Elon Musk wrote on X on Friday that Starlink satellites carrying “~100 lasers per satellite, each operating above 10 Tb/s, with over 100k satellites should be able to outcompete undersea cables”. His post, at 19:32 UTC, quoted a clip of his remarks at the Baron Investment Conference earlier in the day.
In that clip, posted by Sawyer Merritt, Musk said “our current laser links are around a hundred gigabits a second our next generation laser links will be 10 terabits a second a hundred times better”, according to our machine transcript, and then said “we’ll put 10 of them on”. The post says about 100 lasers per satellite; the clip says 10.
Business Pill · Vertical Integration
A short explainer on vertical integration: when a company makes the inputs it uses. It teaches the general idea only and says nothing about any company in this story.
The key insight:Musk said next-generation Starlink lasers would run at 10 terabits per second, up from around 100 gigabits today, and wrote that about 100 per satellite across more than 100,000 satellites “should be able to outcompete undersea cables”; on his post’s figures that is 100 exabits per second of summed terminal capacity, by our arithmetic.
In the clip, Musk said he thinks “Starlink version four will have 10 times the capability of version three especially with the long-distance laser links”. In the post he added: “Speed of light in space is ~40% faster than cable and the path is direct, instead of following the outline of land masses”.
He gave no launch date for version four satellites in either the clip or the post, and SpaceX has not published specifications for them. The figures are Musk’s targets.
Signal speed in a vacuum vs in a typical telecom fibre core (refractive index around 1.4475, per RP Photonics as cited by Wikipedia), thousand km per second; our arithmetic. Musk said speed in space is about 40% faster than cable. Route length and ground hops are not shown.
What Was Posted on X
~100 lasers per satellite, each operating above 10 Tb/s, with over 100k satellites should be able to outcompete undersea cables.
Speed of light in space is ~40% faster than cable and the path is direct, instead of following the outline of land masses.
Elon Musk’s post at 19:32 UTC on 6 November, which says about 100 lasers per satellite at above 10 Tb/s across more than 100,000 satellites should be able to outcompete undersea cables. It quotes Sawyer Merritt’s clip of his conference remarks.
Elon just teased SpaceX’s next-gen V4 @Starlink satellites:
“Starlink version 4 will have 10x the capability of V3 (which itself is 10x better than V2), especially with the long-distance laser links. Our next-gen laser links will be 10-terabits per second. 100x better. And we’ll put 10 of them on.”
Sawyer Merritt’s post at 19:10 UTC with a clip of Musk at the Baron Investment Conference saying next-generation laser links will be 10 terabits a second and that SpaceX will put 10 of them on. The piece quotes the clip from our own machine transcript.
The Arithmetic on His Figures
On the numbers in Musk’s post, 100 lasers at 10 terabits per second on each of 100,000 satellites add up to 100 exabits per second of summed laser-terminal capacity, by our arithmetic. With the 10 lasers per satellite in the clip, the total would be 10 exabits per second. Per satellite, that is 1 petabit per second on the post’s figures or 100 terabits per second on the clip’s.
Those totals are not like-for-like with cable capacity. They count each laser terminal; a link between two satellites uses a terminal at each end, and the traffic a network can carry end to end depends on routing and on ground links, not on summed terminal speeds. TeleGeography’s 2026 Submarine Cable Map depicts 694 cable systems and 1,893 landing stations and tracks lit and potential capacity by route; the page we read gives no single global total to compare against.
The Speed-of-Light Point
Light is slower in glass than in a vacuum. A typical telecom fibre core has a refractive index of around 1.4475, according to the RP Photonics Encyclopedia of Laser Physics and Technology as cited on Wikipedia, which means a signal in the core travels at roughly two-thirds of its speed in a vacuum. By our arithmetic, light in a vacuum is about 45% faster than in that core; Musk said about 40%.
A laser link between satellites therefore has a speed advantage per kilometre. Whether a route is shorter overall depends on the satellite path and the hops to and from the ground, which Musk’s post describes as “direct” but does not quantify.
As we read the clip and the post together, the per-link figure is consistent at 10 terabits per second, while the number of lasers per satellite is 10 in the clip and about 100 in the post.
As we read the arithmetic, summed terminal capacity measures how fast lasers can send, not how much traffic a network carries end to end, so it is not a like-for-like comparison with cable capacity.
As we read the physics, a fibre core index of around 1.4475 puts light in a vacuum about 45% faster, close to the 40% Musk cited.
“Speed of light in space is ~40% faster than cable and the path is direct, instead of following the outline of land masses.”
Three Implications
FOR THE LASER LINKS Musk said current links run at around 100 gigabits per second and next-generation links at 10 terabits, which he called a hundred times better.
FOR THE CONSTELLATION His post assumes more than 100,000 satellites; on its figures, summed terminal capacity is 100 exabits per second, by our arithmetic.
FOR SUBSEA CABLES TeleGeography’s 2026 map depicts 694 cable systems and tracks lit and potential capacity by route; the page we read gives no single global total.
The framework argues: “The digital economy rests on a physical foundation that most people in that economy never see.”
As we read it, subsea cables are part of that foundation, and Musk’s post proposes satellites as an alternative route for some of the traffic. The framework does not discuss Starlink.
What Is Not Established
Version four satellites, 10-terabit lasers and a constellation of more than 100,000 satellites are Musk’s plans and targets, not deployed systems. The clip quotations are from our machine transcript of the 31-second video, not an official transcript. The number of lasers per satellite differs between the clip and the post. The 100-exabit and 10-exabit totals are our arithmetic on his figures and measure terminal capacity, not usable network throughput.
Musk said next-generation Starlink lasers would run at 10 terabits per second, up from around 100 gigabits, and wrote that about 100 of them per satellite across more than 100,000 satellites “should be able to outcompete undersea cables”. On his post’s figures that is 100 exabits per second of summed terminal capacity by our arithmetic, a different measure from cable throughput.
A note on sourcing. Musk’s words are from his X post of 6 November 2026 and from our own machine transcript (faster-whisper) of the 31-second conference clip in Sawyer Merritt’s post, not an official transcript. The refractive index and speed of light are from the RP Photonics Encyclopedia of Laser Physics and Technology as cited by Wikipedia; cable-system counts are from TeleGeography’s 2026 Submarine Cable Map page. Totals, per-satellite figures and the speed comparison are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
