The Hidden Crypto Power Plan

This chapter follows the thread from that appetite to the compute used for AI. We’ll track how crypto mining became a public, incentivized way to acquire massive compute - first for hashing, then, by a change in how the machines were bought, sold, and repurposed, for training. And we’ll do it with skepticism switched on, because the story is too convenient to be purely accidental. How did a currency mechanism end up paying for the muscle of AI training - while the public watched the wrong scoreboard?

The Compute-for-Cash Loop: when mining became a compute marketplace

Crypto mining is often explained as a math contest: machines race to solve a cryptographic puzzle, and whoever wins gets rewarded. But that misses the economic engine humming underneath. Mining turns electricity and hardware into a steady stream of revenue - at least in theory - because the protocol pays out block rewards and, in many networks, miners also capture transaction fees. That means the “business” of mining is not just the cryptography; it’s the conversion of real-world inputs - power, cooling, maintenance, replacement parts - into an income stream that is visible, continuous, and publicly auditable in the form of blocks and difficulty.

That visibility mattered. As mining scaled, it produced something that AI buyers later wanted: large numbers of GPUs and other compute-capable machines already deployed somewhere, already running, already maintained by entire communities. The Compute-for-Cash Loop is the phrase that captures the loop’s real shape: you buy or assemble compute hardware, you run it for long enough to earn the protocol’s rewards, and you treat that revenue as the justification for the next hardware purchase. It’s not a subtle system. It’s a recurring financial logic that rewards anyone willing to source equipment, keep it cool, and accept the risk that the rules of profitability can change.

The early days of Bitcoin mining used specialized gear - ASICs - more than GPUs. That’s the part most people remember, and it’s true. But the broader mining economy did not stay confined to ASIC-only networks. As the industry explored Proof-of-Work designs and as GPUs became a general-purpose workhorse, the hardware ecosystem expanded. GPUs are not just for graphics; they’re parallel processors, and they can be used for many computational workloads beyond mining. That general-purpose character is where the later connection to AI becomes plausible. Once you have fleets of GPUs running under an incentive structure, you also have fleets of GPUs that can be repurposed when demand shifts.

Here’s the counterintuitive turn, and it’s worth stating plainly: the “mining for currency” story trained the compute supply chain more than it trained the miners. The incentives didn’t just pay individuals; they shaped where hardware went, how it was repaired, and how quickly it could be deployed. The same economic machinery that made it rational to keep hundreds of machines online also made it rational to build operational know-how around keeping them online. When AI later needed enormous amounts of parallel compute, that know-how wasn’t theoretical. It was sitting in warehouses and repair shops and shipping lanes.

From blocks to GPUs: how mining “bought” compute time in public

To understand how crypto mining became a public, incentivized way to buy massive compute for training models, you have to look at how compute becomes available. Compute doesn’t arrive as a cloud invoice in every case; sometimes it arrives as a physical resource someone else already paid for. Mining did that at scale, and it did it in a way that was easy for outsiders to observe: rigs running continuously, difficulty adjusting, profitability changing, and equipment cycling through hands.

Rina’s world shows the human mechanics behind those observations. She repairs cards that have been used hard, often in environments optimized for throughput rather than longevity. Mining rigs run at high utilization for long stretches. That stresses power delivery components, thermal interfaces, and fans. When miners switch hardware - either because a new generation offers better efficiency or because profitability falls - older rigs don’t vanish. They get resold, cannibalized for parts, or shipped to places like her shop for component-level repairs. In other words, mining created a hardware churn that kept repair services busy and kept older compute resources from simply turning into e-waste overnight.

That churn is one reason the compute supply for AI didn’t have to be invented from scratch. AI training is compute-hungry in a way that quickly overwhelms small, boutique hardware procurement. Training runs are sensitive to availability: if the cards aren’t there, the work stalls. Mining had already established a pattern of buying and operating large numbers of cards. Even when those cards weren’t “for AI,” they were for intensive computation, and they were part of a broader market for compute hardware.

Now add a second layer: the public nature of mining profitability. Mining returns are tied to protocol parameters like difficulty and to market conditions like token prices. That means the incentives to run compute can surge or fade quickly. When the incentives fade for a particular coin, rigs can be sold off. When incentives surge, rigs are built faster. That creates a rhythm in the hardware market - an availability pulse - that AI customers later tried to exploit through leasing, resale, and eventually cloud offerings that could scale on demand.

A single-sentence fact that helps frame the pivot: many of today’s AI training systems depend on GPUs, and GPUs have long been the default compute hardware for a wide slice of high-throughput workloads. Once mining normalized the idea of running GPUs continuously to earn money, GPU capacity became a commodity with an established operational history. Even if the original workload - hashing - looks different from neural network training, the infrastructure pattern is similar: keep silicon running, keep it cool, keep it fed with power, and keep the uptime high enough that the economics work.

The surprise inside the surprise: why “wasted” mining became usable compute

The most surprising connection is that the same system criticized for “wasting electricity” also helped build the practical muscle for later AI compute. Mining is often portrayed as pure consumption - energy in, nothing but heat out. But what mining actually produces is not heat; it produces a service: the ability to perform a particular kind of compute continuously at scale. The “waste” critique focuses on the energy, but the operational reality is that miners learned how to operate compute at industrial scale - where the cost and logistics of power, cooling, and hardware replacement can be managed.

This matters because AI training has its own version of the same constraint: power and cooling are not side issues; they are the difference between an experiment and a stalled month. You can buy GPUs, but you still have to house them. You still have to keep them from overheating. Mining communities, by necessity, built that playbook around running machines relentlessly. So even when the original goal was cryptographic proof, the operational capability stayed behind.

Why does this reframe the story? Because it shifts the question from “Who decided to help AI?” to “Who already had the infrastructure to make compute happen?” The Compute-for-Cash Loop doesn’t care what the compute is “for” at the protocol level; it cares that the machine runs and earns returns. That makes it possible for compute to transition between uses when markets change. Mining didn’t just finance hardware; it financed the routines and ecosystems that make hardware usable at scale.

Rina can tell you what this looks like without needing any token chart. When miners flood the market with hardware upgrades - or when they abandon certain rigs - the repair ecosystem absorbs the shock. Cards that were once expensive and hard to source become easier to find, because someone has already paid for them and is now exiting. That exit creates a window where compute becomes available at different price points, often through resale and refurbishment. AI buyers, especially those who can’t afford brand-new capacity, can benefit from that window. The compute “paid for by mining” isn’t always a direct transfer of funds into a research lab; it’s a conversion of capital into accessible hardware over time.

The human story behind public incentives: Rina and the hardware cycle

Rina’s shop has a rhythm that mirrors the market, even if she doesn’t live inside it. When mining demand is high, she sees more rigs arrive - sometimes whole systems, sometimes individual cards pulled from failures. When demand cools, the mix changes: she gets more repairs on older models, and the hardware she fixes often comes from resellers trying to salvage value from equipment that’s no longer profitable to run. Either way, the shop is part of the same chain of incentives that began with protocol rewards and ended up with physical compute.

There’s a specific detail that makes the connection feel real: the way she diagnoses faults. Mining pushes components into thermal stress and power cycling. That means her work often turns into a kind of forensic accounting - what failed, what overheated, what degraded after weeks of continuous load. She’s not repairing “AI” hardware; she’s repairing compute that was used because it could earn. But once those repaired cards circulate again, they don’t stay locked in one purpose. GPUs that were kept alive for mining can be used for other workloads, including machine learning, because the physical capability is the same: parallel computation.

In places like hers, the public incentive structure of mining plays out as private maintenance labor. That’s the part outsiders miss when they talk about crypto like it’s only a website or a price chart. Mining is embodied in hardware, and hardware lives or dies based on people who can keep it functioning. The incentive doesn’t just attract capital; it attracts technicians, parts suppliers, and informal networks of knowledge about what fails under sustained load.

One more concrete link: mining scale has always been tied to where power is cheap enough to justify the risk. That geographical reality shaped the hardware ecosystem. Many mining operations clustered near favorable power conditions, and that distribution created regional markets for used equipment, repair, and resale. When AI demand surged, those regions didn’t magically become research hubs overnight. But the compute capacity, the refurb pipeline, and the “keep it running” culture were already there. Rina’s shop is not a data center, but it sits downstream from that same geography-driven supply chain.

So the story of crypto’s compute role isn’t a story of clever philanthropy. It’s a story of incentives that built capacity, then left capacity behind when incentives changed. The public watched for tokens. The hardware kept moving.

What this tells us about AI, money, and the compute that follows

If you step back, the compute-for-cash loop points to a pattern that shows up across tech history: when a system pays for utilization, it ends up paying for infrastructure - even if no one designed it for the purpose you care about later. Mining was a machine for converting electricity and hardware into money. AI training later became a machine for converting money and hardware into capability. Those two machines met through the uncomfortable middle: the cost of running hardware at scale.

And that’s the real conspiracy-leaning thread worth holding onto - not that someone planned everything with a grand diagram, but that incentives are powerful enough to create outcomes that look planned from the outside. Markets don’t need a master plan to route capital toward compute; they just need a reason for compute to exist. Crypto mining provided that reason, publicly and continuously, with enough variability to keep hardware cycling through the world.

The broader lesson is unsettling: the things we treat as “innovation” often ride on top of logistics and maintenance that were never meant to serve the new thing. The compute powering AI doesn’t just come from breakthroughs; it comes from the unglamorous work of keeping machines alive, and from the economic systems that decide which machines get to be built in the first place. When you see mining as a compute marketplace rather than a currency ritual, the AI story starts to look less like a straight line and more like a scavenger hunt through supply chains.

If the most important AI resources were financed by incentives that had nothing to do with AI, then what else have we been mislabeling - because we watched the wrong part of the machine?

End of chapter one. 4 more chapters in the full book.