Unleveraged FCF yield on Fiber

[LearningMachine]

I remember some of the cable investors have mentioned in the past how laying fiber doesn't make economic sense in today's high interest rate environment.

 

The core message I took away from that discussion was that fiber pays at best only about mid-single digit unleveraged yield, and that doesn't make sense when interest rates are higher than that.  @Spekulatius, if memory serves right, I think you might have been one of the participants in that discussion.

 

I see people buying NNI and BOC where they are allocating capital towards fiber install. 

 

Can we jointly figure out what is the unleveraged FCF yield range on laying out fiber?

 

 

Edited February 1 by LearningMachine

[dwy000]

Other than BOC's numbers the other one you can use to measure is Tucows (Ting Fiber).  They're pretty similar in that they go into very small, newer communities (often new builds).  The actual returns are still TBD because they generally assume a very strong level of uptake and pricing that takes years to normalize.  They often also don't take into account the response from incumbents (if there are any) with promo pricing and retention offers that were never factored into the planning models.  I haven't seen any that have hit their projections or have turned cash flow positive to the level where you can judge long term capital returns. 

 

Edit - I took a quick look at Tucows.  Thru 3Q they state they have invested $329m of capex in the biz since 2015 (this includes a couple small acqns).  For just the past 3 yrs they will have EBITDA losses of another $80m or so.  Haven't looked further back.  So all in, they're about $425-450m of spend.  Right now they have 41k subs.  Let's say they can get that up to 60k on the current builds.  They seem to charge about $100/mo with 60% gross margins.  

 

$450m of spend for 60k subs is about $7100/subscriber.  At 20% op margin that's $240/yr which is a 3.4% ROIC.

Edited February 1 by dwy000

[KJP]

2 hours ago, dwy000 said:

Other than BOC's numbers the other one you can use to measure is Tucows (Ting Fiber).  They're pretty similar in that they go into very small, newer communities (often new builds).  The actual returns are still TBD because they generally assume a very strong level of uptake and pricing that takes years to normalize.  They often also don't take into account the response from incumbents (if there are any) with promo pricing and retention offers that were never factored into the planning models.  I haven't seen any that have hit their projections or have turned cash flow positive to the level where you can judge long term capital returns. 

 

Edit - I took a quick look at Tucows.  Thru 3Q they state they have invested $329m of capex in the biz since 2015 (this includes a couple small acqns).  For just the past 3 yrs they will have EBITDA losses of another $80m or so.  Haven't looked further back.  So all in, they're about $425-450m of spend.  Right now they have 41k subs.  Let's say they can get that up to 60k on the current builds.  They seem to charge about $100/mo with 60% gross margins.  

 

$450m of spend for 60k subs is about $7100/subscriber.  At 20% op margin that's $240/yr which is a 3.4% ROIC.

 

What is the incremental OpEx/SG&A after a new subscriber connects?  Assuming this is just broadband, and not video, it should be very low.  So, incremental operating margin, which is probably the best way to assess incremental new build CapEx, is likely much higher than 20%.   Also, 60% gross margin on broadband seems low.  Tucows claims much higher.  See slide 25:  https://ir.tucows.com/wp-content/uploads/2023-Q3-TCX-results-investor-deck.pdf Put those two things together and you'll get much higher (claimed) margins.  See slide 26.

 

I do want to note that I've never been able to reconcile that slide (which they've long had in their presentations) with the segment numbers.  The CapEx/passing in the segment financials always seems higher than what they put on that slide.  [The slide is showing cost per passing, not cost per subscriber, and doesn't include the cost of the drop to the house of a new subscriber.]

 

I also agree with you that there are other costs besides CapEx that must be incurred to get the system to steady state/50% penetration (marketing, etc.) so the EBITDA losses should be included, not just the CapEx, when looking at returns on capital.

Edited February 1 by KJP

[Spekulatius]

The returns depend on uptake and cost to build. I think you can charge $70/month for broadband / give or take. Thats $840 in revenue/ year in re Neue at 80% margins. Uptake is probably 40-60% depending on competitive situation.

 

If you assume 50% uptake, that gets you to $336 annual EBITDA and a 15% min requirement assumes the all in build cost needs to be $2240/connected  location. Thats my napkin math.

 

One thing that worries me is that some of the subsidized  ACAM funded builds seems to have costs of $6k or more per connection. Even with a 50% subsidy (usually, it’s less) , that still a net  $3k /connection and works only with very high uptakes. Some ACAM builds have cost up to $10K/connected home and I think those will be uneconomic.

[Munger_Disciple]

55 minutes ago, Spekulatius said:

The returns depend on uptake and cost to build. I think you can charge $70/month for broadband / give or take. Thats $840 in revenue/ year in re Neue at 80% margins. Uptake is probably 40-60% depending on competitive situation.

 

If you assume 50% uptake, that gets you to $336 annual EBITDA and a 15% min requirement assumes the all in build cost needs to be $2240/connected  location. Thats my napkin math.

 

One thing that worries me is that some of the subsidized  ACAM funded builds seems to have costs of $6k or more per connection. Even with a 50% subsidy (usually, it’s less) , that still a net  $3k /connection and works only with very high uptakes. Some ACAM builds have cost up to $10K/connected home and I think those will be uneconomic.

 

It takes a long time for a new entrant to get anywhere close to 50% penetration (if ever) and that time dramatically reduces IRRs. It took CHTR & CMCSA more than 25 years to get to 50% penetration in broadband when they were monopolies for high speed internet & already had a customer relationship thru' cable video & phone. The new entrant has to fight for every sub against an incumbent who is offering an equivalent broadband service plus cell phone bundles. It is a very tough fight & I think you will see most overbuilders go under or sell to the big boys. 

I used to think overbuilders were the biggest threat to incumbent cable but it turned out the real threat came from FWA despite its far poorer service and that appears to be moderating. The reason for the initial uptake of FWA was the bundling with cell phones IMO & lack of mobile traffic on 5G bands (no longer the case).

Edited February 1 by Munger_Disciple

[dwy000]

1 hour ago, KJP said:

 

What is the incremental OpEx/SG&A after a new subscriber connects?  Assuming this is just broadband, and not video, it should be very low.  So, incremental operating margin, which is probably the best way to assess incremental new build CapEx, is likely much higher than 20%.   Also, 60% gross margin on broadband seems low.  Tucows claims much higher.  See slide 25:  https://ir.tucows.com/wp-content/uploads/2023-Q3-TCX-results-investor-deck.pdf Put those two things together and you'll get much higher (claimed) margins.  See slide 26.

 

I do want to note that I've never been able to reconcile that slide (which they've long had in their presentations) with the segment numbers.  The CapEx/passing in the segment financials always seems higher than what they put on that slide.  [The slide is showing cost per passing, not cost per subscriber, and doesn't include the cost of the drop to the house of a new subscriber.]

 

I also agree with you that there are other costs besides CapEx that must be incurred to get the system to steady state/50% penetration (marketing, etc.) so the EBITDA losses should be included, not just the CapEx, when looking at returns on capital.

I was just back-of-the-enveloping it so you're probably right on some of these.  The 60% gross margin is what they're currently reporting (3Q23) and gibes with the low end of the slide deck (p 26).  

 

Like you, I also struggle to reconcile their numbers.  They literally state in their quarterly KPI page (footnote 2) that they have spent $329m in capex and acquisitions (net of writeoffs) to date and yet they only have 114.5k passings.  That's $2900 per passing and yet their investor deck says their cost to build is $1650 per address.  It's a huge difference.  Likewise, where they get $1000 annual gross margin per subscriber is beyond me - especially when the same page says they're charging $89/mo.  Even at 75% margin that implies they're charging $111/mo for every single sub which seems ridiculously high, especially if you need to get to 50% uptake.  

 

These business plans always look like they were drawn up by an investment banker trying to sell the deal.  The reality never comes close to the plan and the only number that is higher than expected is the spend.  But mgmt can't now say they were wrong so they keep expanding and say it's not wrong, just delayed.

[oscarazocar]

4 hours ago, LearningMachine said:

I remember some of the cable investors have mentioned in the past how laying fiber doesn't make economic sense in today's high interest rate environment.

 

The core message I took away from that discussion was that fiber pays at best only about mid-single digit unleveraged yield, and that doesn't make sense when interest rates are higher than that.  @Spekulatius, if memory serves right, I think you might have been one of the participants in that discussion.

 

I see people buying NNI and BOC where they are allocating capital towards fiber install. 

 

Can we jointly figure out what is the unleveraged FCF yield range on laying out fiber?

 

 

 

NNI is not really allocating much capital towards fiber.  They invested $450 million in capex and net losses in Allo (their fiber business) through 2020, then in late 2020 sold a majority stake in the business to a private equity firm (SDC) for $260 million and retained $130+ million in preferred stock and 45% equity interest (at the time valued at $129 million), so they took out most of their invested capital.  They have contributed minor amounts since, including $8 million in 2023 Q1, but Allo raised debt to fund its growth going ahead and I think there will likely be minimal future capital conributions.

[Munger_Disciple]

3 minutes ago, oscarazocar said:

 

NNI is not really allocating much capital towards fiber.  They invested $450 million in capex and net losses in Allo (their fiber business) through 2020, then in late 2020 sold a majority stake in the business to a private equity firm (SDC) for $260 million and retained $130+ million in preferred stock and 45% equity interest (at the time valued at $129 million), so they took out most of their invested capital.  They have contributed minor amounts since, including $8 million in 2023 Q1, but Allo raised debt to fund its growth going ahead and I think there will likely be minimal future capital conributions.

 

One could argue they practically exited fiber. 

[Intelligent_Investor]

Fiber buildout makes a lot of sense for large telcos trying to replace copper networks to reduce opex, as you are trading a high upfront cost for lower run rate opex, which only really pays off if you have a large enough scale. Not sure how the economics shake out for smaller players

[oscarazocar]

49 minutes ago, Munger_Disciple said:

 

It takes a long time for a new entrant to get anywhere close to 50% penetration (if ever) and that time dramatically reduces IRRs. It took CHTR & CMCSA more than 25 years to get to 50% penetration in broadband when they were monopolies for high speed internet & already had a customer relationship thru' cable video & phone. The new entrant has to fight for every sub against an incumbent who is offering an equivalent broadband service plus cell phone bundles. It is a very tough fight & I think you will see most overbuilders go under or sell to the big boys. 

I used to think overbuilders were the biggest threat to incumbent cable but it turned out the real threat came from FWA despite its far poorer service and that appears to be moderating. The reason for the initial uptake of FWA was the bundling with cell phones IMO & lack of mobile traffic on 5G bands (no longer the case).

 

NNI has disclosed Allo metrics over time and the penetration has been reasonably impressive.  In 2017 they had 71k passings and 20k households served for 29% penetration, then ramped up and by 2020 had 150k passings and 59k households served for 40% penetration.   I spoke with them several years ago and they indicated they would hit their financial targets with 50% penetration and in the 2022 letter they indicated results are ahead of initial underwriting expectations.  My guess is that their Lincoln deal was probably a pretty good one given their long-time presence and deep connections there.

[Spekulatius]

1 hour ago, Munger_Disciple said:

 

It takes a long time for a new entrant to get anywhere close to 50% penetration (if ever) and that time dramatically reduces IRRs. It took CHTR & CMCSA more than 25 years to get to 50% penetration in broadband when they were monopolies for high speed internet & already had a customer relationship thru' cable video & phone. The new entrant has to fight for every sub against an incumbent who is offering an equivalent broadband service plus cell phone bundles. It is a very tough fight & I think you will see most overbuilders go under or sell to the big boys. 

I used to think overbuilders were the biggest threat to incumbent cable but it turned out the real threat came from FWA despite its far poorer service and that appears to be moderating. The reason for the initial uptake of FWA was the bundling with cell phones IMO & lack of mobile traffic on 5G bands (no longer the case).

FYBR has presentations in their IR section with uptake rates for the fiber builds and they do plan to get to 40% in a couple of years and track the progress to some extend. This applies to areas where there is no or at most one broadband competitor and they are mostly targeting DSL customers.

But yes, the uptake time ramp is critical for the IRR, if it really takes 10 years + to get to 40% penetration, then the IRR are probably not that great. I think any fiber build that does not get to 49% penetration is probably it going to have a great returns because I have not seen many fiber builds with much less than $2K cost per connection (all in cost, not what some of those fiber overbuilders companies are posting ).

Edited February 1 by Spekulatius

[Munger_Disciple]

21 minutes ago, oscarazocar said:

 

NNI has disclosed Allo metrics over time and the penetration has been reasonably impressive.  In 2017 they had 71k passings and 20k households served for 29% penetration, then ramped up and by 2020 had 150k passings and 59k households served for 40% penetration.   I spoke with them several years ago and they indicated they would hit their financial targets with 50% penetration and in the 2022 letter they indicated results are ahead of initial underwriting expectations.  My guess is that their Lincoln deal was probably a pretty good one given their long-time presence and deep connections there.

 

I don't know much about the location of these builds; could be in very underserved areas with little competition. In general, the "good' locations" for overbuilds get decent numbers initially, and incrementally returns get harder and harder after the initial uptake. This is especially the case with an incumbent. 

[Yosemite]

The returns are very long dated.  We are talking a decade and depend on the exit strategy.

 

1.    Mesa, Arizona has at least four (4) fiber overbuilders, some are open and some are not,  the returns are going to be poor with more than one overbuilder.   

 

2.    A Telco copper to fiber overbuilder (Frontier and Brightspeed an Apollo PE company) who do not offer up their fiber to third parties    Brightspeed  plan is  3mm homes over 3.5 years.   It's going to spend $3billion on this iniative

 

3.  A single open fiber overbuild (competing against the telco old copper and the cable companies hybrid network)   the margins are high  but the investment return timeline is extreme (10-13 years).     The 10-13 years is coming from Brookfield's projections.

 

As an example Brookfield (Intrepid) is building 78,000ft of aerial and 29,000 ft of buried in Pueblo, Colorado.  (about 20 miles)  passing 41,000 homes for a total cost of $4.6mm (with $2.7mm received in federal grants) .     T Mobile is the initial company but small ISPs might become customers later.       

[KJP]

In today's presentation, Charter projects $2,000 - $2,500/passing (doesn't include cost of the drop) for its greenfield footprint expansions.  See slide 11:  https://ir.charter.com/static-files/ba52d4ff-3a83-415b-8c1a-1f0a9a12df8b

 

On slide 12, they show post-subsidy cost/passing for the rural buildouts.  It's well over $3,000/passing.

 

On slide 13, they show margin profile of rural buildouts.  They claim 70% gross margin with some of the revenue being video and mobile.  So broadband only would appear to be above 80%. 

 

Based on the dot plot on slide 13, average penetration appears to be ~45% at 15 months.  I believe these subsidized rural areas have some of the weakest current broadband offerings in the country.  So, someone overbuilding an average existing cable broadband system with fiber likely would not achieve a similar penetration timetable.

Edited February 2 by KJP

[oscarazocar]

18 hours ago, Munger_Disciple said:

 

I don't know much about the location of these builds; could be in very underserved areas with little competition. In general, the "good' locations" for overbuilds get decent numbers initially, and incrementally returns get harder and harder after the initial uptake. This is especially the case with an incumbent. 

 

Returns on fiber/broadband investments hugely depend on competition or lack thereof.  As others have mentioned, if you are in an area with 3 or more providers, returns will be terrible.  If you are the only option or there is one other mediocre compeitor, returns can be very good.

 

Allo focuses mainly on small towns where they think they can have a high market share.  Click below and you and see the places they are.  It's mostly small towns in Nebraska like Kearny, Crete, and Gering.  This is not competition central.  They went into Lincoln because Spectrum (Charter) had a terrible local reputation and, as mentioned, I believe they got a good deal with the city who wanted competition in the broadband market.

 

https://www.allocommunications.com/communities-that-want-allo/

[dwy000]

After 14 years of stop/start/stop/start in the fiber business it looks like Google is throwing in the towel and seeking external funding for Google Fiber.

 

https://www.reuters.com/business/media-telecom/alphabet-is-seeking-outside-investment-its-gfiber-internet-business-2024-02-06/