Dec 31: An Analysis

An hour by hour recap of what occurred

As written a few days ago, 2022 kicked off with some controversy around a massive increase in voting power on the Internet Computer, which reduced the yield of all neurons considerably. Let’s review what happened:

• The NNS started the day off with 310M in voting power and that remained relatively constant until 17:42 (GMT), at which point the voting power jumped by about 6M.

• Between 17:42 and 17:54 the voting power jumped by an additional 4M.

• By 18:06 the voting power had increased by an astounding 64M to 384M.

• At 18:42 the voting power increased by 2M, followed by 3M at 18:54. The voting power added another 4M at 19:06, 2M at 19:12 and 2.3M at 19:24. Over the next thirty minutes, the voting power increased gradually to the final value of 402M, which is where it has remained over the last few days.

So what happened?

Let’s review what we know:

• The voting power of the NNS increased by 92M on Dec 31st in a process that lasted two and a half hours and ended just before 20:00 GMT.

• We know that the increase is real (ie, not an error on dashboard.internetcomputer.org) because recent voting has been close to 400M.

• We know that at least 30M of that increase in voting power is caused by about 15M ICP in genesis accounts being (a) claimed and (b) quickly increasing the dissolve delay to 8 years.

• We know that ic.rocks currently “knows” about neurons totaling 311M in voting power. We also know that ic.rocks doesn’t “know” about all neurons since there isn’t indexing of neurons on the NNS yet. Historically, the amount of unknown neurons has been about 5% of voting power. However, now it’s much greater and appears to be the same 5% plus an additional 64M in voting power.

So what does this all mean? It’s clear that there was A LOT of staking happening on the NNS on December 31st that ic.rocks doesn’t have visibility into due to the indexing of neurons. This staking should total about 62M in voting power (31M ICP if it is all staked into 8-year neurons). Interestingly, there are many transactions that occurred on December 31st that were over 1M ICP. It’s not clear whether this ICP ended up being staked (although it seems likely), but the transaction history leads back to Neuron 4000 which dissolved shortly after genesis.

Looking through large transaction data (transactions greater than 1M ICP) shows that there were 6 transactions that occurred during the time frame discussed above, which totaled 32M ICP. Unless there were MANY smaller transactions, this suggests to me that the 62M in unknown voting power is cause by these 32M ICP being staked for close to 8 years. If that’s the case, then the total number of ICP in ~8 year neurons is currently 96.7M ICP (20.3% of total supply)!

If anyone wants to do some sleuthing, the receiving addresses of those 6 large transactions were:

• 27bbe9b4f0b00e4b6fe3fb39328358cf82031e82014e0cd0ae60983cc92008f5

• 9a60a7bcf722217debea223fb83e28d3686b79e65d8f785c33b11466c9fdf0fb

• 9666e607a64b9213a40e65bbea15e1e97c907e5536a1443f4df277b1ca390a64

• dce0dcbdb83d339b3beb6967e10f369a5e149f9c5db9cd16051dfbc03dd8d303

• fae073332cb8c6722b6b9a824ec42323d87d07d388f6b0db85b1e521ea0b85e2

• 191ff3f513339860f85e0e4f77a90ec8404835f80f5f6dbcd3db43db4826f0d5

What does this mean moving forward?

With the claiming of the 15M ICP, only 16M ICP remains unclaimed (note - I had incorrectly reported 35M ICP remained unclaimed in the previous article). The ultimate fate of that 16M ICP is still unknown.

In addition, there is now 62M additional voting power on the NNS that I am unable to see until indexed neurons occurs on the NNS. This additional voting power looks like it came from neuron 4000 through many transactions. If anyone wants to play detective, I’d love to know if there is any idea what entities own these accounts or what their purpose is and how much ICP is still in these accounts.

One final note - since neuron 4000 had originally moved these ICP tokens 6 months ago, the 32M ICP that got staked would come from what I considered to be the circulating supply of ICP. While these new neurons are most likely owned by only a few entities, and thus this move was bad for decentralization of the NNS, it also is the exact behavior that many of us want moving forward - tokens taken from the circulating supply and staked for (presumably) 8 years.

To recap

• 15M ICP was in unclaimed genesis accounts that got claimed and quickly staked for 8 years.

• We don’t currently have visibility into the other 62M increase in voting power, however it seems extremely likely that it came from 6 transactions totaling 32M ICP that originated from Neuron 4000.