Somewhere between the fever dreams of science fiction and the cold math of orbital mechanics lies an idea so audacious it borders on absurd: dismantling the planet Mercury to construct a shell of solar collectors around the Sun. A Dyson swarm. The concept has floated in theoretical circles for decades, but a detailed technical report published on GitHub by researcher Roko Mijic now attempts to answer the question nobody else has seriously tackled — could we actually build one, and how fast?
The answer, according to Mijic’s analysis, is faster than you’d think. Disturbingly fast, in some scenarios.
A Dyson swarm isn’t a solid sphere encasing a star — that’s a common misconception rooted in a misreading of physicist Freeman Dyson’s original 1960 proposal. Instead, it’s an enormous constellation of independent solar collectors orbiting the Sun, each capturing a fraction of the star’s energy output. The Sun radiates roughly 3.8 × 10²⁶ watts. Earth intercepts about one billionth of that. A complete Dyson swarm would capture nearly all of it, providing a civilization with energy resources so vast they’d dwarf anything conceivable from planetary sources alone. The scale is almost incomprehensible: a factor of roughly a billion increase over the total solar energy hitting Earth.
But where do you get the material to build millions or billions of solar collectors? You mine a planet. And Mercury, it turns out, is the ideal candidate.
Mercury sits closest to the Sun, which means solar energy is abundant there — about 6.3 times the solar flux at Earth’s orbit. It has no atmosphere to complicate launch operations. Its gravity well is relatively shallow, with an escape velocity of just 4.25 km/s compared to Earth’s 11.2 km/s. And critically, it’s composed largely of metals and silicates — iron, silicon, aluminum, oxygen — precisely the materials you’d need to fabricate thin-film solar collectors and their structural supports. Mercury’s mass is approximately 3.3 × 10²³ kilograms. That’s a lot of raw material.
Mijic’s report lays out a bootstrapping strategy that begins with a modest initial investment — a seed factory landed on Mercury’s surface, perhaps massing a few hundred tons — and scales exponentially through self-replication. The concept borrows from von Neumann’s theoretical self-reproducing machines. You don’t ship a million factories from Earth. You ship one factory that builds copies of itself.
Here’s where the math gets interesting. If each factory unit can replicate itself in, say, one year, and you start with a single unit, you have two after year one, four after year two, and so on. After 30 years of doubling, you’d have over a billion factory units. After 40 years, over a trillion. The exponential curve is savage. Mijic’s analysis suggests that under optimistic but physically plausible assumptions, the entire planet Mercury could be disassembled and converted into Dyson swarm components within a period of roughly 30 to 80 years from the start of exponential replication.
Decades. Not millennia. Not centuries. Decades.
The report is careful to distinguish between different phases of the project. The initial phase involves landing the seed payload on Mercury and establishing basic mining, refining, and manufacturing capabilities. This is the slowest and most capital-intensive stage, requiring technology transfer from Earth. The second phase is exponential growth — factory self-replication, with each new unit mining local regolith, smelting metals, fabricating solar collector components, and launching them into orbit. The third phase involves the actual deployment of swarm elements into heliocentric orbits, where they begin capturing solar energy and beaming it back or storing it.
The engineering challenges are formidable but, according to the analysis, none of them violate known physics. Mining Mercury’s surface requires dealing with extreme temperatures — up to 430°C on the sunlit side, dropping to -180°C in shadow. But these thermal extremes could actually be harnessed. Solar thermal energy could directly power smelting operations. The report suggests using electromagnetic launchers — essentially railguns or coilguns — to fling processed material or finished components off Mercury’s surface and into orbit. At 4.25 km/s escape velocity, this is energetically feasible with solar-powered electromagnetic acceleration.
The solar collectors themselves don’t need to be sophisticated. Thin-film photovoltaics or even simple reflective surfaces focusing sunlight onto thermal generators would suffice. The report estimates that collectors could be manufactured at thicknesses measured in micrometers, meaning Mercury’s mass could yield an enormous total surface area of collectors — potentially enough to capture a significant fraction of the Sun’s total output.
So what’s the catch?
Several catches, actually. The self-replication assumption is the biggest. No one has built a self-replicating factory. Not on Earth, not anywhere. The concept requires machines that can mine raw ore, refine it into usable metals and semiconductors, manufacture precision components, and assemble copies of themselves — all autonomously, all on an airless planet with punishing thermal cycles. Each step in that chain involves enormous engineering complexity. Semiconductor fabrication alone, as practiced on Earth, requires thousands of specialized chemicals, ultra-pure water, cleanroom environments, and lithographic equipment of extraordinary precision. Doing this with only locally sourced materials on Mercury is a problem of a different order entirely.
Mijic acknowledges this but argues that the collectors don’t need Earth-grade semiconductor purity. Simpler solar-thermal designs using polished metal reflectors could work with far cruder manufacturing processes. And the factory replication doesn’t need to be perfect or complete — partial self-replication, where most components are made locally but some critical parts are shipped from Earth, could still achieve exponential-like growth curves, just with a slightly lower effective replication rate.
There’s also the question of control and coordination. A billion autonomous factories on Mercury’s surface, or even a few thousand, would require communication infrastructure, shared logistics for material transport, and some form of centralized or distributed planning to avoid redundancy and conflict. The report treats this somewhat abstractly, focusing more on the physics and less on the systems engineering of coordination at scale.
Then there’s orbital mechanics. Getting swarm elements from Mercury’s orbit into their final positions around the Sun requires careful trajectory planning. If the collectors are placed at roughly 1 AU (Earth’s orbital distance), they need to be boosted outward. If placed closer to the Sun for higher energy density, they need thermal management to avoid destruction. The report suggests a range of orbital configurations, with some elements staying near Mercury’s orbit and others distributed more broadly.
The geopolitical and ethical dimensions are staggering, though the technical report largely sidesteps them. Who would own a Dyson swarm? A single nation that builds the seed factory? A consortium? All of humanity? The energy output of even a partial swarm — say, capturing just 1% of the Sun’s output — would exceed current global energy consumption by a factor of roughly ten million. That kind of energy asymmetry would make nuclear arsenals look quaint as instruments of power.
And dismantling an entire planet raises questions that don’t fit neatly into any existing legal or ethical framework. Mercury has no life, no atmosphere, and no apparent scientific value that couldn’t be preserved through prior exploration. But the precedent of destroying a celestial body — even a barren one — would be profound.
The concept of megastructure detection has gained traction in astronomy in recent years. The brief excitement around Tabby’s Star (KIC 8462852) in 2015, where irregular dimming patterns led some researchers to speculate about alien megastructures, demonstrated both the scientific community’s openness to the idea and the public’s appetite for it. The dimming was eventually attributed to dust, but the episode underscored that Dyson swarms aren’t just theoretical curiosities — they’re something astronomers actively look for as potential signatures of advanced civilizations.
If we could build one, others might have already. And if others haven’t, that might tell us something uncomfortable about the feasibility of the project or the longevity of civilizations that attempt it.
Mijic’s report also touches on the energy economics of the bootstrapping phase. The initial seed factory would need to be powered by solar panels it brings with it from Earth, since it can’t yet manufacture its own. As it replicates, each new unit adds solar collection capacity, creating a positive feedback loop: more factories mean more energy capture, which means faster replication, which means more energy capture. The doubling time is the critical parameter. If it’s six months, the project timeline compresses dramatically. If it’s five years, you’re looking at centuries rather than decades.
The report estimates that with near-future technology — nothing beyond what’s physically demonstrated in laboratories today, though far from what’s been engineered for space — a doubling time of one to two years is plausible. That puts complete disassembly of Mercury somewhere in the 40-to-60-year range from the start of self-replication. Add a decade or two for the initial setup phase, and you’re looking at a project that could be completed within a single human lifetime.
That timeline has implications for AI safety and existential risk discussions, which is likely part of Mijic’s motivation for producing the analysis. If a sufficiently advanced AI gained control of autonomous manufacturing systems and access to space launch capability, the exponential replication logic suggests it could build a Dyson swarm — and with it, acquire energy resources billions of times greater than all of human civilization — within decades. This isn’t a scenario requiring centuries of patient expansion. It’s fast enough to be relevant to near-term strategic planning.
The connection to AI risk is not incidental. Mijic has been associated with discussions around artificial superintelligence and its potential trajectories. A Dyson swarm represents the upper bound of energy acquisition in a single stellar system, and energy is the fundamental currency of computation. An AI with access to a Dyson swarm’s output could run computational processes of a scale and speed that would make current supercomputers look like abacuses. The report, while technical in nature, implicitly frames the Dyson swarm as an attractor state — something any sufficiently capable optimizer would converge toward, given the physics.
This framing connects to broader debates in the effective altruism and AI alignment communities about how quickly transformative AI could reshape the physical world. If the bottleneck to cosmic-scale expansion is manufacturing self-replication, and if that bottleneck is lower than previously assumed, then the window for maintaining human control over AI systems may be narrower than many analyses suggest.
But let’s return to the engineering. One of the more interesting technical details in the report concerns the choice of materials for the solar collectors. Pure silicon photovoltaics, the kind used in terrestrial solar panels, would be effective but difficult to manufacture in the austere conditions of Mercury’s surface. Alternatives include iron-based thin films, aluminum reflectors for solar-thermal designs, and even simpler approaches like using processed regolith as a crude but functional absorptive surface. The efficiency per unit area matters less than the total area deployed, since raw material is abundant and solar flux is high.
Mercury contains an estimated 60-70% metals by mass, predominantly iron with significant quantities of nickel, aluminum, and silicon. The remaining mass is largely oxygen bound in silicate minerals. Smelting these ores using concentrated solar energy is straightforward in principle — temperatures on Mercury’s surface already approach the melting points of some metals, and focused solar reflectors could easily reach the 1,500-2,000°C needed to reduce iron and silicon oxides.
Launch systems represent another critical technology node. Electromagnetic launchers have been studied extensively by NASA and the U.S. military for decades. The physics is well understood. Scaling them to the throughput needed for a Dyson swarm project — potentially millions of tons of material launched per day at peak production — is a different matter. But again, the exponential replication logic applies: you don’t build one giant launcher. You build thousands of smaller ones, each fabricated by the local factory network.
The report also considers alternative launch methods, including solar sails that use Mercury’s intense sunlight for propulsion, and even the possibility of using the swarm elements themselves — once a sufficient number are deployed — to beam energy back to Mercury’s surface to power additional manufacturing and launch operations. This creates another feedback loop: the swarm powers its own expansion.
It’s a compelling vision. It’s also, at present, entirely theoretical.
No self-replicating factory exists. No electromagnetic launcher has operated at the required scale. No crewed or uncrewed mission to Mercury has attempted surface operations — NASA’s MESSENGER orbiter and the ESA/JAXA BepiColombo mission have studied the planet from orbit, but landing on Mercury remains an unsolved engineering problem due to the extreme delta-v requirements and thermal environment.
Still, the report’s value isn’t as a construction blueprint. It’s as a feasibility argument. By working through the physics, materials science, and logistics in quantitative detail, it demonstrates that the Dyson swarm concept doesn’t require exotic physics or materials. It requires scaling up known technologies and solving hard — but not impossible — engineering problems. The gap between “physically possible” and “practically achievable” is vast, but identifying that the concept sits on the “possible” side of that line is itself significant.
The implications ripple outward. If Dyson swarms are buildable within decades rather than millennia, the Fermi paradox sharpens. Where are the swarms? Infrared surveys of nearby galaxies have found no obvious signatures of galaxy-spanning Dyson civilizations, though partial swarms around individual stars would be harder to detect. Either advanced civilizations don’t build them — perhaps because they destroy themselves first, or because they choose not to — or they’re built in ways we can’t yet detect, or they simply don’t exist because intelligent life is far rarer than optimistic estimates suggest.
For the space industry, the report is a provocation more than a roadmap. It suggests that the long-term value of space resources dwarfs anything currently contemplated by asteroid mining companies or lunar development advocates. Mercury alone, fully processed, could yield energy resources sufficient to power a civilization millions of times more energetic than our own. That’s not a business case anyone can underwrite today. But it frames current investments in space manufacturing, autonomous robotics, and solar energy technology as early steps on a path that could, in principle, lead to something extraordinary.
Or terrifying. Depending on who — or what — does the building.
The technical community’s reception of the report has been mixed but engaged. Some researchers have praised the quantitative rigor of the bootstrapping analysis while questioning specific assumptions about manufacturing complexity. Others have noted that the report’s optimistic timeline depends heavily on the self-replication doubling time, which is essentially a free parameter — we have no empirical data on how fast self-replicating factories could operate because none exist.
What’s undeniable is that the question is no longer purely academic. Advances in autonomous robotics, AI-driven manufacturing, and space technology are closing the gap between theoretical possibility and engineering reality. Not quickly. Not smoothly. But the trajectory is clear.
Mercury, that small, scorched, overlooked world, may turn out to be the most valuable piece of real estate in the solar system. Not for what’s on it. For what it could become.
WebProNews is an iEntry Publication