MARA's $13 Cut and $30 Upgrade: Two Analysts, Two Different Companies

When JPMorgan cut the price target on Marathon Digital (MARA) late last November, pulling it from $20 to $13 on the falling value of the company's bitcoin pile, the headline read like a verdict. The very same day, a different bank did the opposite: Compass Point upgraded the stock to Buy and held a $30 target, arguing the sell-off had overshot the fundamentals.

Both can be true at once, because they are not really rating the same company.

The split is the signal

Look at where the Street lands today and the disagreement is hard to miss. TipRanks counts six analysts at Buy, three at Hold, one at Sell — a "Moderate Buy" consensus that sounds neutral but hides a roughly five-fold range between the lowest and highest targets. Morgan Stanley sits at Underweight and in August cut its target to $6 from $5.50, a view the company's investor relations publicly pushed back on earlier this year.

That is not a pack of analysts politely quibbling over a multiple. It is two incompatible models of what this business is worth, because the company itself is mid-transformation.

MARA is one of the largest bitcoin miners in the world and the second-largest corporate holder of bitcoin, with about 53,250 coins worth roughly $4.6 billion at the time of the JPMorgan note. That half of the business is priced off the price of bitcoin and the network hash rate — the computational arms race that determines how hard it is to earn each coin. When JPMorgan cut its target, its stated reason was the collapse in bitcoin's price shrinking the value of that treasury, and it simultaneously raised targets on other miners like Cipher and IREN.

The other half of MARA is the story of 2026: a pivot from pure mining toward AI and high-performance-computing data centers and energy infrastructure, including a joint venture with Starwood to convert power-rich sites. Compass Point's upgrade leaned on exactly this — calling MARA's "nascent AI business" pure upside and arguing that on its mining fundamentals alone the stock was undervalued.

Why the factor stack goes quiet here

Put the two views through a systematic valuation screen and something important happens: the machinery sputters. MARA does not earn money right now. It posted a roughly $610 million net loss in the second quarter, tied largely to bitcoin's fair-value swings and spending on HPC and data-center infrastructure, after a $1.71 billion loss the prior quarter. Negative earnings make a price-to-earnings ratio empty, so the whole sector gets priced on sales — and those multiples are enormous.

That is the point the "is it cheap?" question misses. No stock means anything in isolation; cheapness is only meaningful against your sector. Marathon's listed mining peers trade at roughly 38 times sales (Hut 8), 37 times (Cipher), and 24 times (IREN). Nothing in this sector is cheap by that yardstick, and the companies sitting at those multiples are all losing money too. Any "cheap" reading of MARA has to be argued from what the AI and energy businesses might someday earn, not from what it earns today.

Regime change calls for structure, not conviction

There is a reason the quantitative factors feel unreliable here, and it is not a data glitch. A factor model quietly assumes the company you are comparing stays the same kind of company from quarter to quarter — that a strong scoring miner keeps being a miner, so its history means something. MARA breaks that assumption by design. When a regime change makes historical patterns untrustworthy, the honest answer is to say so rather than bluff a confident grade.

The tape tells the same story. MARA trades around $12, up about 5% on the day, sitting just above its 200-day average with a 14-day RSI near 56 — momentum that looks constructive on the surface. But its average daily range of roughly $0.90 is near 7.5 percent of the share price. On a stock this size, in a sector this binary, the daily swing is the market admitting it cannot decide which MARA it is watching.

So a price target — $13, $30, or $6 — is really a scenario in disguise. JPMorgan's cut prices the miner and its treasury; Compass Point's upgrade prices the future data-center company; Morgan Stanley's underweight prices the execution risk between the two. The useful question for an investor is not which bank is right but which business you are being paid to own. Until the AI-and-energy transition produces earnings with a number attached, the analyst range is not a disagreement to resolve — it is the honest description of a company whose value is a bet on what it will become. In uncertainty like that, the answer is more structure, not a louder prediction.