IPLs in Demand; U.S. Bandwidth Pricing Declines

Posted: 10/1999

IPLs in Demand; U.S. Bandwidth Pricing Declines

RateXchange’s bandwidth market was moderately active last month. Much of the supply was
in U.S. domestic private lines while most of the trading was in international private
lines (IPLs). Tier 3 wholesale carriers and even Internet service providers (ISPs) are
grabbing IPLs as soon as they come to market to expand their reach of direct connections.

Graph: DS-3 Bandwidth RTBX*RPI and RTBX*RFP-360 Indices

U.S. domestic bandwidth pricing recently declined to an average East Coast-West Coast
rate of approximately $27,000 per month per DS-3. The Revealed Price Index (RPI) as of the
end of August has declined below the one-year Revealed Forward Price (RFP-360) indexed
rate set in July. This means that last month carriers were trading contracts of fixed
durations with set prices that when averaged to a rate one year from now, were above the
spot price for the same good today.

In other words, DS-3 bandwidth sellers won in July if they locked in future prices at
spot rates through longer-term contracts. Buyers fared well if they delayed purchasing, or
at least agreed to shorter-duration contracts.

Minutes market activity surged during August. The number of new circuit-based minutes
offers surged from 23 in July to 309 in August. Much of the minutes trading activity is
concentrated in low-volume peripheral country routes such as Lebanon and Vietnam as
margins in high-volume routes have all but disappeared.

Voice over Internet protocol (VoIP) market activity continues to be active as new
cost-competitive offers are snatched up as soon as they hit the exchange. In the age of
declining margins, it seems more carriers are considering VoIP as not only cost-effective,
but of competitive quality for their end users.

We can expect a surge of trading activity in the fall. RateXchange is eliminating its
commissions for lead-generation–being the first to facilitate free bandwidth trading.
Margins are indeed declining.