Good morning, investors, partners, and friends. We're thrilled to bring you the second monthly edition of Mach33’s SpaceX Investor Intelligence report.  The response to our inaugural issue has been phenomenal, with insightful feedback and suggestions pouring in. Your enthusiasm and input are invaluable as we continue to build and refine this product.

As we roll into our third month, we're actively experimenting with content sections, analysis methodologies, and more. This newsletter is still in its developmental stages, and your continued feedback is crucial in shaping it into an indispensable resource for SpaceX investors. So, please do write in and share your thoughts on this edition – your input will directly influence the evolution of this newsletter.

Quick reminder: The proprietary insights contained herein are critical to our investment thesis at Mach33 and should not be shared or forwarded without express permission. 

With that out of the way, let’s dive in – happy reading, all!

In this July release, we’re exploring… 
🛰️ Starlink vs. Fiber 👷🏽‍♂️
🛠️ F9 Grounding
🤝 SpaceX Supply Chain News
and more…

KPI Dashboard

The latest SpaceX metrics for your eyes only….

Fiber vs Satellite

In a world where your toaster wants to video chat and frontier language models’ training runs involve every word produced by humanity, one thing is certain: mankind and its machines have created an insatiable data beast. And adequately feeding this beast will require more and more data transmission capacity: 

• Over the past two decades, we've seen a 100X increase in optical fiber transport capacity (good for a whopping 40% CAGR).

• Projecting this forward, optical transmission capacity would need to scale to over 40 petabits per second per fiber in the next 25 years. 

• This is a significant scaling constraint, akin to Moore’s Law for silicon integrated circuits.

In today’s Main Event, we’re focused on how our demand for data may outstrip our capacity to provide it, and whether the stars (aka, physics and economics) may be aligning for Starlink to save the day. 

The Cost Conundrum: Dirt vs. Sky

Let's break down the economics of getting 1Gbps to your doorstep, comparing traditional fiber optics to a space-based approach. First, we’ll run through high-level cost estimates for delivering 1Gbps bandwidth to customers in an urban 1km² area.

Cost of New Broadband Infrastructure: The cost of laying fiber optic cable ranges widely from $20,000 (urban) per km to $100,000 (rural) per km, heavily influenced by terrain and whether you’re building above- or below-ground (if you dig, unsurprisingly, you’re paying a premium). 

• For today’s purposes, we’ll assume $20,000 per urban km² and $100,000 per rural/remote km². 

• Urban networks typically support 1 Gbps symmetric speeds per household with current technology. 

• Modestly approximating 10 streets per km and a grid city plan, there will need to be at least 10km of optic fiber laid underground per 1km² in order to serve every home. 

• This translates to $200,000 per Gbps per urban km² versus up to $100,000 per  rural/remote km².

Without diverting our analysis, we’ll just flag the high end of that estimate. For reasons of cost alone, it would be uneconomic to extend modern networking technology to vast swathes of the globe (meaning it’s unlikely to ever happen with current terrestrial methods).

Cost of Starlink: Each Starlink V2 mini satellite costs roughly $500,000 and weighs around 800kg. Falcon 9 launch costs are currently roughly $1000/kg. So, it costs ~$1.3M to build and launch a Starlink satellite. We are approximating that each Starlink satellite covers ~400 km² and provides 100 Mbps (0.1 Gbps) bandwidth to a home terminal. The implied cost per Gbps per km² is $30,000. 

The Upgrade Game: Advantage Fiber, For Now

Upgrading existing fiber infrastructure is where traditional broadband currently shines. Network upgrades can provide massive bandwidth boosts at a fraction of the initial installation price, due to the cost efficiencies of swapping out end-point electronics. 

Allow us to explain…contrary to what one may expect, upgrading optical fiber isn’t as laborious and expensive as it sounds. To upgrade network infrastructure, operators primarily enhance ‘active’ components of the network (network endpoints, like electronics and terminals). ‘Passive’ components, such as the fibers themselves, generally stay untouched until they are physically damaged or retired from service (this typically occurs on a two-to three-decade cycle). 

The upgrade du jour: GPON (Gigabit Passive Optical Network) to XGS-PON (10 Gigabit Symmetric Passive Optical Network), which boosts downstream bandwidth from 1.5Gbps to 10Gbps. This only requires upgrading electronics at the ends of the network, without touching the fibers themselves. Active components make up 20-30% of the overall cost of a fiber-to-the-home (FTTH) network deployment​. Coincidentally, the cost of the XGS-PON components are also 20-30% higher than GPON components.

• Thus, we can take the 25% cost midpoints to approximate the cost of an XGS-PON upgrade. 

• This implies a ~$60,000 upgrade to a 1km²  urban area and delivers an 8x increase in bandwidth. 

• Therefore, the cost to increase capacity by one additional Gbps per km² comes out to $7,500. This is significantly cheaper than the current Starlink implied cost ($30,000).

At this point, you might be thinking, "Game over, Starlink." But not so fast…

Approaching the Shannon Limit

Nothing good lasts forever, and step-change improvements to fiber infrastructure are no exception. 

Remember Moore's Law for semiconductors? Fiber has its own scaling law, but we're rapidly approaching the end of the halcyon days. Fiber scaling faces a physical constraint known as the Shannon Limit, which defines the maximum data transmission capacity of a communication channel, constrained by noise and bandwidth. 

As we near this physical barrier, future upgrades will require more drastic measures, such as adding more fiber pairs, using multi-core fibers, or completely ripping up, replacing, and rebuilding infrastructure. 

• This challenge is especially salient with subsea cables, which cannot be increased in diameter due to constraints on the maritime equipment, systems, and ships that lay cables on the ocean floor. 

• To give you a sense of this squeeze, most subsea cables have 12 fiber pairs while terrestrial cables are now deployed with 3,400 fibers. 

• As more space isn’t an immediate option, next-gen subsea cables require further fiber miniaturization or the utilization of more complicated technological processes to increase the number of cores per fiber. 

This is a major bottleneck, and is the key point to grok: undersea cables are the backbone of global data transmission and will become a bandwidth chokepoint. Once we truly cannot improve spectral efficiency any longer, we will have to build new subsea infrastructure to increase space, which will be extremely expensive. Spanning thousands of kilometers, these cables must be built into the ocean – and buried near shores. There are an estimated 1.4M km of undersea cables worldwide. Doubling spatial capacity by laying new lines would cost ~ $140Bn (at $100,000 per km fiber). Although a crude simplification, this is the same as launching ~100,000 Starlink V2 Mini’s, which would hypothetically increase constellation bandwidth 25x and allow for 2.5Gbps bandwidth to current Starlink terminal customers. So, a 2x increase in undersea cable diameter/bandwidth costs the same as a 25x increase in Starlink bandwidth. 

The outrageous costs of building new broadband infrastructure presents a potent, lucrative attack vector for Starlink to grow global market share. 

According to the high-level calculation presented herein, in rural areas or unserved regions, Starlink is already ~3x cheaper than building broadband infrastructure. This is undoubtedly going to lead to huge market uptake in underdeveloped countries and rural areas worldwide (we’re in the early innings of this today). In the near term, though, it will be a tale of two cities. Fiber is king in metros and will continue to dominate in dense urban areas, at least in the near term. Starlink (or competing space-based services, or heavily subsidized terrestrial technologies) will rule in rural and developing regions.

Although Starlink bandwidth would still be cheaper to implement from scratch on a per Gbps per km² basis, current Starlink bandwidth has not scaled enough to support the full demand of a densely populated city. This is evident in the current typical urban bandwidth per user of broadband and Starlink: 1Gbps vs 0.1Gbps, respectively. Although 0.1Gbps is still more than enough to stream 4k video (25mb/s), broadband is still an order of magnitude ahead in bandwidth. 

Our Takeaway

Broadband will undoubtedly remain the data backbone of the most densely populated areas. However, as data demands grow and broadband reaches physical limitations, the cost of upgrading terrestrial networks will become more expensive than upgrading Starlink, even in urban areas. 

The most near term chokepoint identified is undersea cables. With a 40% annual increase in bandwidth demand, there will be increasing pressure to reduce strain on these cables as we approach the Shannon Limit.  We anticipate this will force more international data traffic into space, driving growth for Starlink. Specifically the lower bandwidth traffic will be best suited. In particular, text, voice and mobile data all use the least, ranging from 0.001 Mbps for text to 25 Mbps for data. Satellite internet connection  is known to be problematic in cities due to line of sight issues as a result of building cover. This is much less of a problem for lower bandwidth services. For these reasons we foresee Starlink D2C (direct-to-cell), once rolled out, being far more successful in urban areas than Starlink terminals will be.

Note: the way Starlink bandwidth is delivered to urban areas is still in question. We may see Starlink partner with terrestrial telecom companies to provide augmented services or perhaps purchase/lease cell towers - an analysis we will leave for a future edition.

News Roundup

Takes 

What we are reading, watching, or hearing – and how we think about it.

Eye Candy