Description

We just saw Lumn stock jump from the $1 per share range into the $3 range. I think we shall soon again see a $1 print on Lumen. So Lumen and MSFT made an agreement for Lumen to improve the backhaul connections into the MSFT data warehouses. The release was sold quite cleverly. Corning will sell their new AI fiber to Lumen which has just reserved 10% of this Corning AI fiber capacity in order to fulfill this MSFT need for more AI fiber.

So what is this famous AI fiber? Well corning says they found a way to pack in more fiber into a way smaller cable. Up to four times more fiber strands into the same space of a conduit. Why is this supposed to beneficial to Lumen? Well these backhaul guys tend to run their backhaul inside of conduit. Well, if you have one conduit, then this new AI fiber should allow Lumen to run up to four times the number of fiber strands into the same old conduit. So few requirements for the very expensive process of Lumen having to bury more conduits once it is running out of space in their old conduits. Awesome, more capacity within limited capacity conduits. So awesome right …

I guess from my tone you already understand I do not agree with this simple thesis for two reasons. As we have noticed, backhaul fiber has been a stagnating business ebitda for some time now. I remember the original Level 3 execs advertising in the late 90s “we build excess conduits so we have a lot of opportunity to grow” and their competitors had not … and then oops Gpon arrived which allowed for many light beams to travel through one fiber strand instead of just one. And it seems the industry just keeps Gpon Gbps going up and up and up and up.

Here is a quick overview of the evolution of fiber routers.

• 2004: GPON – down 2.5 Gbps, up 1.25 Gbps
• 2013: NG PON1 – down 10 Gbps, up 2.5 Gbps
• 2016: NG PON2 – down 40 Gbps, up 10 Gbps
• 2018: XGS PON – down 10 Gbps, up 10 Gbps
• 2022: 50 GPON – down 50 Gbps, up 50 Gbps
• Today: 100 GPON has been just tested – Yes that 100 Gbps down and 100 Gbps up. Nokia will bring this router on to the market in 2025.
• Also today: 200 GPON has already been demonstrated and coming soon after 100 GPON. – And yes that will be 200 Gbps down and 200 Gbps up.

 I guess people might notice that the technical standards of these new backhaul GPON routers is literally increasing at rates that are way faster than the actual growth rate of global data.

So lets now address the two issues I see with Lumen.

1. Yes, maybe Lumen got a sweethard deal with MSFT, but I kinda doubt it. And also, yes Lumen tied up 10% of the AI fiber capacity from corning. But what am I supposed to assume what Corning is supposed to do with the other 90% of its capacity, just park it in inventory and not sells it? Nope Corning is planning on selling the rest of their AI fiber. And I see 9 other sellers of fiber strand globally. Prysmian, Fujikura, Sumitomo, Sterlite, OPS, Yangtze Optical, Hentong, Commscope, ZTT. Yes with 25% market share globally Corning is indeed large. But it is not as if it can offer some form of exclusivity to Lumen for its deal with MSFT. And am I also supposed to assume that the competition will not be able to innovate itself?

So yes, this new AI fiber from Corning is an improvement, afterall a backhaul provider can indeed now take this new kind of fiber and when its current conduit capacity is maxed out on a certain traject, just pull out the old fiber and run the new AI fiber, thus avoiding the need to dig again, which is expensive. But this will be the same for all other competitors of Lumen. Lower cost will just drive down prices more, as it has done since the year 2000.

2. And now lets connect this AI fiber issue to the GPON progress. So with 50 GPON routers currently on the market, lets say you have X fiber strands between place A and B. With the current XGS PON 10 Gbps routers, you just buy 50 GPON routers and you just increased your capacity between A and B by a factor of five. And in 2025 when Nokia brings their 100 Gpon product to market you can just make the jump from XGP PON (10 Gbps) to 100 GPON, taking up your capacity by a factor of ten. Indeed demand for data is growing fast, it ain’t even close to growing at a growth rate of a 1000% in a year. The aggregate growth rate for data transmission globally isn’t even growing at a rate of 100% annually. And keep in mind, 200 GPON is right around the corner.

So I believe that backhaul data transmission rates are about to fall off a cliff. So we got our AI fiber that will allow backhaul fiber providers to increase the quantity of fiber strands in old conduits by four times. And then we got new routers come to market  that will be able to take the capacity from 10 Gbps routers to now 50 Gbps routers, next year will be able to take the capacity of a fiber stands to 100 Gbps and soon after the fiber backhaul companies will soon be able to go to 200 Gbps. And I wouldn’t be surprised we get even much faster Gbps speeds after 200 GPON has been rolled out. So in worst case scenario by 2027 when 200 GPON routers hit the market, together the Corning AI fiber we could see a capacity expansion of 800 times, and next year we will already see a times 400 capacity move per conduit. Yes data transmission growth is going to keep going up and up, but it ain’t going to come even close to the 400 times rate that is coming to market next year.

I am a huge fan of last mile fiber owners, but I think backhaul fiber is just a really shitty business. Yes, these capacity improvements indeed allow back haul companies to lower capex, but it is exactly this need for new conduit when capacity between city A and B was maxed out that kept a need for somewhat decent prices, otherwise no one would build more new conduit capacity. But especially since the fact that we now already have 50 GPON routers being sold and that next year 100 GPON routers from Nokia will hit the market, we could likely see such a massive and cheap way to increase capacity per fiber strand that demand for new fiber strand will collapse. And since fiber is a fixed cost asset, the bit of data your fiber did not transport is lost forever. So the incentive will be to have a race to the bottom.

Anyway, lets bring this back to Lumen:

• Lumen’s revenue from backhaul and backhaul related business is 78% of total revenues. We will see the currently already declining revenues on a quarterly basis even accelerate. Yes, Lumens backhaul might touch twice the bits, but Lumens pricing per bit will fall by more, a lot more.
• So Lumen is just reserved 10% of the corning AI fiber for the next two years and will according to its own release, replace a lot of its USA intercity conduit, hence doubling total capacity. So given Lumens liquidity profile, am I to assume MSFT is going to fund this upgrade? I guess we all know the answer to that question. Given Lumen is already struggling with the required levels of free cash flow it generates right now, are we supposed to assume that something will fund this capex?
• Also Lumen has 18 million DSL lines where it is already struggling with its overbuild speed of fiber. But I am supposed to assume that this MSFT will result in a higher ROIC long term?

Anyway, I think this Lumen stock price bump is just nothing more than another meme move by mostly retail. And soon retail will realize that there is literally zero involvement of AI with fiber. Yes AI data bits travel over fiber, just like any other bit of data. And then the stock will soon start trading back to the $1 level and in next few years the collapse in the Lumen backhaul business will force Lumen into chapter 11.

Risk: This now seems it is a meme stock. If they can sell AI fiber as a concept to retail, who knows what for other bs they get to sell to them. But between now and the end of 2026 Lumen will file for chapter 11 protection.

Anyway, this current period so feels like the .com bubble. Remember some burial comapany added .com to its name and the stock price shot up. Feels again a lot like that today.

Btw. the current borrow at IBKR is 0.25% and they got 73.5 million shares avaible to short. I think about 130 million shares are current sold short.

Catalyst

- GPON routers increasing their capacity way faster than actual growth of internet data needs. 

- Continuing needs for massive capex just to stop going backwards faster.

- Some new machination from Lumens PR department, probably selling more of this AI bs could indeed create a short squeeze. Now this situation is not remotely like GME was a few years ago, but still if a stampede is created of people trying to cover a short, it will be painfull in the short term for sure.