What Moves It, and What Only Looks Like It Does

Where the number comes from

The same way bitcoin’s does. Ether trades on many venues with separate order books, and every
quoted price is an aggregate across some selection of them. Aggregators differ in which venues
they include and how they weight them, which is why two sites can show slightly different
figures at the same instant without either being wrong.

One difference is worth knowing. A large share of ether trading happens on the network itself,
through pooled contracts rather than order books. Those trades price against whatever sits in
the pool at that moment, so during fast moves the on-chain price and the exchange price can
separate briefly before the gap is arbitraged away.

The supply side moves, and that is the whole difference

Bitcoin’s issuance is fixed in advance and nothing about it responds to demand. Ether’s does.
New units are issued to validators for securing the network, and a portion of every
transaction fee is destroyed rather than paid out. The two run against each other
continuously.

When the network is busy, the amount destroyed can exceed the amount issued and total supply
falls. When it is quiet, issuance wins and supply grows. This is why the deflationary label
attaches and detaches depending on when somebody checked, and why a supply argument about ether
has to name a period to mean anything.

The practical consequence is that usage and supply are linked. On bitcoin, an argument about
adoption and an argument about scarcity are separate. Here they are the same argument observed
from two ends, which makes the asset easier to model and much easier to model badly.

What staking does to the float

A large quantity of ether is committed to validators, and that quantity is often presented as
supply taken off the market. The mechanism is real and the conclusion is only partly right.
Deposits can be withdrawn, subject to a queue, so the lock is a delay rather than a removal.

More importantly, a substantial share of staked ether sits behind arrangements that issue a
tradeable token representing the deposit. Somebody who wants out sells the token instead of
unstaking. The economic exposure changes hands immediately while the deposit stays put, which
means counting the full staked total as illiquid counts the same supply twice.

The arguments that recur, and what each rests on

The fee-burn argument says activity shrinks supply, which is mechanically true and says nothing
about whether activity will continue. The settlement argument says other networks post their
data here and pay for it, which is checkable and currently modest relative to the whole. The
collateral argument says ether is the reserve asset of the applications built on it, which is
the strongest of the three and the hardest to size.

Against all of them sits one structural risk that has no bitcoin equivalent: the same
applications can run elsewhere. Where they run is a decision made by developers and users
rather than by a protocol, and it is the variable most capable of invalidating the other three
at once.

Worth adding that the burn and the issuance are both published continuously, so the net change
over any period is a matter of record rather than estimate. Anyone making a supply argument in
either direction can be asked which window they measured, and the answer either exists or the
argument does not.

Why comparisons to bitcoin usually mislead

The two are routinely charted against each other as though they were competing versions of one
thing. They are not doing the same job. One is a settlement asset whose case rests on a fixed
schedule; the other is closer to a claim on the use of a network, priced by how much that use
is worth.

A ratio between them is still a useful market signal, because capital does rotate between the
two and the ratio records it. It is not a verdict on which design is better, and reading it as
one is how a market-flow observation gets promoted into a technical argument it cannot support.