Rosenblatt reiterates Buy rating on MARA Holdings stock at $15 By Investing.com

Rosenblatt reiterated a Buy on MARA Holdings with a $15 target, implying about 12% upside from the current $13.39 share price, but cut adjusted EBITDA estimates despite better mining profitability. The company beat first-quarter fiscal 2026 expectations on fleet upgrades and higher hashrate, yet higher expenses and $3.6 billion of debt remain key concerns. Management said initial HPC contracts are expected in the second half of 2026, while an earnings miss elsewhere in the article underscored volatility in the stock.

Analysis

The key signal here is not the noise around one miner’s print, but the market’s growing sensitivity to anything that can compress semiconductor multiples under the label of policy risk. That matters most for NVDA because its valuation now embeds uninterrupted AI capex growth; even a low-probability tax/regulatory headline can trigger de-grossing in crowded momentum books before fundamentals change. In other words, the move is less about near-term earnings risk and more about how fragile positioning is when investors are already paying for perfect execution.

For MARA, the core tension is between improving unit economics and a balance sheet that forces the market to discount every operational win. The strategic pivot toward HPC is optionality, but it is not yet earnings visibility; until contracted revenue appears, the equity remains a levered call option on Bitcoin with a costly overhead structure. That makes the next 1-2 quarters critical: if expense discipline does not show up in cash burn, the market will likely treat any rally as financing-driven rather than fundamental.

The second-order winner from this kind of tape is quality infrastructure and diversified compute, not pure-play miners or the highest-multiple chip names. If AI capex is questioned even temporarily, buyers tend to rotate toward businesses with clearer customer contracts, lower duration risk, and less policy beta. The contrarian view is that the selloff may be overdone in names like NVDA if the market is conflating headline volatility with actual demand destruction; the downside can reverse quickly once the narrative shifts back to earnings revisions and supply constraints.

Key Decisions for Investors

• Maintain a tactical hedge: short NVDA vs. long SOXX or SMH call spreads for 2-6 weeks; use the hedge to express that policy/headline risk can compress multiple expansion without needing a fundamental demand break.
• For MARA, avoid outright longs until Q2/Q3 cash burn trends improve; if expressing upside, use small-sized call spreads into the expected HPC contract window in 2H26, since the equity is still more financing-sensitive than earnings-sensitive.
• Pair trade idea: long diversified AI infrastructure beneficiaries with recurring revenue and short high-beta miners; the goal is to capture the spread between contract-backed compute and balance-sheet-levered optionality over the next 3-6 months.
• If NVDA down-move continues on no incremental evidence of demand deterioration, fade it with tight stops via out-of-the-money put selling or call spreads; implied panic can overshoot by 10-15% before earnings revisions actually roll over.