Recently, JPMorgan upgraded two Bitcoin mining companies. On Nov. 24, Cipher, or CIFR, and CleanSpark, or CLSK, were raised from neutral to overweight. Cipher’s price target was also lifted from $12 to $18, and its shares rose 4.2% in early trading after the news.

CleanSpark gained 4.6% in premarket trading, reaching $10.18.

Why did JPMorgan suddenly raise its ratings on mining companies?

Since late September, Cipher and IREN have signed contracts worth more than $19 billion, locking in 600 MW of critical computing power resources.

These two companies are no longer simply living off coins. They are quietly moving toward a hybrid model that combines crypto mining with high-performance computing, or HPC.

The pure crypto mining business has been getting harder for some time. After the 2024 halving, daily revenue per EH/s is expected to fall 45%, while gross profit is set to drop 53%. At the same time, the network hash rate has surged to 972 EH/s, making competition 50% more intense than before the halving and further squeezing miners’ margins.

That is where HPC has become a lifeline.

Miners already have cheap power and ready-built sites, exactly the resources in short supply for AI computing power. Converting existing facilities for HPC can cut data center deployment time by three-quarters, with relatively low incremental costs.

For these two companies, the transition path is now very clear.

Cipher has already secured partnerships, signing 600 MW of computing power contracts with AWS and Fluidstack. Its share price has recently fallen 45%, a sizable pullback. JPMorgan sees that as a rare chance to buy the dip. CleanSpark, meanwhile, has chosen to expand in Texas, where its new site could add 200 MW of HPC capacity.

The market’s choice is clear, and capital flows say it all. Among the 13 mining companies tracked by JPMorgan, those with exposure to HPC, including WULF, IREN and Cipher, have far outperformed Bitcoin itself.

Even the valuation framework has changed. HPC hosting services are now valued at $8 million to $17 million in equity value per MW. For integrated cloud services, that figure can rise to $19 million.

But not every mining company will get a share of the gains. IREN’s price target was raised to $39, a positive move, but its rating remains underweight.

MARA and RIOT, by contrast, have been less fortunate. Because Bitcoin prices remain weak, their price targets were cut instead. The key factor behind this divergence is whether their HPC capacity is clear and reliable enough.

JPMorgan made a forecast: by the end of 2026, the miners under its coverage could have a combined 1.7 GW of HPC computing power.

IREN and Cipher are expected to lead. Behind that trend is institutional capital’s search for more stable cash flow, since long-term HPC contracts can help insulate companies from swings in Bitcoin prices.

But is this transition really free of risks? These miners will have to compete directly with cloud service giants. How long can their modest advantage in power costs last?

That is the question. And as more players crowd into the market, could HPC fall into the same pattern of overcompetition that once defined the crypto mining industry?

Mining companies are no longer surviving solely on moves in Bitcoin prices. They are beginning to build computing power infrastructure. That role shift has only just begun. In this contest, real victory will not be measured by how many contracts a company signs, but by who can fully break free from the cycle.