Starlink, the proposal for SpaceX to service the internet through tens of thousands of satellites, is a staple in the space news, with storeys on the latest developments appearing every week. The large schema is clear and a reasonably well motivated person (such as your humble servant) can deduce a great deal of information thanks to filings with the FCC. Despite this, even among expert analysts, there is still an exceptionally high degree of uncertainty about this new technology. Reading papers comparing Starlink to OneWeb and Kuiper (among others) is not unusual, as though they were all comparable rivals. It is not unusual to read about well-meaning astronomy concerns about space junk, space law , regulation, and harm. It is my hope that the reader will be both better educated and more inspired by Starlink by the end of this very long article.

 

SpaceX is at the helm of competing launch providers and, at the same time, offers a mechanism for space restructuring. The subtext here is that the traditional satellite industry was unable to keep up with the steadily growing capability and reducing cost of SpaceX on the Falcon launch family, putting SpaceX in a challenging role. On the one side, it saturated a market worth a few billion a year, at most. And on the other hand, to create a massive rocket with almost no paying customers, and then to launch thousands of them to Mars with no immediate economic return, the insatiable appetite for cash was developing.

 

Starlink is the solution to those twin issues. Through creating their own satellites, SpaceX could build and define a new market for highly capable, democratised access to space communication, provide their own rocket with a revenue stream and payload, even as they cannibalise themselves, and ultimately unlock trillions of economic value. Do not underestimate the size of the ambition of Elon. Only a few trillion dollar industries exist: oil, high-speed transportation , communications, chemicals, IT, health , agriculture, government, defence. Space mining, lunar water, and space-based solar power are not feasible industries, despite common myths. Elon has a power play with Tesla, but communications alone offers a reliable, deep satellite and launch market.

 

 

The first space-related proposal for Elon Musk was to spend $80 m on a philanthropic mission to develop a plant on a lander on Mars. It would cost maybe 100,000 times as much to create a Mars city. Starlink is the key bet for Elon to produce the ocean of gold needed to create a self-sustaining city on Mars philanthropically.

 

In 2012, when SpaceX realised that its clients, mainly comsat providers, had better margins than they did, Starlink was conceptually born. In order to put satellites in space, launch providers charge notoriously insane prices, and yet somewhere there was a piece of the action they had missed? Elon dreamt of an internet constellation and got the ball rolling, unable to avoid a near-impossible technological obstacle. The creation phase of Starlink has had its difficulties, but at the end of this article, you, the reader, will probably be shocked at how few problems there have been, considering the magnitude of the underlying vision.

 

Why do we need to supply the Internet with an immense constellation of satellites? And Why now?

 

The internet has grown from scholarly curiosity to the single most disruptive piece of technology ever created in my lifetime. This is not the focus of an extended internet debate, but I would assume that global internet demand and the wealth it brings will continue to rise rapidly by about 25 percent a year.

 

But today, from a tiny handful of geographically-isolated monopolies, almost all of us get our internet. In the US, the nation has been carved up by AT&T, Time Warner, Comcast, and a handful of smaller players to discourage competition, charge exorbitant prices for poor service, and bask in near universal hate.

 

There is a compelling explanation for anti-competitive behaviour among internet service providers, in addition to overwhelming greed. The internet's underlying infrastructure, microwave cell towers and optical fibre, are incredibly costly to instal. Just how miraculous the internet's data-transfer properties are is easy to forget. My grandmother's first job during the Second World War as a Morse code operator was a medium that fought for preeminent strategic importance with homing pigeons! It is so disembodied, so incorporeal for most of us to travel the knowledge superhighway, that we neglect that those bits have to cross our physical universe with all its borders, rivers , mountains, seas, hurricanes, natural disasters, and other annoyances.

 

What we need is a way to raise the flow of data from the turbulent surface of Earth and into space, where in 5 years, a satellite would easily orbit the Earth 30,000 times. This seems obvious — so why didn't anybody try it before?

 

 In the early 1990s, the Iridium constellation invented and implemented by Motorola (remember them?) was the first global LEO-based communications network. Its niche ability to route tiny data packets from asset trackers turned out to be its main use by the time it was introduced, as cellular phones were inexpensive enough that the use of satellite phones never took off. In 6 separate orbital planes, the Iridium network had 66 satellites (plus a few spares)-the minimum amount required to fully cover the globe.

 

If 66 satellites are enough for Iridium, why are tens of thousands being designed by SpaceX? What makes SpaceX different?

 

SpaceX came from the other way, from launch, to this venture. As a result of their groundbreaking efforts in booster recovery, their cheap launch has absolutely cornered the market. There is not much money to be made to undercut their competition excessively, so being their own customer is the best way to leverage their excess ability. Since SpaceX is able to launch its satellites at about a tenth of the price (per kg) of the original Iridium constellation, they are capable of tackling a far more inclusive market.

 

The world-wide internet from Starlink would offer high-quality internet connectivity to every corner of the world. Internet availability would depend, for the first time, not on how close a specific country or city comes to a strategic fibre path, but on whether it can see the sky. Unlike their own variously inept and/or corrupt government telecom monopolies, entrepreneurs worldwide will have free access to the global internet. The monopoly-breaking potential of Starlink would catalyse tremendous positive change, taking billions of people into our future global cyber culture for the first time.

 

If I can deviate from a paragraph, what does that mean?

 

For people who are growing up today in an age of ubiquitous access, the internet is like the air we breathe. It's still there. But that's to forget its immense ability to bring about positive change – a change that we're in the midst of right now. The Internet is capable of helping people keep politicians accountable, connect with people elsewhere, exchange ideas, create new things, and unify the human race. The history of modernity is one of expanded human data sharing capabilities. First of all, by speeches and epic poetry. Then write, allowing the dead to communicate to the living, data storage, and asynchronous communication. The printing press, which allowed the mass production of knowledge. Electronic correspondence, which hastened the passage of data around the globe. From notebooks to mobile phones, our personal note taking devices have gradually become more sophisticated, each of which is an internet-connected machine festooned with sensors and increasingly able to anticipate our needs intelligently.

 

As a cognitive aid, a person capable of using writing and computers is substantially more capable of transcending the limitations of their poorly formed wetware. What's more exciting is that humans have both a good note-keeping system in mobile phones and a mechanism for exchanging thoughts with other people. Although people may have relied on speech historically to express the ideas they have written in their notebooks, the trend today is for notebooks to share the ideas created by humans, an inversion of the conventional schema. A type of collective metacognition is the logical extension of this process, mediated by personal devices ever-more closely integrated with our brains and each other.

 

While there is space in this world to be nostalgic for the diminution of our connection to nature and the loss of solitude, it is important to remember that technology, and technology alone is responsible for the vast majority of humanity’s emancipation from the “natural” cycles of ignorance, premature avoidable death, violence, malnutrition, and tooth decay.

 

How?

 

Let 's address the business case and design architecture of Starlink.

 

In order for Starlink to be a good company, its profits needs to surpass its construction and operating costs. Satellite firms historically have front-loaded expenses in capex, using advanced specialist funding and insurance mechanisms to launch a satellite. The design and launch of a geostationary communications satellite will cost $500 m and take 5 years, but the industry works similarly to the construction of a jet or container ship. With a relatively cheap operating budget, large outlays and barely enough sales to cover funding costs. The biggest drawback of the original constellation of Iridium, on the other hand, was that Motorola forced the operator to pay ruinous licencing fees, bankrupting the whole effort within months.

 

Traditional satellite companies have had to serve specialist clients and charge high prices for their data in order to operate these companies. Despite the poor latency and relatively low bandwidth of a satellite link, airlines, remote outposts, aircraft, war zones and vital infrastructure pay about $5 / MB, which is 5000 times higher than the cost for a conventional ADSL connexion.

 

Starlink aims to compete with commercial terrestrial ISPs, so it must be capable of delivering data for less than $1 / GB, preferably even less. Is it possible here? Or rather, we can wonder, if it is possible, how is this possible?

 

A cheap start is the first ingredient in the mix. Currently, Falcon offers launches for up to 24 T for about $60 m, which works out to be $2500 / kg. However, this is a lot more than their internal marginal expense. On several reflown boosters, Starlink satellites are to be deployed, so the marginal cost per launch is the cost of a new second stage (maybe $4 m), fairings ($1 m) and ground handling (~$1 m). This works out to be more than 1000 times cheaper than a traditional comsat launch, about $100k per satellite.

 

However, most of Starlink will be published on Starship. Indeed, as shown by updated FCC filings, the history of the Starlink constellation gives us some insight into how the internal architecture formed when Starship became real. The constellation 's total number of satellites rose from 1584 to 2825 to 7518 to 30,000. Or, if you think that the sum is accumulating, still more. For the first step of growth, the minimum viable number of satellites is 6 out of 60 planes (360), with 24 out of 60 planes (1440) required for complete coverage within 53 degrees of the equator. 24 Falcon launches could cost as little as $150 m internally. The Starship, on the other hand, is planned to launch up to 400 satellites at a time and at a comparable cost per launch. After five years, Starlink satellites are planned to be replaced, so 15 Starship launches a year are needed for 6000 satellites. This could cost as little as $100 million a year, or $15 thousand per satellite.

 

While each of the Falcon-launched satellites weighs 227 kg, while carrying third-party instruments, the Starship-launched satellites could weigh as much as 350 kg and be somewhat larger without exceeding the launcher 's capacity.

 

What's the expense of satellites? The Starlink Sats are somewhat peculiar as satellites go. They are incredibly simple to produce in mass, constructed, stacked and launched in a flat configuration (stackellites?). The manufacturing cost goal should, as a rule of thumb, be about the same as the unit launch cost. A large cost difference will be a sign of misallocated engineering capital, as the marginal cost enhancement to savings on the cheapest side is too minimal. Does $100k for the first few hundred sound fair per satellite? Is a Starlink satellite, in other words, of similar complexity to a car?

 

To answer this question fully, we need to understand why, despite not being a thousand times as complicated, a geostationary comsat could cost a thousand times as much. More generally, why does hardware for space cost so damn much? There are many explanations for this, but the most popular in this case is that if the geostationary launch (pre Falcon) costs more than $100 m, to make any money, the satellite needs to work reliably for many years. It is a painful, drawn out process that takes years and requires hundreds of people to ensure a high likelihood of operability on the first and only paper. Costs add up and if the launch is that costly anyway, it's easy to justify additional procedures.

 

On the other hand, by designing hundreds of satellites, iterating rapidly on early design defects, and using mass manufacturing methods to control costs, Starlink violates this model. Personally, I have no trouble imagining a Starlink manufacturing line where a technology can integrate and zip all together in an hour or two, maintaining the necessary replacement rate of 16 satellites per day. There are many fancy components in a Starlink satellite, but I see no reason why the overall production cost of the thousandth unit off the line does not fall to $20k. Indeed, Elon tweeted in May that the cost of satellite development was already less than the cost of launch.

 

Let's select an intermediate scenario, using round numbers, to analyse "time to revenue." For five years, a single Starlink satellite costing $100k to build and launch can work. Is it going to pay for itself, and how fast?

 

A Starlink satellite will complete 30,000 orbits in five years. The satellite would spend much of its time in each of these 90-minute orbits over the uninhabited ocean, and maybe just 100 seconds over a densely populated area. It can transfer data and receive revenue as quickly as it can during that brief window. Assuming 100 separate beams can be supported by the antenna, and each beam can transmit at 100 MB per second using advanced coding such as 4096QAM, the satellite generates revenue per orbit of $1000, assuming a $1 / GB subscriber cost. This is sufficient in just a week to earn the $100k deployment expense, significantly simplifying the structure of the money. The remaining 29,900 orbits, after fixed costs are accounted for, are benefit.

 

These assumptions will obviously differ a lot in either direction. But in any case, a significant business opportunity is provided by being able to deliver a competent communications constellation to LEO for $100k, or even $1 m, per unit. A Starlink satellite will produce 30 PB of data over its lifetime at an amortised cost of $0.003 / GB even taking into account its ludicrously low utilisation fraction, with virtually no marginal cost increase for transmission over a longer distance.

 

Let's quickly compare it to two other models for consumer data distribution to understand how convincing this model is: traditional optical fibre-based, and a satellite constellation provided by a company that doesn't specialise in launching.

 

The SEA-WE-ME 4 is a major submarine cable which was commissioned in 2005 and runs from France to Singapore. It can transmit 1.28Tb / s, and it costs around $500 m to deploy. If it runs for 10 years, equal to 100% capacity, with a 100% capital cost overhead, then the price per bit works out to be $0.02 / GB. Transatlantic cables are shorter and a little cheaper, but in a long line of people who need money to deliver data, the undersea cable is only one person. For Starlink, the middle road calculation is 8 times cheaper, all in, than just an undersea cable.

 

How will it be possible? All the complex electronic switching gear found linking optical fibres together is included in each Starlink satellite, but they simply use empty vacuum instead of a costly and fragile cable to carry data. Transmitting information across space dis-mediates all the cosy moribund monopolies and enables customers to communicate with far fewer hardware in between.

 

Let's contrast that with OneWeb, a rival constellation creator. OneWeb aims to launch a constellation of approximately 600 satellites with a variety of commercial suppliers for about $20,000 / kg. Each satellite would weigh 150 kg, suggesting a launch cost of about $3 m for the best case unit. The cost of satellite hardware is forecast at $1 m per, with a total cost of constellation production of $2.6b by 2027. A peak data rate of 50 MB / s, preferably per each of the 16 beams, was demonstrated by the OneWeb test satellites. We find that each OneWeb satellite produces $80 / orbit revenue and a total of $2.4 million in 5 years , following the same process as the Starlink estimate above, perhaps only covering launch costs when data services to remote areas are included. This averages out to be approximately $1.70 / GB.

 

Starlink is 17x cheaper or faster than OneWeb, which would mean a comparable cost of $0.10 / GB, Gwynne Shotwell was recently quoted as saying. This refers to the cost of data distribution in the initial configuration of Starlink, with less configured manufacturing, launch of Falcon, and minimal data offerings only in the northern USA. It turns out that SpaceX has an insurmountable advantage over the competition: a more capable satellite can be launched today for 15 times less per unit. In fact , it is possible to imagine SpaceX launching 30,000 satellites by 2027 for a total cost of less than $1b, most of which will be self-funded by that point. Starship will lift this to a factor of 100 or more.

 

For OneWeb and other optimistic constellation creators, I 'm sure there is an analysis that is less grim, but I'm not quite certain how it will be organised.

 

Recently, Morgan Stanley estimated that Starlink satellites would each cost $1 million to build and $830k to launch, a figure described by Gwynne Shotwell as' waaaaayyyy off.' I find it curious that the projected costs are close to our OneWeb guess, but approximately 10 times the original version of Starlink 's calculation. Starship and satellite mass production usage could reduce the cost of satellite deployment to about $35k per unit, an incredibly inexpensive number.

 

One final point is to compare Starlink 's revenue per watt of solar power generated. According to photographs on the website, each satellite's solar array is about 60 sqm, which means they generate an average of around 3kW, or 4.5kWh, over an entire orbit. With a ballpark estimate of $1000 in revenue per orbit, around $220 / kWh is produced by each satellite. This is 10,000 times the wholesale cost of electricity generated by solar, showing once again that a losing proposition is space-based solar power. It is a huge value-add to modulate microwaves with data!

 

Architecture

In the previous section, I glossed over a significantly non-trivial aspect of the Starlink architecture; the way in which it deals with the decidedly non-uniform population density of humans. Every Starlink satellite can produce concentrated beams that produce spots on the Earth's surface. Both subscribers within a single beam spot need to share the bandwidth. The size of this spot is determined by fundamental physics – its width is basically (satellite altitude x microwave wavelength / antenna diameter) which, for Starlink satellites, is planned to be at best a few kilometres long.

 

Many cities have a population density of about 1000 people per square kilometre, although some are much higher. More than 100,000 residents in parts of Tokyo or Manhattan may be within the footprint of a single beam. Fortunately, there is a dynamic broadband wholesale market in all such high density areas, not to mention highly developed mobile phone networks. Furthermore, when a highly designed constellation has several satellites overhead at any one time , the data rate can be improved by spatial separation as well as by frequency allocation. In other words, hundreds of satellites might aim their most effective beam to the same area, and subscribers in that area would use ground terminals that would break the demand between those satellites.

 

Although remote, rural and suburban areas will provide the most important market for the initial stages of the constellation, further launches will be financed entirely by offering better service to high density cities. This is the opposite of the more traditional business expansion scenario, where competitive city-based services ultimately experience diminishing margins as they continue to expand outwards into poorer, sparser regions.

 

After the data has been produced from the satellite subscriber, what happens to it? Satellites transmit this data directly back to dedicated ground stations near the service areas in the initial version of Starlink. "Bent pipe" is the name of this configuration. In the future, Starlink satellites will add the capacity to use lasers to communicate with each other. Satellites would be at full capacity over dense towns, but data will spread over two dimensions via the laser network. In practise, this implies that the satellite network has immense latent backhaul power, so subscriber data can be "regrounded" anywhere it fits. I expect, in practise, that SpaceX ground stations would be collocated outside metropolitan centres with carrier hotels.

 

In situations where the satellites are not asking, it turns out that satellite to satellite communication is non-trivial. The most recent FCC filing provides 11 satellite families with different orbits. Surrounding satellites are travelling at the same altitude, inclination, and eccentricity within a given family, meaning that the laser links can reasonably easily track the satellites immediately surrounding them. But the closing speeds between families are measured in km / s, so interfamily contact, if any, must be conducted with short-lived fast-steered microwave connexions.

 

The topology of these families of orbits is super esoteric and not really important to the business case, but I find them beautiful, so I'm going to include them. Skip to "Fundamental physical limits" below if you're not that excited about this segment.

 

The mathematical entity identified by two radii is a torus, or donut. Drawing circles on the surface of a torus, either parallel or perpendicular to the form, is very trivial. It which interest you to discover that on the surface of a torus, both passing through the central hole and around the outside of the torus, there are two other families of circles that can be drawn. This pattern is known as the circles of Villarceaux, and I used it when I designed the toroid for the Coup de Foudre 2015 Burning Man tesla coil.

 

While orbits are usually ellipses, not circles, in the case of Starlink, a comparable construction applies. In a sequence of orbital planes, all at common inclination, a group of 4500 satellites form a continuously moving sheet over the Earth's surface. At a given latitude, the northward moving sheet turns around and heads south once more. The orbits would be very slightly eccentric in order to prevent collisions, so that the northward moving sheet will be a few kilometres higher (or lower) than the southern moving sheet. Together, as illustrated in this exaggerated illustration, the two sheets form a blown out torus formation.

 

 

Recall the contact is between neighbouring aspiring satellites inside this torus. In general, as their closing speeds are too high to reliably detect with a laser, there are no long-lived contact ties directly between satellites in the northward moving and southward moving sheets. Then, the route for transmitting data between sheets goes across the top or bottom of the torus.

 

In particular, 30,000 satellites with a large gap across the orbit of the International Space Station will be located in 11 nesting tori! This diagram illustrates, without exaggerated eccentricity, how all these sheets stack up.

 

 

 

Finally, the optimum flight altitude is worth considering. There is a trade-off between low altitude, allowing greater beam sizes and higher data speeds, and high altitude, allowing the entire World to be reached by fewer satellites. Over time, the FCC filings of SpaceX have trended towards lower altitudes, especially as the growth of Starship allows the rapid deployment of larger constellations possible.

 

Low altitude has other benefits, including a substantially decreased risk of collision with debris or detrimental effects of hardware failure. The lowest flying Starlink satellites (330 km) will burn up in a matter of weeks following loss of orientation power because of increased atmospheric drag. Indeed, 330 km is smaller than almost all other satellites, and the use of the built-in Krypton electric thruster as well as a compact configuration would be sufficient to sustain altitude. In theory, a sufficiently pointy electrically propelled satellite might stably orbit as low as 160 km, but when there are a few other data rate-increasing tricks to try, I don't expect SpaceX to travel that low.

 

Fundamental spatial borders

Although it seems unlikely that satellite deployment costs will ever drop well below $35k, the ultimate physical limitations on satellite efficiency are less obvious, even with sophisticated digital production and entirely reusable Starships. Centered on the round numbers of 100 beams each capable of transmitting 100MB / s, the above study assumed a peak data bandwidth of about 80Gbps.

 

The Shannon-Hartley theorem defines the ultimate limit on channel size, which is given by log(1+SNR) bandwidth times. Bandwidth is also restricted by the available bandwidth, while because of imperfect antennas, SNR is limited by the available power on the satellite, background noise, and channel cross chat. Processing rate is another salient restriction. The new Xilinx Ultrascale+ FPGAs have up to 58Gbps GTM serial bandwidth, which is a reasonable approximation without creating custom ASICs for current channel network performance constraints. Even then, 58Gbps, most likely in the Ka or V bands, would require a significant frequency allocation. There are more cycles available in the V band (40-75GHz), but there is more atmospheric absorption, especially in areas with high humidity.

 

Will 100 beams be practical? This topic has two aspects: beam diameter and density of phased array antenna components. The wavelength divided by the diameter of the antenna defines the beam distance. A digital phased array antenna is also a specialised technology, but the width of a reflow oven (around 1 m) is determined by maximal practical sizes as it calls for difficulty using RF interconnects. The wavelengths of the Ka band are about a centimetre, indicating a beam diameter of 0.01 rads at half the actual full width. About 2500 different beams will work in this area assuming a useful beam solid angle of 1 steradian (similar to the field of view of a 50 mm camera lens). Linearity means that 2500 beams inside the array will need at least 2500 antenna components, which is feasible, but not easy. When used, it can even get very warm!

 

All up, a staggering amount of storage, about 145 Tbps, is 2500 channels each supporting 58 Gbps. Total internet traffic is expected to average 640 Tbps in 2020, for comparison. For individuals concerned about inherently poor satellite bandwidth, this is good news. Global internet traffic may exceed 800 Tbps if the 30,000 satellite constellation is operational by 2026. If 50 percent of this is covered at any point by ~500 satellites over heavily populated cities, so each satellite will have a peak data rate of roughly 800 Gbps, 10 times more than our initial simple calculation, theoretically generating 10 times the revenue.

 

A 0.01 radial beam occupies an area of 10 sqkm for a satellite in a 330 km orbit. In this zone, extraordinarily heavily populated cities such as Manhattan may have 300,000 inhabitants. The overall data demand is 2000 Gbps if they are all watching Netflix (7 Mbps in HD), approximately 35 times the actual hard cap set by FPGA serial outputs. There are two avenues, one of which is physically feasible, around this.

 

The first is to launch such satellites that at any moment there are more than 35 in the atmosphere over the high demand fields. This means a satellite density of 0.0002 / sqkm, or 100,000 if uniformly spaced across the globe, by using 1 steradian again for a rational addressable sky patch and 400 km for an average orbital altitude. Note that the selected orbits of SpaceX raise density over the heavily populated 20-40 degrees of latitude and 30,000 satellites begin to look like a magic figure.

 

The second suggestion is much cooler, but not probable, unfortunately. Recall that the diameter of the beam is determined by the width of the antenna array. Perhaps if, much as radio telescopes like the Very Wide Array, multiple antenna arrays on separate satellites merged their forces to synthesise a narrower beam? This solution is fraught with difficulties since it would be important to precisely track baselines between satellites to sub-millimeter accuracy in order to maintain the beam process as the satellites travel about. And even if this were feasible, owing to a lack of satellite density in the sky, the resultant beam would have incredibly poorly confined side lobes. The satellite beam diameter could be limited to a few millimetres in size on the ground (it could detect a mobile phone antenna), but millions of them would be because of weak intermediate nulling. A representation of the Thinned-Array Curse is this loss.

 

It points out that the division of channels by angular separation as the satellites are scattered across the sky produces sufficient data rate increases without breaching the laws of physics.

 

Using cases

 

What is the Starlink consumer profile? Hundreds of millions of residential subscribers with a pizza box-sized antenna on their roof are the usual case for usage, but there are major possibilities for other sales sources.

 

Land stations won't need phased array antennas in remote and rural areas to optimise bandwidth, so smaller consumer terminals are feasible. This range from IoT asset trackers to satellite pocket-sized tablets, emergency beacons, or monitoring devices for scientific animals.

 

Starlink can provide primary and back-up backhaul for cellular networks in heavily populated urban environments. Every cell tower can have a top-mounted high-performance ground station, but it can use ground-supplied electricity for the last mile for amplification and transmission.

 

Finally, the extremely low latency provided by VLEO satellites is used even in congested areas during the initial roll-out. In order to get critical knowledge just a little faster from any corner of the world, finance institutions are ready and able to pay top dollar. The vacuum speed of light is around 50 percent higher than in glass, but the journey taken by Starlink data is slower because of the hop into space, more than compensating for the disparity over longer distances.

 

Impacts

This blog's final segment deals with impacts. Although this blog is meant to discuss misconceptions around the continuum, some of the more controversial are the future impacts. It is my goal here to include data without too much editorialization. I don't have a crystal ball, nor does SpaceX have any inside details.

 

For me, the most important consequence is improved access to the internet. Even my native Pasadena, a thriving, multi-million-strong technological city that is home to many observatories, a world-class university, and the largest NASA hub, has very small internet supply options. Internet companies have been rent-seeking services around the US and the rest of the world, aiming to skim off their $50 a month in a comfortable, uncompetitive environment. Any asset that comes out of the housing wall is arguably a utility, but the standard of internet access is less standardised than water or power or coal.

 

The trouble with the status quo is that the internet is still young and rising increasingly, unlike water , electricity, and gas. The ways we use it are also changing us. The internet's most transformative use has not yet been invented yet. Yet the chance of competition and creativity is choked by "bundled" internet plans. For no other cause than a mistake of birth, trillions of people are left behind by the internet transition, or their country is a long way from a major undersea cable. The Internet is already provided by geostationary satellites at prohibitive prices in vast areas of the world.

 

Starlink, spraying parts from the sky constantly, totally disrupts this model. I don't know a great way to place billions of unconnected people online. SpaceX is on the road to become a provider of internet content and, perhaps, rivalling Google and Facebook as an internet business. I guarantee that you didn't see it coming.

 

It's not clear that satellites on the internet are the way to go. SpaceX, and only SpaceX, is in a position to create a vast internet constellation quickly, and only SpaceX has the ambition of wasting a decade trying to crack the government-military hegemony of launching space. And if Iridium had beaten mobile phones to market by a decade, using traditional launch services, it may never have reached large acceptance. Without SpaceX and its unique business model, there is a fair possibility that there will actually never be a global Internet satellite.

 

Astronomy is the second main influence. There was an outpouring of criticism from the international astronomy community that a massively expanded number of satellites would spoil their access to the night sky following the launch of the first 60 Starlink satellites. There is a saying that the astronomer 's output is directly proportional to his telescope 's size. Performing astronomy in the modern era without exaggeration is a super difficult task, a relentless arms race between improving research and increasingly increasing light pollution and other sources of noise.

 

Thousands of brilliant satellites shot through the picture are the last thing any astronomer needs. In fact, owing to large, mirrored panels that could reflect the sun's light over small patches of the Earth, the first Iridium constellation was somewhat infamous for its output of "flares." They might sometimes be as bright as a quarter moon, and sometimes even harm sensitive astronomical sensors. There are still legitimate questions regarding the incursion by Starlink into radio bands used for radio astronomy.

 

It is easy to see hundreds of satellites flying overhead on a clear evening if you download a satellite tracker app. After sunset or before sunrise, satellites are visible, but illuminating the satellite while the sun's rays already pass overhead. Later during the night, any overhead satellites are within the shadow of the Earth and are virtually unseen. They're small, incredibly far away, and they fly very easily. It's not unlikely for them to cover a distant star for less than a millisecond, but I guess it would be a huge hassle to even notice it.

 

Many of the outcry over Starlink light emission arose because the plane of the first launch was closely connected to the terminator of the Planet, ensuring that Europe, which was in autumn, had a clear view of the satellites flying overhead during the evening sunset. Furthermore, calculations based on early FCC filings revealed that even after the end of celestial dusk, satellites circling at 1150 km would be observable. There are three twilight phases; civil, nautical, and celestial, happening when the sun is 6, 12, and 18 degrees below the horizon, respectively. The sun's rays are about 320 km above the horizon at the zenith, just above the atmosphere, at the close of celestial twilight. Based on the Starlink website, I assume that all satellites are going to be deployed below 600 km. In this scenario, at dusk, but not long after nightfall, satellites may be visible, significantly minimising the possible effect on astronomy.

 

Orbital debris is the third preoccupation. In a previous post, I found out that, owing to atmospheric friction, satellites and debris below 600 km would deorbit within a few years , significantly decreasing the risk of Kessler syndrome. For trying to launch thousands of satellites, I think SpaceX gets a lot of hatred, as if their creators never thought of debris. It is impossible for me to envision a simpler way of doing it for debris mitigation as I look at the specifics of the Starlink implementation.

 

At 350 km, satellites are launched, and flown using an onboard thruster to their target orbit. Instead of lurking at a higher altitude for thousands of years, any satellite which dies on launch will de-orbit within weeks. Free entry assessment requires this implementation technique. In addition, if they lose attitude control, the Starlink satellites are flat in the cross section, ensuring a significant rise in orbital drag.

 

It is not well known yet in space launch, SpaceX pioneered the use of alternate fixtures for frangibolts. In order to launch stages, spacecraft, fairings, etc., nearly every other rocket supplier uses explosive bolts, raising the risk for orbital debris. Instead of making them float around indefinitely, SpaceX often purposefully de-orbits the upper stages, eventually decaying and disintegrating in the harsh space setting.

 

The last issue I 'm trying to record here is the possibility for SpaceX to substitute another monopoly for the current internet monopoly. SpaceX will already have a complete launch monopoly in a dynamic marketplace. Just the demand of competing policymakers for guaranteed military access to space keeps their costly and outdated rockets in service, mostly constructed by massive monopoly defence contractors.

 

It can definitely be imagined that SpaceX will launch 6000 of its own satellites annually by 2030, plus a few spy satellites for the sake of ancient times. The inexpensive and stable satellites of SpaceX will sell "rack space" for instruments constructed by third parties. Without needing to pay the cost of constructing the whole satellite platform, any university that can build a space-rated camera will bring it into orbit. And Starlink is becoming synonymous with satellites as historical producers disappear into obsolete obscurity as a result of this pervasive enhanced access to space.

 

For far-seeing firms so overwhelming a modern industry, there is a precedent that their commodity becomes synonymous with the notion. Hoover.-Hoover. From Westinghouse. From Kleenex. From Google. Frisbee. Xerox. Kodak.  From Motorola.

 

Where this may become troublesome is where the first competitor to market is engaging in anti-competitive tactics to protect its market share, but this is always legal after President Reagan. By pressuring other constellation developers to fire ancient Soviet rockets, SpaceX could maintain its Starlink monopoly. The United Aviation and Transportation Company's related behaviour, as well as price fixing on postal routes, contributed to its break-up in 1934. Fortunately, SpaceX is unable to hold an total monopoly on entirely reusable missiles indefinitely.

 

More worryingly, the deployment of tens of thousands of LEO satellites by SpaceX could be construed as a common co-option. For private profit, those previously public, unowned orbital slots will be permanently occupied by a private company. While it is true that the innovations of SpaceX provided the mechanism for commercialising the previously unremarkable vacuum, much of the intellectual capital of SpaceX was built on billions of dollars of publicly funded research.

 

On the one hand, to protect the proceeds of private investment, research and development, we need legislation. Innovators will be unable to finance ambitious projects without this protection, or will move their businesses to places that provide that protection. The public suffers from the loss of the production of wealth in either case. On the other hand, we need laws to protect people from rent-seeking private entities that would annex public wealth, the nominal owners of common property, including the sky. Neither hand, on its own, is correct, or even possible. Starlink 's development gives us all the chance to find our way to this new market's safe middle ground. When we've maximised the rate of innovation and common wealth creation, we will know we've found it.

 

Final Thought

 

Launch has been around for a long time, but without Starlink, Starship, a launcher cheap enough to be interesting, is not feasible.

 

For a long time , human spaceflight has been around, and if you're a fighter jet pilot who is a brain surgeon as well, you can go. Human space exploration has a credible, near-term path from the orbital outpost to fully industrialised deep-space cities with Starship and Starlink.