📰 Key Takeaways

Bitcoin mining companies are pouring big money into AI and high-performance computing (AI/HPC), but there’s still a massive gap between capital deployed and actual revenue. According to BlocksBridge Consulting’s latest Miner Weekly report, 15 public bitcoin miners and AI data center companies posted combined capex of $30.7 billion in the most recent 2026 reporting period, up 42.6% from full-year 2025’s $21.53 billion. Looking just at the 9 comparable miners, H1 2026 capex totaled $5.11 billion, while directly reported AI/HPC revenue over the same period came to just $341.2 million — a capex-to-revenue ratio of roughly 15 to 1. BlocksBridge’s capex estimates are based on cash purchases and allocations toward productive assets like hardware, real estate, and equipment, net of proceeds from asset sales and refunds. Despite the sizable gap, AI/HPC revenue growth is accelerating — the 9 miners posted combined Q2 revenue of $205.8 million, up 52% quarter-over-quarter, with Core Scientific, TeraWulf, and Bitdeer showing the most pronounced growth. BlocksBridge notes that while power contracts and land give miners a natural head start, turning those assets into “AI-ready” capacity still requires substations, buildings, cooling systems, and networking gear — and some business models even require GPUs — making the upfront investment bar extremely high. Meanwhile, bitcoin’s price jumped over 13% this week, climbing back above $72,000. Whether the warmer market sentiment eases pressure on miners still running large-scale mining operations remains to be seen. See the original article for full details.

💬 JudyAI Lab Take

Bitcoin miners are speeding up their pivot into AI/HPC, but the earnings numbers show this road is still very expensive to travel — and that gap itself deserves more attention from AI builders than the pivot slogans do.

BlocksBridge’s latest report lays out the reality: capex across 15 public miners hit $30.7 billion in 2026, up 42.6% year-over-year, but the 9 comparable miners’ AI/HPC revenue came to just $341.2 million — a capex-to-revenue ratio as high as 15 to 1. This reflects a trend that’s often underestimated: having “power and land” doesn’t automatically mean you’re “AI-ready.” Converting mining-era infrastructure into capacity that can actually handle AI/HPC workloads requires substations, cooling systems, networking gear, and sometimes even GPU procurement — every piece is a capital-intensive rebuild, not a simple repurposing. It’s also worth noting that despite the huge gap, these 9 miners’ Q2 revenue grew 52% quarter-over-quarter, showing demand is starting to catch up with the capital being poured in — the transition is expensive, but it’s not going nowhere.

For readers tracking AI infrastructure, it’s worth watching how the ratio between capex and actual revenue shifts in these “transition-stage” companies’ earnings — that tends to be a better signal of whether the pivot is really landing than the stock price is.

📅 Original Article Info

🔗 Further Reading