BTC mining companies are collectively switching to AI data centers. Whose story is more likely to materialize?

Have you ever thought: why can a Bitcoin mining company suddenly sign a $4 billion contract with NVIDIA?

To put it bluntly, it's because these mining companies hold the most scarce resource in the AI era – electricity.

Last year, everyone was laughing at crypto miners for their "cyclical gambling," but this year, they've transformed into hot commodities for AI data centers. The reason is simple: training a Frontier Model doesn't require smarter engineers, but bigger power plants. AI data centers start at 500MW, 1GW, and building new power plants + substations + cooling systems takes at least 3 years in queue – while the mining companies' facilities are ready to go as soon as you turn them on.

But pivoting to AI isn't as simple as changing a sign. Two leaders, IREN and CIFR, are both doing this, but in completely different directions:

$IREN(IREN.US) ($4B market cap, current price $8.40 / US East 5/14 close): $2.6B cash, 5GW of power secured, AI Cloud quarterly revenue of $33.6M, up 94% QoQ – they are taking the "buy and operate our own cards" route. On 5/14, they just secured a $4B five-year contract with NVIDIA, including Blackwell GPU deployment. On the same day, they also completed a $3B convertible bond offering (with only a 1% coupon, maturing in 2033). This company's AI story is already being realized with real contracts and real revenue.

Counterpoint: Quarterly net loss of $48M (although $40M of that was a non-cash impairment loss on mining equipment), and the convertible bond cap call still has a floating loss of $3.7M. To put it bluntly, the story is good but they're still burning cash.

$Cipher Digital(CIFR.US) ($9.1B market cap, current price $2.29): BTC mining quarterly revenue of $48.4M (missed expectations) + EPS of -$0.28 (also missed) – looking at the mining business alone, it's inferior to IREN. But CIFR is taking a different path: "We don't operate AI compute ourselves, we only act as the 'power + facility' supplier for the big clients." They have signed three hyperscale customer leases, with a total Contracted Revenue of $1.4B, average annualized NOI of $87M, and a pipeline of 3.3GW to be energized between 2027-2030. The story is bigger, but it's all after 2027.

In other words –
IREN is "seeing AI revenue growth curve now, but heavily reliant on a single major customer, NVIDIA"
CIFR is "zero AI revenue now, but three hyperscale long-term contracts locked in until 2036"

Short-term momentum is better for CIFR – up 29.8% on 5/5, with a 10-day gain 4 times that of IREN. But realization is higher for IREN – major customer contracts are already generating revenue, not just PPT.

You shouldn't go all-in on one of these two. If you are bullish that "AI compute shortage will last until 2027," IREN is more suitable as a core holding; if you are willing to wait for the 2027 pipeline energization and look at a longer story, CIFR actually has higher odds. Neither is cheap (IREN PE 122.78, CIFR is losing money so no PE), and neither should have a heavy position. The worst thing is to trade them as "BTC concepts" – these two are already AI data center stocks, and the pricing logic has completely changed.