Celestia has a technical problem, which I observe,

Celestia thinks that an Ethereum expansion network similar to Layer 2 can publish newly generated data to the Celestia chain instead of directly to Ethereum , which can save more than 90% of handling fees.

But,

None of Celestia's DA layers (Data Availability layers) can reach the same level of availability assurance as Ethereum, and a certain degree of trust assumption as well!

The whole Ethereum Layer 2 liquidity is concentrated on orthodox projects such as Arbitrum or Optimism, and the main versions of these projects will surely not be integrated with Celestia.

Arbitrum Orbit is just a minor version similar to Arbitrum Nova, you think it can compete with OP, Base or zkSync? Even if Celestia can attract such these "secondary Layer 2", it doesn't seem to bring much value capture. (Arbitrum Nova currently has only ~$22M in TVL).

There will also be EIP-4844, where Ethereum will be even more powerful.

TIA is ambitious but there are lot of things to do. If TIA is adopted on a large scale, the Ethereum Foundation will face competitive pressure.

And we need it for better blockchain experience.

Time will tell.