SpaceX Transcript: The Roadmap to IPO, xAI Integration, and the First-Principles Revolution of Space Manufacturing
June 21, 2026 - Exclusive Interview with Former SpaceX Chief Rocket Manufacturing Engineer Lewis
Introduction and the ChatGPT Moment in Space
Xiaojun: Hello everyone, I am Xiaojun. Standing in 2026, SpaceX has just completed the acquisition and integration of xAI, and it is widely expected to complete an IPO this year. Today, we invited Lewis, the former chief rocket manufacturing engineer of SpaceX and now GP of Aris Fund, to talk about it. Is this a prelude to the expansion of human civilization? Here is my interview with Lewis.
Lewis: Imagine you start driving today from here to Los Angeles. Right when you pull into your driveway, your very last drop of fuel runs out. That is how a rocket operates—it is completely actuarial. If SpaceX really goes public this year, I think the mainstream market will finally foot the bill for the space sector. To scale up production and development, the first team SpaceX brought in was actually from Mini Cooper. When someone told Elon Musk that a task was impossible, he just said very calmly, Okay, then I accept your resignation. Our competitor is ourselves, but we closely monitor China's aerospace development.
Lewis’s Journey to SpaceX and Early Milestones
Xiaojun: Hello, Lewis. Why don't you say hello to the audience and briefly introduce yourself?
Lewis: Hello everyone, my name is Lewis. I am the co-founder of Aris Fund, and I am also a GP. We are an early American venture capital firm based in Los Angeles, specializing in hard technology projects in the United States, whether in space, energy, or what is now called physical artificial intelligence. Regarding my early career, I joined SpaceX in 2012 and stayed there for 7 years, leaving in 2019. When I left, I was in charge of 7 departments and more than 3,000 rocket parts. After leaving SpaceX, I joined a battery company. In 2019 and 2020, we were lucky to catch the electric vehicle trend and had a successful exit in 2020. My current partner in Aris Fund comes from Tesla; he was an executive who helped build the Fremont factory and founded the Shanghai Gigafactory. That is where we met, realized the golden hard tech opportunities in the United States over the next 20 years, and decided to start angel investing in 2020. We established our first fund last year to support hardware entrepreneurs.
Xiaojun: We invited you today because you know the aerospace industry so well. It feels like the industry is at a wonderful turning point. SpaceX acquired xAI at the beginning of this year, and everyone thinks they will go public in 2026. How do you view this point in time?
Lewis: From a market perspective, it feels like this year is the ChatGPT moment for space. People were curious about the industry before, but lacked conviction. Since Elon Musk mentioned a potential SpaceX IPO, we have received a steady stream of inquiries asking for SpaceX shares and space tech investments. From a capital perspective, this is a massive turning point. However, looking at actual space development, it is just an ongoing execution of a roadmap that started over a decade ago. I joined SpaceX in 2012 and had my first meeting with Elon in 2013. He repeated the mission constantly: the first step is to recycle the rocket in a stable, quantifiable, and economical way. We thought it might take 10 years to dominate the market. We started in 2013, and the first successful orbital rocket recovery occurred on December 21, 2015. That is a highly memorable day for everyone at SpaceX. We launched, reached orbit, and successfully recovered the rocket, which is now displayed at our headquarters in Los Angeles. That was the real turning point, demonstrating that a reusable space age was possible, even though industry experts had claimed it defied economic logic.
The Long-term Roadmap: From Falcon 9 to Starship
Xiaojun: What was the SpaceX roadmap back in 2012?
Lewis: In 2012 and 2013, right when our first rocket successfully launched, the company could barely manage one launch per year. Yet, we already established a Mars team. Elon recognized that our current rockets could not meet the needs of exploring the universe, so we needed a stronger team and a rocket designed for that mission. What we now call Starship was born back then. He is always executing one generation while planning the next. Starlink, Starship, and space data centers were planned over ten years ago. This year is indeed a ChatGPT moment for the public and for capital, but the execution has been underway for a decade. The roadmap was not written as a rigid prediction of 2025 or 2026, but the master plan has always been simple: expand the universe for human exploration. He never positioned SpaceX as just a rocket company.
First-Principles Thinking: Starlink, Space Data Centers, and xAI Integration
Xiaojun: In 2012, no one expected the explosion of large language models. How do you view the SpaceX acquisition of xAI? Is this a prelude to the expansion of human civilization?
Lewis: It was not explicitly in the 2012 master plan because nobody predicted the AI boom, but the core logic fits. We wanted to build a low-orbit communication network to create a new data highway. Starlink was planned in 2015 after we achieved rocket recovery. From a physical first-principles perspective, transmitting messages in the vacuum of space is twice as fast from point A to point B compared to the most advanced fiber optics on Earth. That is a natural physical advantage. Now, what does AI need? Huge computing power and massive data centers. In the United States, building physical data centers is incredibly slow and difficult because of bureaucratic permits, zoning laws, and a severe power grid deficit. The US power grid consists of only three major grids—East Coast, West Coast, and Texas—and most facilities are over 30 years old, with the latest nuclear plants built in the 1990s. Even before AI, the power grid had a 25% to 40% deficit. You cannot easily take electricity away from residents to feed AI data centers. Space solves these constraints. There are no Earth-bound permits, space has endless solar energy with a conversion rate that is 10% more efficient than on Earth, and there are no geographic restrictions. Integrating xAI with SpaceX and potentially building space data centers is logical. People previously speculated Tesla would acquire xAI because of autonomous driving and physical AI. It is possible that all of Elon's companies—Tesla, SpaceX, Neuralink, Boring Company, and xAI—will eventually integrate under a single holding company, X.com, which is a domain he bought back in 2014. If you look closely, they all connect to a grand master plan. Neuralink provides the brain-computer interface for human-AI interaction, Boring Company provides underground shelter systems for space environments with high radiation, and xAI provides the intelligence. They are not accidental standalone companies; they are clues that could join together under X.
The Grok Model and Physical AI
Xiaojun: The xAI vision is to understand the nature of the universe, and its logo symbolizes a black hole. How does it compete with other AI giants?
Lewis: From a competitive perspective, each AI giant has a specific footing. Gemini has Google search, Claude targets coding and business-to-business tools, OpenAI targets the general consumer public, and Grok has unlimited data from X. Grok's ultimate strength within Elon's ecosystem lies in physical artificial intelligence—the world model and entity AI. Last year, they built a massive 100,000 GPU super data center, and we will see the results of those investments in 2026. The organizational cultures of SpaceX and xAI are highly compatible because they both share Elon's ethos: move fast and break things, with rapid execution and iteration. If we send humans to explore outer space, our most reliable companion will be AI, similar to the Moss system in science fiction.
Lewis's Background and Stanford Days
Xiaojun: Let’s discuss your background before SpaceX. How did your journey begin?
Lewis: I was born in Taiwan, China. Because my parents worked for a German company, we moved abroad when I was very young. I lived in Germany, Hong Kong, and then Guangzhou, where I spent 8 years and finished high school. In 1999, I applied from Guangzhou and was admitted to Stanford University for a Bachelor of Mechanical Engineering. Back then, international students were a tiny group; Stanford admitted only two students from the entirety of China that year—myself and a classmate from Beijing. After graduating in 2003, I had the choice to be an early Google employee or to join a Fortune 500 company as an automated production engineer. At that time, startups and venture capital were not yet mainstream. The career fairs at Stanford featured giants like Boeing, Microsoft, General Motors, and Ford, while startups like Google were relegated to small side tables. I followed the mainstream and joined Epson at their northwest headquarters in Portland, Oregon.
Xiaojun: What did you do at Epson?
Lewis: I worked on the automated assembly lines for ink cartridges, which was Epson's most profitable business. While a similar manual production line in China required 55 workers, we designed fully automated lines in the US that required only two workers. I became the youngest executive in the Epson Group globally in my early twenties, managing a factory of over 70 people. My boss was Japanese-American and trusted me completely. I stayed there for a few years, but eventually felt the need for a new challenge. I decided to pursue an MBA at Oxford in 2006 to pivot into investment banking. However, I graduated in 2008, on the exact day Lehman Brothers went bankrupt. The financial crisis hit Europe and America hard, and my job offer was withdrawn. It was the first time I faced extreme difficulty finding a job. I eventually moved to Hong Kong to work for a Canadian supply chain consulting firm. In 2010, my boss and I started a food manufacturing business. My family has three generations of food industry experience, so we brought supply chain engineering principles to food. Within two years, we opened 7 stores and a central kitchen, generating great income.
Getting the Call from SpaceX
Xiaojun: How did you transition from a food business to making rockets?
Lewis: In 2012, a close friend from Stanford, who was among the first 100 employees at SpaceX, called me. He told me SpaceX was flying upward and invited me to join them in Los Angeles to build rockets. I had never even heard of rockets being built in Los Angeles, and I was not an Elon fan back then. In 2012, Elon was not the globally famous figure he is today; he was just a successful entrepreneur who had invested his PayPal fortune into two highly risky companies, Tesla and SpaceX. At that time, I faced two choices: join Amazon Japan, which was just starting up and offered a highly secure and lucrative path, or join SpaceX to explore the universe. Everyone around me urged me to choose Amazon. They asked, what are the chances in space? It was a tiny industry launching just a few satellites a year. However, I realized that if SpaceX actually succeeded, it would change the world forever. I knew I would regret it for the rest of my life if I did not join. So, I chose SpaceX and joined as an engineer in the Dragon spacecraft department when the company had just over 1,000 employees.
The Flat Organization and the Move to Falcon 9
Xiaojun: What was the organization of SpaceX like in the early days?
Lewis: It was incredibly flat and had extremely high talent density. Compared to Boeing or Lockheed Martin, which had 10,000 people working on similar projects, SpaceX had a 1-to-10 ratio. Every department was managed by working engineers who were personally responsible for design, procurement, and production. If someone was hit by a train, it left a massive gap because there was very little redundancy. The team was young, mostly in their twenties and thirties. We had a joke that if you survived at SpaceX for six months, you were a veteran. The turnover was high because the environment was demanding. Initially, I worked on the Dragon spacecraft, specifically designing the interior cargo holds and astronaut interfaces. It was essentially like designing the world's most advanced, lightweight IKEA furniture. The cabinets had to withstand high gravity and pressure, requiring custom material engineering. After a year, because of the success of my projects, I was promoted to Falcon 9 manufacturing. In 2012, there were only 77 rocket launches globally across all nations, and SpaceX launched maybe once a year. Yet, Elon set a company-wide goal to scale up to launch at least 40 times a year. It sounded insane. To achieve this, SpaceX hired a production team from Mini Cooper. Mini Coopers are the car series with the highest number of SKUs and design variations. Their assembly lines are designed for high volume but extreme variability. This was exactly what early Falcon 9 needed because we iterated constantly; no two Falcon 9 rockets were identical because we applied data from each launch directly to improve the next one.
First-Principles and the Coke Can Analogy
Xiaojun: How did you manage that scale-up?
Lewis: I was put in charge of the high-pressure helium tanks on Falcon 9. These are large, two-meter-long tanks that must be incredibly light but withstand tens of thousands of psi. Originally, we purchased them from external suppliers for over $100,000 each, and they could only deliver a few per year. Elon told us to build them ourselves. We engineered our own production line, cut the costs by 90%, and achieved the volume needed for our launch schedule. We were incredibly proud and presented the finished product to Elon. He did not say Good or Bad. Instead, his first question was, Have you ever seen the production of Coke cans? He explained that a Coke can is also a pressure vessel. It is manufactured at a rate of thousands of cans per minute for pennies. Even with our 90% cost reduction, we were nowhere near that efficiency. This is a perfect example of his first-principles thinking. He does not compare you to the aerospace industry; he compares you to the physical limits of manufacturing.
Working with Elon and the Crisis of 2015-2016
Xiaojun: What is it like to have a meeting with Elon?
Lewis: Elon only meets with you for three reasons: your project is late, you cannot build it, or it needs more money. He is an extremely intense and demanding boss, but he is genuinely the chief engineer. He understands the physics of every component. I remember a meeting where a senior engineer was presenting a design progress report that was behind schedule. The engineer eventually told Elon, what you want us to do is physically impossible. Elon looked at him very calmly and said, Okay, then I accept your resignation. The room went completely silent. The engineer just said okay and walked out. Elon did not yell, scream, or throw chairs; his calm dismissal was far more terrifying. Yet, he is also highly sincere. When we succeeded, he would be as happy as a child. I spent 22 iterations revising our slide deck before presenting our automated high-pressure tank production line to him, and he approved millions of dollars for it. The hardest period for the company was 2015 and 2016. On June 28, 2015, which was Elon's 44th birthday, we had a major Falcon 9 failure. The rocket exploded shortly after launch. It was a dark period because we had just achieved our first successful recovery, and suddenly we failed on the world stage. Because I managed almost all Falcon 9 departments, the failure was directly linked to my area of responsibility. We spent months working 12 hours a day, 6 days a week—a standard 996 schedule in the US without any overtime pay—to find the root cause. We tested every possible scenario and eventually resolved a highly obscure design anomaly. Since then, we have completed over 600 successful launches.
Leaving SpaceX for the Electric Vehicle Boom
Xiaojun: Why did you leave SpaceX in 2019?
Lewis: On December 31, 2018, Elon sent an email to the entire company offering free Full Self-Driving to anyone who ordered a Tesla that day. I ordered a Model 3, which was my first electric vehicle. Driving that car was a profound experience. From a mechanical engineering perspective, I realized that the electric vehicle platform was vastly superior to internal combustion engines. I knew that the next massive boom would be in electrification. So in 2019, I left SpaceX and joined an early-stage battery company focused on heavy-duty commercial trucks, which we successfully exited in 2020.
The Space Ecosystem and Advanced Manufacturing
Xiaojun: How do you map the space industry ecosystem from upstream to downstream?
Lewis: Upstream is the physical hardware: building rockets, engines, satellites, and space stations. Midstream is the control and data transmission layer: satellite management and downlinks. Downstream is the application layer, such as Starlink providing high-speed global internet. But the future of downstream is much larger. The unique weightless and vacuum environment of space allows us to manufacture things that are impossible to make on Earth. For example, we can manufacture perfect, flawless silicon wafers for next-generation microchips without spending billions simulating a vacuum on Earth. We can also manufacture perfect artificial corneas for medical use. The space industry will divide into two ecosystems: the vertical SpaceX ecosystem, which is like Apple, and the open Android ecosystem. A group of former SpaceX and Tesla employees—the SpaceX Mafia—are now starting hard tech companies in the US, building supersonic planes, mini nuclear reactors, and advanced chips. They know how to build physical hardware in the US from zero to one, bypassing the software-only focus of the last 20 years.
The SpaceX Mafia and the Future of Space Stations
Xiaojun: What about China's aerospace progress?
Lewis: Elon's exact words when we asked about competitors were, Our competitor is ourselves and our speed of progress, but we closely monitor China's aerospace development. China has massive national team resources and can iterate manufacturing incredibly fast. The future of space will be dominated by the US and China. Regarding our own investments, Aris Fund recently invested in a space station startup founded by former SpaceX engineers. They are designing the world's first space station with controllable gravity. In a zero-gravity environment, even simple tasks like using the toilet are incredibly complex engineering problems that require vacuum suction. By utilizing centrifugal rotation to create controllable gravity, we can solve these fundamental human challenges in space. This is a pure engineering problem, not a scientific one, and it will unlock the next era of commercial space stations as the International Space Station prepares for decommissioning in 2030.
Conclusion and Book Recommendation
Xiaojun: How does Elon manage his time across so many companies, and what book would you recommend to our audience?
Lewis: Elon does not micromanage; he only shows up to solve the most difficult engineering bottlenecks. His weekly schedule when I was there was brutal: SpaceX meetings every 20 minutes on Monday, fly to Tesla on Tuesday and Wednesday, back to SpaceX on Thursday, and work on weekend projects like Boring Company or Neuralink. He would sleep on the Tesla factory floor after 11:00 PM to inspect the night shift operations. He is a truly singular leader. I recommend the book Silicon Valley Iron Man because it focuses on the engineering and business execution that makes him unique, rather than his personal or political life. I believe SpaceX is a trillion-dollar company that will become the modern equivalent of the East India Company during the Age of Discovery. I have never sold my core shares, and I do not plan to. Thank you for having me.
Xiaojun: Thank you, Lewis. Our studio is called Language is the World Studio. Hearing your stories about outer space and advanced engineering truly shows how language and communication explore the greatest things in our universe. Thank you for this comprehensive conversation.