

On a logarithmic scale the price follows a straight line closely over 16 years (power law, R² = 0.96). This describes the past and is not a price forecast.

$1$10$100$1k$10k$100k$1000k201020122014201620182020202220242026PricePower-law trend

5 · Proof of work and network security

Behind every Bitcoin block sits real, measurable computation; hence the name "proof of work". That work (the "difficulty") has climbed from 1 in 2009 to about 133 trillion today. The cost of attacking the network has risen with it.

11K1 M1 Mrd1 T1000 T201020122014201620182020202220242026Mining difficulty
Hash rate (now)
958 EH/s
compute power
Difficulty
133 T
×133T since 2009
Attack cost/hour
$140 M
51% attack
Attack model
51%
rough estimate

The verdict

None of these five observations predicts a future price, and Bitcoin remains risky and volatile. The tulip-bubble pattern (one rise, one collapse, no recovery) does not appear in 16 years of data. The data show repeated recoveries, falling volatility, a stable long-term trend and a difficulty that has risen sharply since 2009. The price currently sits 33% below the all-time high of October 2025.

Methodology & data

All price figures come from our own price database (daily close since 2010, USD as the historically traded price). Drawdown = distance from the highest daily close so far. 4-year rule = return across all buy dates for each holding period. Volatility = annualised standard deviation of daily log returns per year. Trend = linear regression of log(price) over log(days since genesis). Network data from mempool.space, refreshed hourly.

Frequently asked questions

01Is Bitcoin a bubble like tulip mania?

Tulip prices fell in 1637 and never recovered. Since 2010 Bitcoin has crashed more than 50% 8 times and reached a new all-time high after 7 of them; the deepest drawdown was 93%.

02Has holding Bitcoin ever lost money over the long run?

No. In the data since 2010 no 4-year window ended in a loss; the worst entry ended at +31% after 4 years. This describes the past and says nothing about future periods.

03Isn't Bitcoin getting more volatile?

No. Annualised volatility has fallen from roughly 159% in the early years to about 46%. That points to a maturing market.

04What is "proof of work" and why does it matter?

The Bitcoin network is currently secured by about 958 exahash per second, which is real, measurable computation and energy. A 51% attack would cost roughly $140 M per hour. This measures network security, not the valuation of the price.