Nvidia was worth more than all of crypto

Nvidia had recently become the most valuable company on Earth, per the report: $3.41 trillion in market cap, against a $2.41 trillion crypto market at the time.

Nvidia didn't get there by accident.

Over the four quarters before the report was written, it pulled in about $80 billion in revenue. $34.5 billion of that, 43% of the total, came from AI chips alone.

The report's own napkin math puts roughly $1.3 trillion of that market cap on AI chips specifically.

Nvidia isn't the whole AI market, and that 43% AI-chip share means 57% of what it sells is something else entirely.

It's also poured billions into R&D on chips built for AI work, the H100 "Hopper" among them.

The market cap comparison is the headline, but most of what follows covers decentralized GPU networks, DePIN, the alternative to renting compute from centralized providers.

Five reasons builders skip the centralized GPU providers

So what's wrong with the centralized options?

Five problems come up again and again:

• Confusing, expensive pricing, especially on-demand. Hidden costs on top, the way Nvidia's own DGX Cloud works.
• Availability gaps. Not every machine type is around when you need it, which makes scaling for heavy compute demand harder.
• Scaling takes massive upfront cost.
• Integration limits. GPUs locked to specific VMs, awkward to plug into other cloud services or internal systems.
• Thin model selection from providers that specialize narrowly, leaving little room when you need flexibility.

On the decentralized side, five networks stood out:

• Render Network: 4,367 GPUs connected, over 82,000 TFLOPS.
• Akash Network, as of February 2024: 150 GPUs, nearly 100 of them A100s, plus RTX 3000 and 4000 cards, useful for AI/ML training and consumer-grade rendering.
• Gensyn.ai: a decentralized GPU network built for AI applications, its own trustless layer-1 protocol.
• io.net: built on Solana for consumer and enterprise AI/ML, verifies connected GPU providers through Proof of Time Lockup.
• Aethir: decentralized GPU cloud for AI, machine learning, and cloud gaming.

AI and crypto are interconnected narratives, the report argues, like guava paste and cheese, a classic Brazilian pairing.

They need each other, on its reading. Be bullish on one, it says, and you should be bullish on the other.

Four places show the overlap already happening: AI chatbots onboarding people into crypto wallets, crypto rewards funding open-source AI development, AI catching fraud while crypto secures the transfer, and AI reading market trends to sharpen DeFi strategy.

AI and robotics are fundamentally deflationary, the report argues, a real challenge for monetary systems built on debt.

As automation increases, wages could fall, making debts harder to pay. Left unaddressed, that could tip into a systemic financial crisis.

This was written before the DePIN GPU space had two more years to run. What's changed about it since, from where you're sitting?