In this episode of the Shawn Ryan Show, Baiju Bhatt discusses his vision for solving data center challenges through space-based infrastructure. Bhatt addresses the growing concerns about terrestrial data centers—their escalating energy consumption, strain on power grids, water usage for cooling, and competition with communities for land and resources. He proposes a solution: solar-powered data centers operating in orbit, where continuous sunlight provides abundant energy and the vacuum of space enables cooling without water.
Bhatt explains the technical details of his company's Big Iron satellites and the Clydesdale rocket engine designed to launch them cost-effectively. He outlines plans to deploy a constellation of 20,000 satellites by the early 2030s, aiming to make orbital computing cost-competitive with terrestrial data centers. The conversation also covers broader themes of commercial space access, the limitations of government-dominated space programs, and Bhatt's goal to build independent infrastructure that enables diverse companies to pursue their own orbital missions.

Sign up for Shortform to access the whole episode summary along with additional materials like counterarguments and context.
Data centers are drawing increasing scrutiny for their escalating power and resource consumption, triggering concerns about environmental sustainability.
Baiju Bhatt observes that data center energy consumption is rapidly escalating, with projected gigawatt usage set to explode by 2028 and 2029. This surge places immense strain on existing power grids, raising fears about infrastructure capacity. Shawn Ryan notes that many Americans are concerned about the energy data centers consume, particularly as cooling these facilities demands vast quantities of water. This raises worries about water depletion, pollution, and rising utility costs for local residents, with Ryan highlighting first-hand accounts of communities facing water pollution and aquifer depletion tied to data center operations.
Bhatt describes data center expansions as fundamentally involving procurement of land and vital natural resources—energy, water for cooling, and delivery infrastructure. This competition pits data centers directly against local communities, often resulting in opposition. Ryan stresses that terrestrial data centers impose significant burdens on communities, with the need for large land parcels, substantial water access, and reliable energy intensifying competition for scarce resources.
Bhatt explains that satellites positioned in polar, sun-synchronous orbits are illuminated by the sun almost continuously. This allows solar panels on these satellites to generate more power per unit area than terrestrial panels, making the technology even more economical than traditional solar energy.
Each Big Iron satellite is massive—larger than a football field, weighing around 15,000 kilograms, and producing about a megawatt of power. Bhatt emphasizes the unique design: the entire upper stage of the rocket transforms into the satellite, eliminating redundant components. Inside are five racks of high-performance computing hardware and six high-bandwidth laser links for data transmission.
Bhatt addresses the cooling challenge: in the vacuum of space, only radiative heat transfer is available. The rocket's fairing unfolds to serve as a giant radiator, emitting thermal energy into space and eliminating the need for water-based cooling. Residual ammonia from fuel tanks can also recirculate through GPU cooling loops.
The vision includes building a constellation of 20,000 Big Iron satellites generating a combined 20 gigawatts—equivalent to Tennessee's entire power grid. Importantly, Bhatt notes, orbital data centers generate and consume all their power in space, placing no strain on ground-based grids.
The company aims to revolutionize space access by building rockets designed for reliability, manufacturability, and cost-efficiency.
The Clydesdale engine, a gas generator cycle engine producing 300,000 pounds of thrust, is specifically designed for ease of manufacturing and robustness. Bhatt points out this is enough to hurl 50 Ford F-150 trucks to orbit. The team built the gas generator from scratch in under five months, emphasizing swift, reliable production rather than immediate high performance.
The launch vehicle is 86 meters tall with a 4.7-meter diameter, designed to deliver up to 15 metric tons to low Earth orbit. Propelled by liquid oxygen and methane, it uses nine Clydesdale engines on the booster stage and one on the upper stage, streamlining production and maintenance.
Every component serves multiple purposes to minimize redundancy and mass. The fairing transforms into a radiator in orbit, ammonia propellant tanks double as GPU cooling systems, and cold gas thrusters also deploy solar panel booms. The booster is designed for reusability, with upper stage engines retired for parts after their mission, ensuring each piece works as efficiently as possible.
Bhatt describes a vision where orbital compute becomes cost competitive with terrestrial data centers by 2031 or sooner. As terrestrial costs climb, orbital costs will decline through economies of scale and reusability improvements. Economic viability requires maximizing revenue-generating equipment relative to structural mass, with dual-purpose parts limiting weight and reducing costs.
The company aims for its first orbital launch in 2028, with engine hot-fire testing by mid-2027 and booster testing between 2027 and 2028. Bhatt highlights that achieving this rapid pace requires prioritizing proven technologies and mature designs, focusing on the most direct path to a working product.
Each Trailhead facility includes five launch pads supporting launches every two minutes when weather allows. Each can be constructed for approximately $1 billion and deploy up to 660 megawatts of computing capacity annually. Multiple facilities will be necessary to scale up to 20,000 satellites.
Rocket manufacturing must become a scalable, repeatable industrial process. The company's 260,000-square-foot facility in Kent, Washington, serves as the first production hub, where the team expects to identify and overcome manufacturing challenges through hands-on execution before wider constellation deployment.
Bhatt, drawing on his lifelong exposure to aerospace through his father's NASA career, founded Cowboy Space to build independent space infrastructure and enable broad commercial activity beyond government programs.
For most of recent history, space access has been dominated by government programs, with pricing and missions dictated by shifting political priorities. Bhatt observes that the industry has operated "for the government, by the government, at government pricing," stifling long-term progress. He also notes that private sector consolidation has created monopolistic control, with one major company prioritizing its own missions. To break this dependence, Cowboy Space is developing independent launch capability and orbital infrastructure, enabling companies to pursue their own missions without being beholden to external agendas.
Cowboy Space introduces Big Iron Express, capable of sending third-party payloads of up to 18 tons into low Earth orbit. This launch service explicitly opens orbital access to external companies, lowering the threshold for commercial participation. Bhatt emphasizes that regular, diversified launch options are crucial to financial sustainability and economic abundance in space.
Bhatt argues that launch facilities should be seen as infrastructure assets like airports, attracting private capital and business services. Empowering innovators to identify profitable space applications will foster a diverse economy driven by individual initiative. True space commerce requires infrastructure, resources, and services to generate genuine economic value and job creation, and it is only by enabling more people to access space that humanity will unlock the full potential of the orbital economy.
1-Page Summary
Data centers are drawing increasing scrutiny for their rising consumption of power and resources, triggering widespread concern about their environmental footprint and sustainability.
Data centers' energy consumption is rapidly escalating, with Baiju Bhatt observing a dramatic rise in power drawn by data centers in recent years. By 2028 and 2029, the projected gigawatt usage is set to explode, adding to the already substantial demands seen in 2023. This surge in energy requirements places immense strain on existing power grids, raising fears about whether the current infrastructure can meet future needs without comprehensive restructuring.
Shawn Ryan notes that many Americans are increasingly concerned about the amount of energy data centers are using, echoing fears about power grid resilience and reliability. The issue extends beyond electricity; the cooling of these massive facilities demands vast quantities of water. This need raises significant worries about water depletion, damage to water quality, and rising utility costs for local residents. Ryan highlights how people's water and electric bills have been affected, and also mentions first-hand accounts of communities facing water pollution and aquifer depletion tied to data center operations.
Constructing and operating these installations is a vast real estate challenge. Baiju Bhatt describes data center expansions as fundamentally involving the procurement of land and vital natural re ...
Energy Consumption and Environmental Impact of Data Centers
Baiju Bhatt explains that satellites are positioned in polar, sun-synchronous orbits—also called terminator orbits—where they are illuminated by the sun almost continuously. This continuous sunlight allows solar panels installed on these satellites to generate more power per unit area than solar panels on Earth. Traditional solar energy is already among the most affordable forms of energy production; by putting solar panels in sun-synchronous orbit, energy output is further increased, making the technology even more economical.
Each Big Iron satellite is massive—larger than a football field and weighing around 15,000 kilograms. Every satellite produces about a megawatt of power, enough to supply roughly 800 homes. This quantity of power is significant in the context of data center operations. Bhatt emphasizes the unique design: Instead of filling the top of the rocket with stacks of satellites, the entire upper stage of the rocket itself transforms into the satellite, eliminating redundant components and reducing overall orbital mass.
Inside each Big Iron satellite are five racks of high-performance computing hardware, such as Nvidia NVL72 processors. The satellite is also equipped with six high-bandwidth laser links designed to transmit data to other satellites in the constellation or down to ground stations, supporting fast, flexible communication.
Bhatt addresses a crucial challenge of running data centers in space: cooling. Unlike on Earth, where heat is dissipated through conduction, convection (e.g., fans), and sometimes radiation, in the vacuum of space only radiative heat transfer is available. To tackle this, the rocket’s shell, or fairing, unfolds and serves as a giant radiator, emitting thermal energy into space. This design eliminates the need for traditional water-based cooling equipment. Additionally, residual propellant from the ammonia-based fuel tanks used for orbital maneuvering can recirculate through GPU cooling loops, helping transfer heat to the radiator panels for efficient dissipation.
Solar-Powered Space-Based Data Centers Cooled by Space Physics
The company aims to revolutionize space access by building rockets designed for reliability, manufacturability, and cost-efficiency, focusing on modular, dual-purpose components throughout the system.
The core of the new rocket is the Clydesdale engine, a gas generator cycle engine specifically designed for ease of manufacturing and robustness, rather than maximizing performance from the outset. Each Clydesdale engine produces 300,000 pounds of thrust, which Baiju Bhatt points out is enough to hurl 50 Ford F-150 trucks to orbit—a scale of power that is difficult to imagine.
The gas generator, a central element of the engine that drives the turbo pumps, was quickly developed and tested at NASA's Stennis facility. The team built this critical component from scratch in under five months, and their first earnest test was recently completed. The emphasis for the Clydesdale is not immediate high performance but swift, reliable production to bring orbit delivery capability online as quickly as possible.
The launch vehicle is 86 meters tall with a 4.7-meter diameter, deliberately chosen to push the boundaries of road-transportable rockets. Propelled by liquid oxygen and methane and constructed from stainless steel, the rocket is designed to deliver data centers—up to 15 metric tons—to low Earth orbit (LEO).
The rocket's booster stage is equipped with nine Clydesdale engines, while the upper stage employs a single Clydesdale to insert payloads into orbit. This shared engine architecture between stages not only streamlines production but also simplifies maintenance and refurbishment.
A hallmark of the vehicle's design is modularity and multifunctionality—every component is engineered to serve more than one purpose, minimizing redundancy and mass.
The rocket's fairing, serving as an aerodynamic shell during ascent, transforms into a radiator once in orbit. Instead of each satellite requiring its own radiator, the nose cone structure is reused as a large, effective radiator, saving weight and complexity.
Ammonia propellant tanks, initially intended for orbital maneuvering, also double as GPU cooling systems in space. Residual amm ...
Rocket Development: Cost-Effective Orbit Delivery With Clydesdale Design
Baiju Bhatt describes a vision in which orbital compute becomes cost competitive with terrestrial data centers by 2031—or sooner, if possible. As terrestrial compute costs climb, Bhatt anticipates that orbital compute costs will decline through economies of scale and continuous improvements in reusability. The fundamental challenge is to achieve cost parity for a GPU hour in space versus on the ground, despite the apples-to-oranges differences in physics and infrastructure requirements. Terrestrial computing depends on real estate, power, and cooling systems, while orbital compute relies on rocket structures, methane-fueled engines, and multi-year orbital deployments. Economic viability requires maximizing the proportion of revenue-generating equipment relative to the mass devoted to structural and propulsion systems. The goal is to achieve an entire vehicle design where parts perform dual roles, limiting weight and reducing costs.
The company aims for its first orbital launch in 2028, demanding both an accelerated timeline and extreme focus. Bhatt highlights key milestones, including engine hot-fire testing by mid-2027 and booster testing between 2027 and 2028. Achieving this rapid pace requires a philosophy centered on hands-on capability—"the best way to get good at building rockets is by building rockets." Instead of extended planning, the company prioritizes proven technologies and mature designs to meet aggressive deadlines, always focusing on the most direct, minimum path to a working product. The design prompt starts with one question: How can cost parity with ground-based computing be achieved as efficiently as possible?
The Trailhead facility concept supports high-frequency orbital data center launches. Each Trailhead includes five launch pads, with four active at any time while one is under refurbishment. This setup supports launching rockets every two minutes, provided weather allows, ensuring readiness for frequent sun-synchronous orbital insertions. Each Trailhead can be constructed for approximately $1 billion and is designed to deploy up to 660 megawatts of computing capacity annually. The use of sun-synchronous orbits allows for consistent booster return to the same site, streamlining refurbishment and enhancing reusability. Multiple Trailhead facilities in U.S. coastal regions ...
Path to Cost-Competitive Orbital Compute: Economics, Scalability, Modular Launches, and Reusability
Baiju Bhatt, drawing on his lifelong exposure to aerospace through his father’s career at NASA, has observed the transformation and stagnation of the space industry. He recalls that as a child, NASA seemed more like a museum showcasing historical achievements, raising the question of why meaningful progress had slowed. These reflections ultimately informed the founding of Cowboy Space, a company devoted to building independent space infrastructure and enabling broad commercial activity beyond government programs.
For the majority of recent history, access to space has been dominated by government-driven programs, with pricing, missions, and progress dictated by shifting political priorities. This dependency means that as government interests waver, so do the resources and momentum behind space activities, stifling long-term progress and deterring widespread industry participation. As Bhatt points out, the industry has largely operated “for the government, by the government, at government pricing,” perpetuating cycles based on public-sector whims rather than sustainable commercial incentives.
Bhatt observes that consolidation in the private sector has also become a barrier, with a single major company controlling most orbital access and prioritizing its own missions. To break this dependence and foster a commercially vibrant space environment, Cowboy Space is developing independent launch capability and orbital infrastructure. Their vision includes operating a data center in space and building an energy grid that supports a variety of corporate uses, enabling companies to pursue their own missions and self-betterment without being beholden to external agendas. Bhatt stresses this infrastructure allows people to build and profit in space in the “most American of senses”—through their own initiative and enterprise.
Sustainable commercial incentives and free-market participation are essential for the long-term prosperity of the orbital economy. When only governments or monopolistic providers control space access, true economic abundance and opportunity remain out of reach. Bhatt envisions a generational shift where young people see viable careers and entrepreneurial opportunities in space, driven by recurring, market-based reasons to operate and innovate beyond simply executing government-driven missions.
Cowboy Space introduces the Big Iron Express, a new launch offering capable of sending third-party payloads of up to 18 tons into low Earth orbit. This launch service explicitly opens orbital access to external companies, lowering the threshold for commercial participation and encouraging new business models, research projects, and technological advancements beyond government or singular corporate interests.
Bhatt emphasizes that regular, diversified launch options are crucial to financial sustainability and economic abundance in space. If only one entity controls the gateway to orbit, it prevents the development of a robust economy and stifles innovation. The success of Cowboy Space’s business and the larger vision for the industry depend fundamentally on expanding who can d ...
Commercialization of Space: Enabling Corporate Orbital Access and Driving Economic Opportunity Beyond Government Programs
Download the Shortform Chrome extension for your browser
