Tesla's "Final Boss" You've Never Heard Of / Starbase, LA by the Numbers ⚡️

While daily market noise often distracts from long-term ambitions, SpaceX's Starship and Starlink projects represent an unprecedented technological leap poised to redefine global connectivity. The newly announced Starbase Louisiana facility aims to support over 30 daily Starship launches, leveraging its remote location to manage the vehicle's extreme noise, while the shift of all future Florida Starlink missions to Starship signals a massive bandwidth inflection. Specifically, a single Starship V3 launch delivers 60 Tbps of capacity—roughly 7% of the entire current constellation and 20 times that of a Falcon 9—which will be crucial for supporting the millions of Cyber Cabs requiring built-in Starlink connectivity. However, this explosive growth faces a critical regulatory bottleneck, as SpaceX is rapidly approaching its 15,000 Gen 2 satellite cap with only about 7,100 slots remaining. Consequently, the approval of their Gen 3 proposal under new, faster Part 100 licensing rules is essential to unlock the next phase of Starlink's expansion and sustain this ambitious launch cadence.

• ▶ 0:05 The bet tracker dashboards have been migrated to Cursor, with the Robotaxi Tracker showing 270 Tesla vehicles registered in Texas (89 added in August alone) and the Tesla Semi Tracker now tracking 1,267 verified orders, including delivery updates from Einride.
• ▶ 1:12 Dylan urges viewers to zoom out from daily news cycles and stock price noise, emphasizing the unprecedented ambition of projects like Terafab and Starbase, whose long-term impact will far outlast the current moment.
• ▶ 2:26 Dylan introduces sponsor Bolt.new, demonstrating how he built a fully functional fantasy football tracking app in roughly 2 hours using only his voice and a browser, with automatic pick tracking, live standings, and a password-protected admin panel.
• ▶ 4:53 Sponsor wrap-up: the code "electrified" unlocks a full month of bolt.new Pro for viewers.
• ▶ 5:00 The Starbase Louisiana news is acknowledged as already public, so the focus shifts to analysis and context rather than re-reporting headlines.
• ▶ 5:10 Due to the topic's depth, the host will break the Starbase Louisiana discussion into structured sections for easier digestion.
• ▶ 5:23 Elon Musk claims Starbase Louisiana will ultimately have over a dozen launch towers enabling more than 30 Starship flights per day, described as the biggest launch site on Earth.
• ▶ 5:42 Starbase Louisiana's remote location (~30 miles from the nearest town, ~50 miles from a major city) is critical because Starship is "insanely loud," making multiple daily launches from a more isolated site a much bigger deal than the raw cadence suggests.
• ▶ 6:36 SpaceX has stated that the recent pad 40 mission was the last planned Falcon 9 Starlink launch from Florida, with all future Florida Starlink missions shifting to Starship—a move flagged as a significant signal toward Starlink bandwidth inflection.
• ▶ 7:16 Cyber Cab demand is framed as a secondary but important consideration, setting up the broader analysis of how Starship launch cadence affects Starlink's ability to serve large-scale Tesla fleet connectivity.
• ▶ 7:21 Every Cyber Cab will be equipped with a built-in Starlink unit, creating a direct link between Tesla's robotaxi rollout and SpaceX's satellite internet constellation.
• ▶ 7:26 While Cyber Cab production will take years to ramp into the millions, it represents a clear long-term demand driver for Starlink capacity.
• ▶ 7:29 Starlink's goal is to scale to millions of users, but downlink capacity is the more meaningful operational metric than raw satellite count.
• ▶ 7:42 Current snapshot: ~11,000 satellites delivering ~850 Tbps of total downlink capacity to ~12 million subscribers, which theoretically equates to ~71 Mbps per user.
• ▶ 8:12 The 71 Mbps figure is a theoretical nameplate number—real-world usable bandwidth per user is significantly degraded by weather, gateway availability, backhaul constraints, spectrum congestion, and signal interference.
• ▶ 8:34 A single Starship V3 launch delivers 60 Starlink V3 satellites at ~1 Tbps each, adding ~60 Tbps of downlink per launch.
• ▶ 9:03 One full Starship V3 launch adds capacity equal to roughly 7% of the entire current Starlink constellation's total capacity, illustrating the massive network growth impact of each flight.
• ▶ 9:07 Current Falcon 9 / V2 mini launches deliver roughly 3 terabits per second (Tbps) of capacity to the Starlink constellation.
• ▶ 9:19 Starship V3 launches are expected to deliver approximately 20 times the capacity of current missions, equating to roughly 60 Tbps per launch.
• ▶ 9:27 SpaceX estimates that around 1,000 V3 satellites are required to meaningfully change the performance and behavior of the Starlink network.
• ▶ 9:48 Falcon 9 historically dedicated 75% of launches to Starlink, but SpaceX's AI segment makes the actual Starship mix currently unknown.
• ▶ 10:09 The analyst applies a 50% Starlink share assumption for Starship over the next 1-3 years, reflecting lower allocation than Falcon 9 due to expected diversification.
• ▶ 10:16 The remaining ~50% of Starship capacity will go to non-Starlink programs (NASA HLS, tankers, Starlink Mobile, and Star Mind), with Star Mind expected to become the largest category by the end of the decade.
• ▶ 10:57 SpaceX is currently authorized to operate up to 15,000 satellites under the Gen 2 approval, which represents only about 50% of their original request, limiting expansion as launch cadence increases.
• ▶ 11:08 The "Generation" framework (Gen 1, Gen 2, Gen 3) is a regulatory construct rather than a hardware designation, functioning as "boxes" that govern operational parameters—number of satellites, permitted orbits, and usable radio bands—filled with various satellite types.
• ▶ 12:03 SpaceX's original 2020 Gen 2 request sought approval for up to 30,000 satellites, but six years later they still have only half that capacity authorized, constraining their ability to fully scale Starlink services.
• ▶ 12:30 The Gen 2 authorization cap is set at roughly 15,000 operational satellites, while SpaceX holds approvals for approximately 19,000 satellites total between Gen 1 and Gen 2.
• ▶ 13:05 The current operational fleet consists of ~3,200 Gen 1 and ~7,900 Gen 2 satellites, with SpaceX no longer actively launching under the Gen 1 approval.
• ▶ 13:22 With 7,900 Gen 2 satellites already operational against the 15,000 cap, SpaceX has room for only about 7,100 additional operational satellites under current rules.
• ▶ 13:55 SpaceX has no headroom for growth under its existing Gen 1 approval, creating a critical regulatory constraint.
• ▶ 14:05 New Part 100 rules are being introduced to accelerate satellite licensing, and SpaceX has requested its Gen 3 proposal be processed under these faster rules.
• ▶ 14:18 The Gen 3 proposal is more complex than previous generations, involving larger satellites and the use of new spectrum bands.
• ▶ 14:40 At 100 Starship launches per year, dedicating half to Starlink would deploy 3,000 V3 satellites annually, adding 3,000 Tbps of capacity—equivalent to 3.5 times the entire current Starlink constellation in a single year.
• ▶ 15:14 SpaceX already has significant regulatory headroom under the existing Gen 2 approval, which permits an additional 15,000 satellites beyond those launched to date, without needing further approvals.
• ▶ 15:31 1,000 Starship flights per year, with 500 dedicated to Starlink launches, would add 30,000 new satellites and 30,000 terabits per second of capacity annually.
• ▶ 15:31 A single year of launches at this cadence would add 35 times the capacity of the entire current Starlink constellation.
• ▶ 15:42 A scenario of 5,000 Starship flights per year enables 2,500 dedicated Starlink launches, deploying 150,000 V3 satellites annually.
• ▶ 15:55 This adds 150,000 terabits per second of capacity per year—176 times the capacity of the entire current Starlink constellation in a single year.
• ▶ 16:12 At this cadence, deployment rate exceeds any near-term demand scenario, marking the point where the strategy is no longer constrained by build or launch capacity.
• ▶ 16:12 The analysis would hit authorization limitations, as SpaceX has not even filed for 150,000 Starlink satellites yet.
• ▶ 16:26 Hypothetically assuming 200,000 terabits per second constellation capacity and today's 71 Mbps per-user figure, the math implies approximately 2.8 billion Starlink kit subscribers.
• ▶ 16:44 Direct-to-cell subscribers would be additive to the 2.8 billion figure, and these numbers are a sense-check rather than a realistic projection.
• ▶ 16:56 At full constellation saturation, the launch mix will shift primarily toward Star Mind, with Starlink launches serving mainly to replace de-orbiting satellites and maintain the operational constellation.
• ▶ 17:12 To sustain a fully saturated Starlink constellation, SpaceX would need on the order of 500 Starship flights per year dedicated solely to replacement satellites.
• ▶ 17:12 These figures are illustrative to convey the immense scale involved, with a deeper dive into Star Mind and related numbers reserved for a future video.
• ▶ 17:25 The current figures are based on a conservative assumption, representing roughly half of the publicly stated long-term ambition.
• ▶ 17:32 The full ambition targets 10,000 Starship launches per year, which vastly exceeds the 2025 global total of just over 320 orbital attempts.
• ▶ 17:40 These baseline numbers are placeholders and will be refined once more concrete data becomes available.
• ▶ 17:50 SpaceX's new Louisiana launch site is pivotal for achieving Starship rapid reusability, marking a major inflection point for the aerospace industry.
• ▶ 18:18 Louisiana's geographic position enables launches at roughly 190° over the Gulf of Mexico, providing unrestricted range for polar, Sun-synchronous, and space-based data center orbits—advantages Cape Canaveral and Boca Chica lack.
• ▶ 19:20 With a target cadence of 1,000 Starship flights per year requiring ~1 million tons of methane and 3.5 million tons of liquid oxygen, Louisiana's proximity to the Henry Hub natural gas hub enables on-site fuel production that other sites cannot support.
• ▶ 19:47 Very few locations on Earth offer the necessary combination of coastline, empty land, safe trajectories, wide azimuths, and high-volume natural gas access, making SpaceX's first-mover advantage in securing this site a nearly insurmountable competitive moat.
• ▶ 21:02 Tesla FSD has a recurring failure mode with thin, horizontal linear objects like railroad crossings, chains, and caution tape, all sharing the same characteristic of being thin, linear, and horizontal.
• ▶ 22:53 CNN pooling layers can entirely average out thin horizontal lines because pooling combines pixels into larger groups, while vertical lines survive and even grow thicker as they get reproduced across pooling layers.
• ▶ 24:36 Tesla's stereo front-facing cameras naturally produce a strong disparity signal for vertical lines due to parallax (the two cameras capture vertical lines at different positions), giving FSD a second advantage in detecting vertical over horizontal objects.
• ▶ 25:36 Thin horizontal objects like chains can be just millimeters wide, while voxel resolution is 1–5+ cm, causing them to slip between voxels or get averaged out.
• ▶ 25:51 Environmental clutter such as painted road markings confuses the network, and it will likely dismiss thin horizontal objects as road markings, especially since training data for such hazards is poorly labeled and insufficient.
• ▶ 26:34 Detecting thin horizontal lines remains an open research problem with no reliable CNN solution found, though the speaker plans to revisit the topic with Tesla's V15 release.
• ▶ 27:30 Tesla (TSLA) closed at $348.75, down 1.71%, while SpaceX closed at $141.50, up 4.5%, against a broader Nasdaq 100 decline of 0.7%.
• ▶ 27:44 Dillon Loomis recommends Bolt.new as a low-friction solution for turning ideas into apps, especially suited for less technical users.
• ▶ 28:00 Closing remarks wish viewers a wonderful weekend and express a huge thank you to Patreon supporters.

• ▶ 0:00 Dashboard Updates and Sponsor Introduction (0:00 - 4:58) - - Welcoming viewers, updating the bet tracker dashboards, discussing the immense scale of Elon's projects, and presenting a sponsor segment for Bolt.new.
• ▶ 5:00 Starship Launch Cadence and Starlink Capacity Planning (5:00 - 17:46) - - Exploring Starbase Louisiana's location advantages, Starship launch cadence, Starlink V2/V3 capacity, and projecting subscriber metrics and satellite launches under various scenarios ranging from 100 to 5,000 Starship flights per year.
• ▶ 17:50 SpaceX Louisiana Launch Site Details (17:50 - 20:56) - - Analyzing the strategic advantages of SpaceX's new Louisiana launch site for Starship operations and rapid reusability.
• ▶ 20:58 Tesla FSD: The Horizontal Linear Object Problem (20:58 - 25:34) - - Breaking down why Tesla's Full Self-Driving system struggles with thin horizontal obstacles, covering CNN pooling limitations, stereo camera parallax, and motion parallax.
• ▶ 25:34 CNN Limitations and the Open Research Problem (25:34 - 27:17) - - Discussing the extreme detection challenges of millimeter-thin chains and presenting the unresolved research question of whether CNNs can reliably detect thin horizontal lines.
• ▶ 27:28 Market Closing Data and Closing Remarks (27:28 - 28:06) - - Sharing the day's closing stock prices for Tesla and SpaceX, and giving a final recommendation for the episode's sponsor, Bolt.new.