Methodology

1. Find Average Fee on Starknet
2. Get the number of blocks sent to L1 to update the state
3. How Starknet Helped Ethereum Fees

Short Definitions

Generally, a transaction’s cost depends on several factors

• Data - the amount of data sent to L1
• Cost of computation
• Storage cost
• L2 network transmission cost for the transaction

the Starknet mechanism lets users/applications use the sequencer to get an estimate for the fee required to accept the transaction:

Blocks To L1

Starknet bundle multiple transactions together, settle them out of the main network and send only the resulting state to Ethereum.

Starknet Mainnet began sending blocks to L1 in November 2021 to update the state. On average, 12 blocks have been sent to L1 each day, but in May the number was up to 33. A reason for the increase could be the May market crash.
A total of 2828 blocks have been sent to L1 as of right now.

Transactions

let's see how many transactions were done in Starknet and sent as bundles to L1.

the green line shows the Starknet transactions, in the first week the daily volume was under 500 transactions, but we can see an increasing trend, reaching over 5K per day in June.

Additionally, Ethereum L1 transactions are depicted with a blue line and have been falling since November, from 1.3 million to one million per day, as a result of the Bearish market.

Fees

Let's examine how the transaction fees on L1 and L2 differ.

This Logarithmic Chart shows the Average Fee per transaction on L1 and L1.

In the beginning, L2 Fees were around $5, but today, they have dropped significantly to around 10 cents, which is significantly less than L1 Fees.

The L1 Fee has also decreased due to market conditions and fewer transactions and is currently under $10.

Let's compare these better with Ratio:

We can observe that L2's fees are significantly lower. Users greatly benefit from paying considerably less money for fees because the Fees on Ethereum are 100X–50X higher than Starknet L2.

How Helped Ethereum and Users?

As stated, Starknet groups several transactions into a single block and sends it to L1 as a single transaction, resulting in a single fee on L1 for all of those transactions.

Let's see how much Fees were to be paid if all these transactions were supposed to occur on L1.

The blue bars represent the Fees if all transactions were done on L1 instead of L2, and the orange bars represent the Fees currently paid for these L2 transactions that have only one updated state transaction on L1.

L1 fees are substantially higher than L2 fees, as shown by the blue bars.
Thus, the Ethereum network will be less busy, computing will be lower, so L1 fees will be lower, and the network will be faster as a consequence.

How much Fees are Saved

Based on the lower fee on L2, let's examine the amount of Dollars paid overtime for L2 transactions, and compare it with the fees if those transactions were supposed to take place on L1 with much higher fees.

The brown Area is dollars Paid for Fees on Starknet already, and the blue area shows the Fees for these transactions if they were supposed to take place on L1.

As of today, $3.24M has been paid for L2 transactions. It would be about 33X higher if there hadn't been an L2 Starknet, which would have cost $105.4M to pay as fees for these transactions.

so Starknet saved about $100M as Fees for users.

Conclusion

• Blocks sent to L1 were usually 12 per day, but in May they reached 30.
• Recently, Starknet's daily transactions exceeded 5k, and they peaked at 10k on some days.
• In recent weeks, Starknet fees have averaged under 20 cents, about 100X-50X lower than L1.
• For Starknet transactions, $3.2M has been paid as fees.
• Starknet saved about $100M in fees and helped Ethereum become faster and cheaper.