Subsea Cables & Internet Infrastructure

This project excites many powerful groups in Europe. Scientists want to equip the cable with sensors to study the Arctic Ocean while the EU wants to strengthen its influence on the Far North and also create a unique, low latency communication link with Asia that bypasses North America. It is in large part about infrastructure sovereignty. The benefits are quite clear. Indeed, they are at first glance compelling. Right now the EU has given a few million Euros to a consortium of carriers and educational networks to design it and perhaps conduct the geophysical survey. Unfortunately, the reality is more complicated. Indeed, the project has two Achilles Heels. A single cable is likely to be down a good deal of the time. That is the track record of Arctic cables: outages take in many cases 4 to 9 months to fix. So it is necessary to build a ring, which means two diverse subsea cables. So the total project cost doubles. But that is just the beginning of the challenge. Most cable ships cannot...

The Michael Burry view is that companies are depreciating their GPUs over 5 year periods despite technological obsolescence in 18 months or less. This means negative profits and rates of return despite corporate claims to the contrary.  But heat stress also depreciates GPUs. AI Chips run marathon sessions during model estimation (training) often exceeding a month and GPUs often fail due to heat stress: "Imagine you had a 10,000 or even a 20,000 GPU data center. You should expect on the statistics a chip to fail about every 3 or 4 hours. So long before I get to the point where I’m rapidly turning these over because there’s a new generation of chips, I’m turning over a vast chunk of my chips just because they’re failing under thermal stress." Source: https://paulkrugman.substack.com/p/talking-with-paul-kedrosky. GPUs used to estimate large language models are like ultra-marathon (50, 100, and 166 kilometers) runners on a 38 degree day.  It took Facebook 54 days to estimate the ...

 Dear Readers,  The International Telecom Union is hosting a meeting in Portugal in February to debate and discuss the resiliency of subsea cables. I think the debates will be lively. I encourage you to attend. I hope to be there myself. As I am not a member of any big international organization or company in the industry, it may be challenging to get attendance approval.  Regards,  Roderick. 

A point: Ashburn Equinix or 56 Marietta, Atlanta. Z point: Madrid Interxion 2. Term: 3 Years Bandwidth: 100G. Service: Layer 1. MRC: $5,990. NRC: $0.

The consortium has been coy regarding its plan to bypass the Red Sea. My initial guess was that SMW6's main trunk would land in Oman and head across the Saudi Arabian desert to Egypt. I know more now. The main trunk will land in Bahrain and Kuwait and then go overland to Jeddah. If you view the map, you can see there is a highway making an almost straight line from Bahrain via Riyadh to Jeddah. This approach makes sense because it uses an existing right of way and hence sharply reduces deployment costs. I don't know whether existing terrestrial SA fibre was used or new stuff blown through an empty conduit. Any Saudi terrestrial capacity will be very expensive although pricing may have been tempered by the desire to get the consortium to adopt the route as part of the main trunk. It will be interesting to see if Blue-Raman, Africa-1, 2Africa, and other stalled projects follow a similar path. Such an outcome would be upsetting for many Middle Eastern in the vicinity of the Red Se...

The LEO satellite business model is one of the most interesting topics in telecommunications today. LEO satellites are not a 'new technology'. It is really just a new approach to providing Internet access without any intellectual or patent moat to prevent the entry of new competitors. There could be a know how moat, but the number of LEO satellite networks entering the market over the next 5 year suggests it is very shallow. LEO satellites provide Last Mile access using low Earth orbits ranging from 300 kilometers (180 US miles) to 2,000 kilometers (1200 miles). Each bird as they are affectionately called in the industry takes only 90 to 120 minutes to complete an orbit. Their orbits can take a variety of shapes. For example, a LEO orbit could be elliptical in order to achieve a closer approach to the Earth at some points along the path for more detailed image or data collection. Or the LEO could fly longitudinally from North Pole to South Pole and back again. Hence LEO satelli...