Brighter with Herbert Recap: 'Gun Range,' 'Spam' and SpaceX's Compute Build-Out

The latest full video on the official "Brighter with Herbert" YouTube channel, "Elon Just Revealed The Strangest SpaceX Project Yet!", was posted on November 3, 2026 and runs about 41 minutes. Host Herbert is joined by Jeff Lutz, whom the description introduces as a former supply chain executive at several Fortune 100 companies who now runs his own consulting firm. The "strange project" in the title turns out to be the names of two new data center buildings in SpaceX's Colossus 2 complex in Memphis, "Gun Range" and "Spam." From there the two discuss how fast SpaceX is adding computing power, how quickly it might pay for itself, the debate over data centers in space, and how much all of it could cost. Both are enthusiastic about Elon Musk's companies, so please read this as one side of the discussion.

This is an independent summary written by GNOUP. It is not affiliated with, sponsored by or endorsed by Brighter with Herbert, Herbert, Jeff Lutz, Elon Musk, SpaceX, xAI, Tesla, Nvidia, Google, OpenAI, Microsoft, Amazon, Meta, Sequoia or anyone else named in the video. The views, forecasts and claims described below belong to the speakers and the people they quote, not to GNOUP. Nothing here is investment, financial, legal or tax advice, and nothing here is a recommendation to buy or sell any security. The channel's own disclaimer says its content is for discussion only and that the host is long Tesla and other stocks. Please consult a licensed financial professional before making financial decisions.

Key takeaways

These points come from the video and its description. GNOUP has not checked the figures or claims.

• Why "Gun Range" and "Spam." Herbert says Musk replied to one of his posts with new details on Colossus 2. As the hosts read it, the "Gun Range" building was literally a local police gun range that was acquired, and "Spam" is a building that used to belong to Hormel, the maker of Spam. Per Musk's reply as read on the show, Gun Range is tracking to come online by the end of November and Spam, set up for inference (running AI models rather than training them), in December.
• The chip count. The description and the hosts say Musk expects Colossus 2 to have around 1.2 million Nvidia Blackwell chips by year end. Herbert walks through a tally from Brian Wang and Cern Basher that he says puts the site on a path to roughly 3 gigawatts of computing capacity by the end of the year, and he estimates each gigawatt could bring in $30 billion to $50 billion a year in revenue. These are the hosts' and commentators' estimates.
• Speed as the edge. Jeff argues that Musk's biggest advantage is hands-on execution: as Jeff describes it, Musk is working on site and clearing permitting, funding and technical roadblocks himself, decisions that he says take much longer at other big companies. Jeff calls compute the biggest bottleneck across Musk's companies right now.
• The compute "flywheel." Jeff's thesis is that SpaceX can rent out new computing capacity to other companies, earn back the cost within months or about a year, and later shift that already-paid-for capacity to its own uses, such as Grok, self-driving cars, robotaxis and Optimus robots. He compares it to mowing lawns now to fund the biggest landscaping business later.
• Telling the cost story. Jeff thinks SpaceX will probably still need to raise more capital, and says the company should show investors clearly how fast its spending turns back into cash (what finance calls the cash conversion cycle). In his view, the stock could be held back if that story is not told well.

Newsworthy insights

These are claims, predictions and opinions as the video presents them. GNOUP has not verified them, and other analysts may disagree.

• Bigger than rivals? The hosts show a chart from an AI research group that they say ranks Colossus 2 as the world's most powerful AI data center, ahead of OpenAI's Stargate site in Abilene. Herbert reads Musk's reply that other supercomputer projects are unable to reach their goals while SpaceX is achieving its own. Jeff asks a fair question in return: how SpaceX's pace compares with what Microsoft, Google, Amazon and Meta are adding, which he estimates at 1 to 3 gigawatts a year combined.
• Compute in orbit. Herbert reads Musk's prediction that the amount of computing power put into space each month will probably exceed what is added on the ground around the end of 2028. Herbert calls that timeline aggressive; Jeff says demand from physical AI devices will be far larger than today.
• Google's haiku from space. The hosts note that Google's Project Suncatcher test satellite, which they say launched on a SpaceX rocket, ran an AI model in orbit and sent a haiku back to Earth. Jeff, who says he knows people on that team, calls it a good first step and expects more companies to want rides to orbit.
• "Up to 100 times more." On the other side, the hosts quote a Google energy executive saying that space data centers can cost up to 100 times more than ones on the ground, partly because of launch costs and the difficulty of maintaining hardware in orbit. Jeff pushes back on a claim from a Sequoia partner that big tech firms lack hands-on hardware skills: he says the hyperscalers do have deep data center experience, but argues SpaceX can decide and move faster and has its own rockets.
• A $700 billion model. Per the description and the hosts, Sequoia partner Shaun Maguire built a model with SpaceX revenue rising from about $51 billion this year to more than $700 billion in 2028. Herbert notes that Maguire himself says the error bars are wide and that Sequoia is a major SpaceX shareholder.
• The capex question. An analyst at the newsletter Milk Road estimates, as read on the show, that SpaceX could spend about $470 billion on capital projects over the next five quarters, roughly twice what Amazon, Google or Meta each spend, while still carrying significant debt. Herbert reads Musk's reply that SpaceX's capex efficiency is "greatly underestimated."
• Risks named on the show. Jeff says government or geopolitical pushback could slow the build-out, and warns that China, with abundant power and manufacturing, could become a strong competitor if it solves its chip constraints.

Watch the video

Elon Just Revealed The Strangest SpaceX Project Yet! on YouTube (official channel "Brighter with Herbert," @BrighterwithHerbert, posted November 3, 2026, about 41 minutes)

Episode guide

Chapter times come from the video's own chapter list (titles lightly reworded). Each time opens the video at that point on YouTube.

• 0:00 The compute story gets bigger
• 2:25 Musk's details on the Colossus 2 buildings
• 4:59 Channel promotion (not summarized here)
• 5:26 Why Jeff thinks Musk out-executes rivals
• 10:22 The compute flywheel and payback
• 11:59 GPU counts and gigawatt math
• 14:41 Colossus versus the hyperscalers
• 18:17 The orbital compute debate
• 20:20 Google's haiku from orbit
• 23:02 What space data centers cost
• 29:35 The capex estimate and how to fund it
• 31:02 Explaining capex efficiency to investors
• 33:52 The "double tap" on compute
• 36:26 Geopolitics and China
• 40:31 Wrap-up

About this summary

This recap was prepared from the video's title, its description and chapter list on YouTube, and its automatic transcript. Automatic transcripts can garble names, so we kept to points we could follow clearly and left out names we could not confirm, including the Google energy executive and the Milk Road analyst. Posts, replies, charts and estimates discussed on the show are described as the video presents them; we have not checked them against the original sources. We also left out the channel's website, membership, merchandise, referral, product and course links, a political aside and a few sharper remarks about other companies. GNOUP has no connection to any product, offer or company mentioned and does not promote them. For the exact words and context, watch the video. This article is general information only. It is not investment, financial, legal or tax advice, it does not recommend buying or selling any security, and it is not a substitute for advice from a licensed financial professional.