I read it closely, because I spend a lot of my week talking with founders weighing whether to raise through a token. It is a good snapshot — and it rewards a careful read. Here is what the numbers actually say.

The tokenized asset market — real-world assets brought on-chain, not counting stablecoins — sat below $3 billion in mid-2024. Today it is around $34 billion. Ten times larger, in under two years.

It is a real market. It is also still an early one. The global bond market is over $140 trillion; tokenized bonds are about $15 billion of that — roughly 0.01%. So we are looking at the first sliver of something, not a finished shift.

And the $34 billion is concentrated. Tokenized US Treasuries and commodities make up about two-thirds of it. Within commodities, almost everything is gold — close to $5 billion of a $5.1 billion category.

That concentration is not an accident. The assets that moved on-chain first were the ones easiest to move.

Gold is a clean example. It is:

A US Treasury is similar — clear price, deep demand, simple ownership. You do not need to reinvent these assets to put them on-chain. You wrap them, and they work.

The assets that took years to scale were the complicated ones. Tokenized venture capital took more than seven years to reach $1 billion. Real estate and operator cash-flow assets — where ownership is layered and trust is the hard part — are still a small share of the market . Not because they matter less, but because they are harder (believe me, it is hard, see what we are doing @ Propex.app).

This is the part of the report worth slowing down on.

a16z draws a line between two things that both get called "tokenization." Most tokenized assets today are held on-chain but not actually used on-chain. Of $15.2 billion in tokenized bonds, only about 5% is deployed in DeFi applications. The rest sits there as a digital record.

In a16z's own words, much of what we call tokenization today "is actually closer to digitization." A separate measure, from Pantera, reaches the same conclusion — it ranks more than three-quarters of tokenized assets in its lowest tier of "on-chain-ness."

The distinction is simple, and worth keeping:

Moving a record on-chain without using what a blockchain can do is a little like moving a filing cabinet to the cloud — the storage is more modern, but the document still behaves the same way. That describes most of the market today. It is a reasonable first step. It is not the finished product.

If you are a founder thinking about a token, the useful takeaway is simpler than any forecast.

A token earns its place when it lets an asset do something it could not do before — be owned in fractions, settle instantly, be governed by the people who use it, be built on by others. If a token only creates a digital record of something that already worked, that is digitization, and it does not need a token.

The forecasts, for what they are worth, vary widely. McKinsey puts the market at $2–4 trillion by 2030; BCG and Ripple at over $9 trillion; Standard Chartered above $30 trillion by 2034. They disagree on the number, but they agree on the direction.

The $34 billion we have today is the accessible part — the assets that were straightforward to bring on-chain. The harder, more interesting work — bringing complex, real-economy assets on-chain in a way that genuinely changes how they move — is still ahead.

For anyone building now, that is where the room is.

Over the next few issues I'll go deeper on what separates a real token network from a digital record, and how a token sale can be done properly.

If that's useful to you, stay with me.