1)       Truly a special situation.  Pink sheets. Quiet mgmt. No 06’ financials.  Filed 03-05 #s in July 07.  No 10K filed in 6 years

2)       Enormous operating leverage: 50-60% + incremental ebitda margins conservatively & ~25% y/y VAR verified growth allow high margin of error for forecasting off 05’ #s

3)       Bandwidth needs forecast to grow 75% + y/y next few years dropping to 50% y/y growth for 20 years due to video, VPN, IP phones.  ABVT most under radar play and industry has rationalized w/ 200 players 5 years ago to ~ 12 today.

4)       Irreplaceable assets & customers in top 14 metro markets such as NYC, SF, & DC trading at ~ 60% of replication cost.

5)       Potential to get taken out by LVLT or other.  Strong, net cash balance ~ 9% of mrkt cap 06E. at least $2/ shr NOL - potential $12

6)       Peers suggest business is going very well and mgmt is pleased.  Experts see assets as attractive and highly capable

Description:

AboveNet owns and operates 1.5 million fiber miles (3,852 route miles) in the top 14 United States metro markets and London to provide private optical network services to 1,300 + buildings comprised of large enterprises/institutions.  It has a highly attractive customer base including top financial & pharmaceutical institutions, 200 + Fortune 500 companies, universities, ISPs etc.  Though bandwidth is a commodity, ABVT provides a unique service by allowing its customers to privately and directly connect with their data centers, disaster recovery sites, satellite offices and other points over a private network rather than sending/receiving the data through multiple public internet peering points.  The advantage is security, reliability, scale, and costs.  ABVT charges for each unit of data that travels across its network and we believe it will be a large beneficiary of the greatly increasing trends in bandwidth needs. 

Truly a special situation that is quickly resolving with the 03-05’ financials filed in mid July 07 and an 06’ 10K on the way.  A long standing, inactive, & concentrated shareholder base (Kluge, billionaire) combined with a multi-year SEC investigation into accounting controls that was resolved in March 2007 are the reasons why ABVT has not filed a 10K since emerging from bankruptcy in 2003.  We believe this presents an enormous opportunity to get an edge on the business. Brief conversations with management suggest getting a 10K in order is a high priority and could be a few months/quarters  away, though they will not confirm.  While the SEC investigation is now old news, it gives us comfort that an ex-VP strongly believed it was not foul play but a poor legacy accounting system / atrocious paper documentation to match – “it was the wild west when it came to documentation.”    

Enormous operating leverage of 50-60% + incremental ebitda seems conservative given facts & research

           03’ bankruptcy projections estimated 50-60% incremental ebitda à included data center biz of lower op. leverage à sold 06’

           04/05 had an 84% incremental ebitda margin and 77% incremental gross margin vs. 86% incremental gross margin 03/04

           Ex-VP and VP atcompetitor felt 80% incremental for current subs was fair, 30% for new

           Peer CCOI has ~75% incremental ebitda blended, and ABVT should be lower but not far off

Although we do not have much financial information with which to project, there are factual points to hang our valuation hat.  04/05 shows ABVT had an 84% incremental ebitda margin.  In addition, bankruptcy projections from 2003 projected an average of 56% incremental margins till 2010 - a time when the industry was much less rational on pricing and when ABVT owned significant data center assets of lower incremental ebitda margins they did not dispose of till 05 & 06.  Former ABVT VP & a VP at a peer felt On-net (current customers / buildings where cap ex is spent) have 80% + incremental ebitda w/ new customers being 30%.  Current customers have high incremental margins b/c there is very little cost associated with each additional unit of bandwidth consumed.  New customers, on average, require some capital expenditure to construct the private network so the incremental margin is lower, but industry observers claim > 25% IRR.  Specifically, two senior managers at peers felt that of 25% revenue growth, 20% is new customers and 80% is existing customers.  This blend would imply 70% incremental ebitda margins - well below what we assume on avg for the next 5 years

Conversations with five senior managers at peers suggest ABVT is likely growing revenues 25% a year. A survey of industry experts suggested revenue is growing between 20 -40% for metro fiber businesses, with 25% being a mid point for the next couple of years.  Y/Y 04 to 05’ revs grew 16% for ABVT and we believe demand trends have increased greatly since this time.  In addition to comments from industry observers, the drivers of bandwidth growth outline below logically support this forecast. 

Bandwidth needs are expected to grow 50-80% CAGR for the next 10 years according to Bernstein Research and others. Video, increasing graphic quality, IP phones, VPN networks, disaster recovery requirements, & other drivers are creating tremendous bandwidth demands.  The CEO of Dell recently pointed out that YouTube alone consumes as much bandwidth as the whole internet did in 2000.

Relative MB needs:

Website                  1

Song                       15

Music video          43

Full screen TV      1731

Full screen HD      8654

The industry built for this growth in the late 90s but was too early; after nearly 10 years demand has finally caught up to supply.  After a massive over build, price wars, and a number of high profile bankruptcies (including ABVT) the industry has rapidly consolidated from 200+ players 5 years ago to ~12 today, and continues to consolidate.  The industry has been growing annual unit volumes 50%-75% w/ 20-25% price declines and with further consolidation expected this could improve, despite it being a commodity.

Irreplaceable assets & high quality customers in top 14 metro markets trading @ 60% of replication cost. During the late 90’s ABVT spent close to 5B building out its fiber network across the top metro markets in the US.  Our research indicates ~30% of the network ($ basis) is in Manhattan with  20% in DC. 

Metros served & estimated % of Network:

Manhattan                               30%

Washington DC                        20%

Dallas                                     10%

Chicago                                  10%

San Francisco                           8%

Los Angeles                             5%

* Remainder Phoenix, Miami, Atlanta, Seattle, Houston, Philadelphia, Denver, Boston, London

The barriers to build new networks in these cities is very high including permits, construction moratoriums, restricted conduits, disruption to business etc.  VAR from VP of a peer suggests “no one is going to re-dig Manhattan” to build new fiber networks and “many cities have build moratoriums, especially in restricted conduits like the Manhattan to NJ conduit under the Hudson River that ABVT owns.”  In some unique instances, Verizon has been adding fiber to Manhattan etc. at $100 / ft for the installation  (NY Times, 2006).

http://www.nytimes.com/2006/08/14/technology/14verizon.html?ex=1313208000&en=469743f49334baa6&ei=5088&partner=rssnyt&emc=rss

Small metros can run $35 to $50 a foot.  Applying an average of $65 / ft across ABVT’s 3,852 route miles would yield 1.32B of network value, or about 60% upside.  Although we do not typically view replication cost as a strong investment case for telecom businesses, the scarcity of these assets and the quality of the markets served excites us.  On top of asset valuation, take out multiples on peers have been increasing, reinforcing our view of scarcity. 

As ABVT resolves its current standing it could become an attractive take over target.  The industry has consolidated greatly in the last 5 years and the pace has not slowed, with Level 3 (LVLT) being the serial acquirer in the space.  Take out multiples have been climbing steadily, with recent transactions at 15x ebitda and 3x revenues for comparables.  Colleagues familiar with LVLT’s network believe ABVT assets could be highly synergistic, as you shutdown overlapping networks removing the costs but keeping the revenue.   Applying the 15x take out multiple to a medium case 08E ABVT ebitda the upside is ~ 80%.

In conclusion, ABVT is a complicated special situation with premier assets and a strong risk/reward that will be valued more in line with peers and possibly taken over by a competitor as financials are made current and management is allowed to share the story.  Upside to our valuation NOT yet factored in:

a)       NOL that could be worth $12 per share (we use $2 / share).  There are limitiation factors due to ownership changes and it's somewhat unclear.

b)       6m of annual cash ebitda from deferred revenue where costs are essentially sunk so the stream is pure cash generation (& growing)

c)       Potential the data center assets were producing much less ebitda than our estimates, as they were sold for $45m, which would only imply a 5x ebitda sales price when public peers haved traded at 15 -35x.  So we might be taking away too much ebitda from the sale. 

Risks:

       1) With 05’ being most recent financial year our estimates could be off but even if ebitda is only 50m à downside limited.  And our estimates are      not including cash ebitda of 5 to 6m ($7/shr upside on 15x multiple) from deferred revenue fiber lease IRUs or NOL. 

      2)   High operating leverage – valuation depends on growth

      3)  Broadband is a commodity – pricing could become more competitive and offset the growth in volume but + to see consolidation

           4) Lack of real communication with management à they will not speak in detail till a 10k has been filed

           5) Economic recession would hurt financial institutions, reducing work force, reducing IT spending needs à monitor this closely