Bitcoin reaching $100,000 per unit has me thinking: What is the value of computation? Apologies in advance - this post may get a bit abstract.
Bitcoin offers a clear, quantifiable example of converting energy into computational value. Whether an investment or a product, Bitcoin's securtity structure - Bitcoinmining - may tell us about the growing demand for power-hungry datacenters.
- Bitcoin as a Model for Computational Value -
The Bitcoin network’s reward system for securing the blockchain is straightforward: across all miners 3.125 BTC is awarded every 10 minutes, along with some smaller transaction fees. At current prices, that translates to about $16 billion annually to secure the network in 2025.
This cost is borne by Bitcoin users, through inflation in Bitcoin’s supply and direct fees. In return, users gain the security and reliability of the blockchain. Here, the relationship between computational effort and economic value is clear, transparent, and precise.
- More Complex Cases -
Computation in other domains - AI/ML applications, streaming services, and VR/AR simulation - lacks Bitcoin’s reward simplicity. These services still deliver value, but their rewards are shaped by evolving market dynamics rather than fixed protocols.
GoldmanSachs recently estimated that tech firms, corporations, and utilities will spend $1 trillion in capital expenditures to support AI in the coming years. OpenAI, for instance, now charges $200/month for its premium ChatGPT tier - a $180 premium over the next plan. Clearly, computation holds value when companies are willing to invest at this scale and customers are willing to pay.
Yet, unlike Bitcoin, where rewards are predefined, the value of these computational services is still being negotiated by the market. How the returns from $1 trillion divided among stakeholders - data center operators, hardware manufacturers, AI model developers - remains to be seen.
- Commodities & Competition -
Returning to Bitcoin mining, the lowest-cost producer wins. Miners compete based on access to cheap electricity (often renewable), efficient hardware, and the low-cost financing. No need for unique selling points, product marketing, or tiered pricing. With fixed rewards, Bitcoin security is purely a cost-driven race.
By contrast, companies in AI/ML and other computational services market differentiated value in their products. They emphasize brand identity, innovation, and unique features. But how different are these services in practice? Can you tell the difference between ChatGPT (OpenAI), Bard (Google), Alexa (Amazon), Siri (Apple), and Co-Pilot (Microsoft)?
I’ll be watching whether computation becomes a commodity. How will value be ultimately distributed between stakeholders - data center operators, hardware makers, model developers?
Comments