IPv4 Market Transfer Prices: update

A couple of years ago I used data from the kind folks at IPv4Auctions.com to plot trends in IPv4 market prices. They now publish that data,so I thought I'd see what prices are looking like. The same caveats from the previous post apply.

Here's a chart plotted from that data:

I've added a couple of trendlines. The two purple extrapolations are a linear and a second-order polynomial (i.e., quadratic) projection based on the scattered plots of /24 prices, and the orange is a linear projection based on /22 prices. I did two projections for /24s because it's important to show that there's a range of possible futures (and this dosn't necessarily encompass the actual future); it's impossible to know ahead of time which curve will be the most predictive. I added the /22 projection to see that prices are changing at different rates for differet sizes of blocks.

When I look at this chart, it seems to me that the price-per-address has been all over the place in the past, and although the average price has risen, the pricing is clustering around $15/address (for /24). This suggests a maturing, or at least a more well-informed, market: since data like this is public, the market is more transparent, and one would expect similar prices for similar blocks. It also looks to me like prices for larger blocks are rising faster, and in time may even converge, so that the price per address is the same, regardless of block size.

Let's dig into that further. It's hard to see in the first chart because there are so many data points, so now let's zoom in on /19 - /22 since January 2016:

Does it look like a step function to you? Not a perfect one; for /22, it looks like the prices bounces a bit. Other ranges seem to be steady for a time, then suddenly rise by US$0.50 or US$1.00. Once the price steps up, transactions below that price are very rare, though they are occasionally higher.

My guess would be that prices are being set by sellers: an organization makes a block of addresses available for sale, and sets their minimum price. So an organization made a dozen /22s available for sale in March 2016, and sold them all in a month for $9.00 per address. A couple of sales closed then around $11, $13, and even $15.75, so the seller(s) raised their price by a dollar, and six months later by another two dollars. Does it look like there were fewer transactions at the higher price? That would say something about the demand curve.

The price per address for larger blocks has fairly consistently been US$0.50-1.00 less per address than the next larger size. That may have changed; look at the recent clustering of /20 and /22 prices. I would guess we're due for a step in /22 prices (US$12.50 is too low when some buyers are bidding US$13.50). It does look like prices are converging, though. With transactions made simple by IPv4auctions.com and others, there's less reason to settle for a lower price per address for one larger block: the incremental cost to dividing them up is smaller than it once was. As prices rise, brokers may find it profitable to trade in mid-size blocks, though I have the impression that /16s are still the bread and butter of the industry. Importantly, everything in this paragraph is speculation.

I'm certainly not ready to call a top to the market. Referring back to the projections in the first chart, I'd expect prices to continue rising for a while.

As always, I'm very interested in others' interpretations of this data, and projections about the market. Follow and Tweet me!

Lee Howard
23 June 2017