I wanted to build a mech. Not design one for a game or draw one for fun. Actually research what it would take to build one. So I did.

I TRIED TO DESIGN A REAL COMBAT MECH Here's Why We Need 100 More Years A love letter to the future that refuses to be built yet I wanted to build a mech.

Not design one for a game or sketch one for fun. Actually research what it would physically take to build one — the power systems, the materials, the AI, the structure. I wanted to know where reality ends and science fiction begins. It started, like most things in my life, with Batman. Specifically the Batwing — that sleek, impossible aircraft that somehow feels real even though it shouldn't. Wayne Enterprises feels like an actual defence contractor. Every panel and thruster has internal logic. I wanted to find the line between that kind of grounded fictional engineering and what we can actually build today. 

So I spent weeks researching. Not sketching cool designs. Digging into materials science, fusion energy, bipedal locomotion, AI architecture, and aerospace engineering. I wanted to answer one question honestly: if someone handed me unlimited funding right now, could I build an 18-metre bipedal combat mech? The answer is no. Not even close. We need roughly 100 years of technology that does not exist yet. Here is what I found — and where we actually are.

 I. The Power Problem: We Don't Have an Engine for This Why Everything Starts with Fusion Everything starts with power. An 18-metre mech weighing around 32 tonnes — capable of ground movement, flight, weapons discharge, and AI processing simultaneously — needs somewhere between 80,000 and 100,000 kilowatts of continuous output. For reference: an F-35 fighter jet at full afterburner produces roughly 70,000-80,000 kW. But a jet is optimised purely for thrust. A mech needs to walk, carry weapons, power sensors, run an AI system, and maintain structural integrity under combat stress — all at the same time. The power demand is not comparable. It is categorically different. The only energy source that could realistically meet this need is miniaturised fusion. Not the fission reactors we have today — those are too massive, too fragile, and too politically complex to put inside a walking weapons platform. Fusion is what you need: energy-dense, theoretically compact, and clean enough that the thermal signature does not become its own vulnerability. "The only energy source dense enough to power a real mech is fusion. We do not have working fusion yet." The problem is fundamental. ITER — the world's largest fusion experiment, built by 35 nations over three decades — is still in assembly in southern France. It has not produced a single watt of net energy yet. Private companies like Helion Energy are targeting demonstration reactors in the late 2020s, and some are making genuine progress. But even if Helion succeeds in the next five years, the gap between 'demonstration reactor in a warehouse' and 'miniaturised fusion core small enough to fit inside a mech torso, survive ballistic impact, and restart after damage' is not a decade of engineering. It is probably 50-70 years of materials science, containment physics, and manufacturing development that has not begun yet. Without the reactor, the mech is a very expensive statue. Everything else in the design — the structure, the weapons, the AI, the sensors — is irrelevant until the energy problem is solved. System Current State ITER (fusion demo) Under assembly — no net energy yet Helion Energy Targeting first plasma mid-2020s Miniaturised fusion core Theoretical — no prototype exists Mech-scale power delivery 50-70 years minimum from today The fusion roadmap — where we are vs. what a real mech needs

 II. The Walking Problem: Boston Dynamics vs 32 Tonnes Why Scale Is the Enemy Assume we solve fusion. Now we have to make the thing walk. Boston Dynamics' Atlas is the most advanced bipedal robot in the world. It runs, jumps, does backflips, and navigates uneven terrain with remarkable agility. It weighs 89 kilograms and operates for roughly 90 minutes before its battery fails. Now scale that up to 32 tonnes and 18 metres tall. The structural physics involved in bipedal locomotion scale catastrophically with size. Every time a mech takes a step, the ground contact force is approximately 2.5 times its body weight — roughly 80 tonnes of force transmitted through a single foot contact. The ankle joint alone must survive stresses that no material we currently manufacture can endure repeatedly without catastrophic fatigue failure. This is the Square-Cube Law — one of physics' most brutal constraints on engineering ambition. As you scale an object up, volume and mass grow as the cube of the linear dimension, but structural cross-section grows only as the square. A structure twice as tall must support eight times the mass through only four times the cross-sectional area. Insects can lift many times their own body weight. Elephants struggle with their own. A real mech is not just a bigger Atlas. It is a categorically different engineering problem. "The Square-Cube Law is the mech's first enemy. Physics scales mass faster than it scales strength." Carbon nanotube composites are the most promising structural candidate. In laboratory conditions, they are roughly 100 times stronger than steel at one-sixth the weight — exactly the combination needed for a mech frame that must be strong enough to survive ground forces and light enough to be moved by any plausible actuator. But laboratory conditions are doing enormous work in that sentence. We cannot manufacture carbon nanotube structural components at the scale needed for an 18-metre frame. Current production yields centimetres of material, not metres. The production infrastructure, quality control methods, and manufacturing processes needed do not exist. Developing them — at scale, at cost, with consistent mechanical properties — is a 30-50 year project even with sustained investment. Scale Challenge Why It Matters Ground contact force ~80 tonnes per step — no joint material survives this repeatedly Actuator power density Current motors cannot move 32 tonnes at speed Carbon nanotube production Lab scale only — metres of material, not metres Ankle joint fatigue life No material rated for 80T cyclic loading exists Self-weight locomotion Square-Cube Law makes every metre of height more expensive The structural physics problems — each one is unsolved at mech scale

 III. The One Technology That's Actually Almost Here The Neural Interface Arrives First Here is where the research surprised me most. In my design, I built an AI command system called ATHENA OS with a neural-link interface providing less than 10 millisecond response latency between pilot thought and platform response. I assumed this would be the most science-fictional element of the whole project. The fusion reactor felt almost plausible by comparison. Direct brain-to-machine interfacing at millisecond speeds seemed like it belonged in 2126, not anywhere near today. I was wrong. Neuralink implanted its first human patient in January 2024. That patient — Noland Arbaugh, a 29-year-old quadriplegic — can control a computer cursor with thought alone. He plays chess. He browses the internet. The neural signal is read, decoded, and translated into machine commands in real time, with latency already in the millisecond range. "The neural interface is the one technology in my design that is not theoretical. It exists in a living human being right now." A full neural-link capable of managing the ATHENA OS architecture — processing sensor data from 300km range, coordinating weapons systems, managing power routing, and executing mobility decisions — is still decades away. The leap from cursor control to mech control is enormous. But the foundation is real. The physics works. The biology is being solved. Of all the technologies in the Aegis Prime design, the neural interface is the one I would bet money on arriving within 20-30 years. The fusion reactor has no working prototype. The carbon nanotube frame has no manufacturing pathway. The neural link has Noland Arbaugh moving a cursor with his thoughts. The future has a hierarchy, and the brain-machine interface is near the top of it.

 IV. The Materials That Are Surprisingly Real Shape Memory Alloys and the Morphing Vehicle While researching the mech, I also designed a second platform — the Triton X-1, a civilian-military tri-modal vehicle capable of seamless transitions between road, air, and submarine modes. The morphing mechanism I used — Shape Memory Alloys driving structural reconfiguration in under 12 seconds — turns out to be closer to reality than almost anything in the mech design. Nitinol — a nickel-titanium shape memory alloy — is real, commercially available, and in active use today. It powers the stents that keep cardiac patients alive. It shapes the wires that straighten teeth over months of slow, constant force. It is used in aerospace actuators and deployable satellite structures. When deformed and then heated, it returns to its pre-programmed shape with remarkable force and precision. The 12-second full-body mode transition I designed for the Triton X-1 is aggressive. The engineering to achieve it — coordinating hundreds of SMA actuators across a vehicle body simultaneously — is genuinely hard. But it is an engineering problem, not a physics violation. The material exists. The actuation principle works. The challenge is integration and control, not fundamental science. "The Triton X-1 might exist in 40-50 years. The Aegis Prime needs 100+. One is an engineering problem. The other is still a physics problem." VTOL aircraft exist — the V-22 Osprey, the F-35B, a growing number of electric air taxis. Small submarines exist. High-performance electric vehicles exist. The Triton X-1 is not any of these individually; it is all of them in a single morphing chassis. That integration is enormously difficult. But it does not require physics we do not have. It requires engineering we have not yet done. The contrast between the two platforms I designed reveals something important: the mech is blocked by missing fundamental science, while the morphing vehicle is blocked by missing engineering. Those are very different kinds of impossible — one with a clear path forward, one still waiting for breakthroughs that may not come on schedule. Technology Status Today Estimated Arrival Shape Memory Alloys Commercial — Nitinol in medical/aerospace use Already here VTOL aircraft Commercial — V-22, F-35B, air taxis Already here Neural interface (basic) Human trials — Neuralink 2024 20-30 years (advanced) Tri-modal vehicle (Triton X-1) Engineering challenge, not physics 40-50 years Miniaturised fusion reactor No working prototype at any scale 50-70 years Mech-scale carbon nanotube frame Laboratory only, centimetre scale 30-50 years Full combat mech (Aegis Prime) Multiple unsolved physics problems 100+ years The technology readiness roadmap — from today to the Aegis Prime V. 

What I Actually Built The Research Paper That Became a World After all this research, I did something strange. Instead of writing up my findings as a conventional analysis, I formatted them as a classified military research document set in the year 2126 — complete with a fake journal header, a DOI number, performance specifications grounded in the real research, certification compliance tables, and a comparative analysis between the two platforms. Not to deceive anyone. To stress-test my own consistency. Formatting the research as a real document forced internal discipline that a loose essay would not have required. Every number had to make sense relative to every other number. 

The fusion reactor's 98,000 kW output had to be sufficient for the weapons loadout I specified. The sensor range had to be physically justifiable given the AI processing architecture. The structural weight had to be compatible with the actuator specifications. If anything was inconsistent, the document format exposed it immediately. "Writing fiction as if it were fact forces you to be internally consistent in ways that pure speculation does not require." The document — the Aegis Triton Research Article — is a technical comparative assessment of the RX-0A1 Aegis Prime and the Triton X-1. It reads like a real classified defence research paper because the underlying research was real. The platforms are fictional. The engineering logic is not.

 This approach has a name in serious speculative engineering circles: it is called a technology readiness analysis dressed in narrative. DARPA uses a version of it when commissioning speculative capability documents. Science fiction writers with physics backgrounds use it. The goal is always the same — to identify which imagined technologies are extrapolations of real science, and which are simply wishes. The Aegis Prime, it turns out, is mostly wishes. The Triton X-1 is mostly extrapolation. That distinction matters more than the fictional frame around either of them. VI. What This Research Actually Taught Me A Technology Roadmap Hidden Inside Fiction The mech is impossible today. But the research revealed something more useful than impossibility — it revealed a dependency map. 

The technologies required for the Aegis Prime are not equally distant. They have an order. They have prerequisites. Some are almost here. Some are nowhere near. Fusion power is the critical path for everything. Without it, the weapons cannot fire at sufficient energy, the mobility cannot be sustained, the AI cannot be powered at the required computational level. Every other technology in the design — the sensors, the structure, the neural interface, the weapons — can be engineered once energy is solved. None of them can compensate for its absence.

 The neural interface will arrive first. Within 20-30 years, brain-machine interfaces capable of complex machine control are plausible. This will change everything about how humans interact with systems — not just mechs, but vehicles, aircraft, factories, surgical robots, and computers. The mech may never exist, but the interface that would pilot it is coming regardless. The morphing vehicle is the sleeper technology. Shape memory alloys and hybrid propulsion are advancing without the fanfare of fusion or neural links. In 40-50 years, a vehicle that transitions between road, air, and water modes using smart materials and hybrid energy is not implausible. It will not look exactly like the Triton X-1. But the concept will exist. And the mech? The mech needs all of these, plus structural materials we cannot make, plus an actuator technology that does not exist at mech scale, plus an energy source that has no working prototype at any scale. It is not one breakthrough away. It is seven or eight, each dependent on the previous. "The mech is not one breakthrough away. It is seven or eight, each one depending on the last. That is what 100 years means." One hundred years is not a pessimistic estimate.

 Given what we know about how long materials science advances take — from laboratory discovery to commercial production — and given the current state of fusion research, 100 years is probably realistic and possibly optimistic. The Aegis Prime lives in 2126 for a reason. . Batman Started This The Batwing works as a fictional vehicle because the people who designed it — the artists, the film engineers, the prop designers — cared about internal consistency. Wayne Enterprises behaves like a real defence contractor. The technology has plausible physics. The vehicle has limitations that get exploited by the plot. It is grounded enough to feel real even when it is doing things that are not. That is the same discipline I tried to apply to the Aegis Prime and the Triton X-1.

 The question I kept asking was not 'what would be cool?' but 'what would actually work?' The research showed me where those two questions diverge — and the answer was fascinating. The cool things and the realistic things overlap more than you might expect at 40 years. They diverge dramatically at 100. The flying submarine car is in the overlap zone. The 18-metre nuclear-powered bipedal war machine is firmly in the science-fiction zone — not forever, but for longer than any of us will be alive to see. Someone has to start the research, though. And sometimes the most honest way to do research is to write the paper from the future and then work backwards to find out what it would take to get there. The mech is not coming for 100 years. But the research into why it cannot come yet? That was worth every hour. "Write the paper from the future. Work backwards to find out what it takes to get there." — END — Mystic Quill | Research & Writing by Selva Ganesh K | 2026

I documented the full technical research as a classified military research paper set in 2126. Read the full document below:

→ Open full document in new tab