MARA · MARA Holdings, Inc.

Current Thesis

The Bitcoin-miner-into-AI/HPC pivot is the story everyone wants to own, but the tape just rejected it. The May rally from $12.44 (2026-05-15) to $15.29 (2026-06-01) a +23% "diversification" bounce has fully round-tripped. MARA closed ~$12.32 on 2026-06-05, down -11.67% on the day, printing back at the May swing low after the 2026-06-02 Q1 double-miss. Underneath it, the actual driver keeps breaking: BTC fell below $62K on 2026-06-05 (~$1.5B long liquidations), trading ~$61.5K–$62.5K on 2026-06-07, versus $63.7K only three days earlier. Q1 revenue was $174.6M (-18% YoY from $213.9M), EPS -$3.31 against ~-$1.51 expected, free cash flow ≈ -$327.5M, dragged by a ~$1.0B mark-to-market writedown on the BTC stack. The Starwood 1GW→2.5GW AI/HPC leg is still a press release. There is no accelerating narrative to ride this is a falling knife against a declining BTC tape, with optionality that has not yet converted to revenue.

Bull Case

• Dated AI/HPC optionality: Starwood Capital partnership (announced 2026-02-26, +17% on the print) targets ~1GW near-term scaling beyond 2.5GW. A single site converting to contracted AI compute revenue re-rates the name from miner to digital-infrastructure.
• Owned power: the Long Ridge Energy acquisition plus the Starwood campus give MARA owned megawatts the binding constraint for AI buildout that pure-play GPU clouds rent rather than own.
• Operational scale intact: record hashrate 72.2 EH/s (+33% YoY, Q1 2026) shows the mining engine still executes even as BTC price guts the P&L.
• A constructive Street tail exists: BTIG Buy $27 (2026-06-01); the blended average target sits ~$17.78, well above spot, framing the name as a battleground rather than a write-off.
• Product flow: GraniteShares launched MARA income ETFs (2026-06-01), keeping structural demand on the ticker.

Bear Case

• The core is shrinking: revenue -18% YoY, a -$3.31 EPS loss more than double the prior year's -$1.55, and ~-$327.5M free cash flow (Q1 2026). A decelerating top line cannot be momentum-traded.
• BTC is the real position and it's falling: below $62K on 2026-06-05 with $1.1B+ of 24h liquidations; MARA holds 35,303 BTC, so each leg lower compounds the writedown that already cost ~$1.0B in one quarter.
• The pivot failed its first validation window: Q1 was billed as the first financial proof of the AI transformation and showed no AI revenue. Promise, not print.
• Analysts cut into the pivot: Morgan Stanley $7 Underweight (cut from $8.50, 2026-06-02), Bernstein $23→$17 (2026-06-03), Cantor $10. A $7–$27 spread means nobody has a thesis with conviction.
• The bounce was counter-trend hope into a binary: price ran +23% into the Q1 print on diversification optimism, then gave the entire move back within three sessions.

Setup & Price Structure

• Last ~$12.32 (2026-06-05 close, -11.67%), intraday low in the low-$12.30s. The whole $12.44→$15.29 May bounce is gone; price sits right on the 2026-05-15 swing low.
• The higher-low sequence that defined the May rally broke when price round-tripped the entire move there is no intact bullish structure left to lean on.
• Price has lost the 20-EMA (≈ $13.50 area); the 20-EMA now sits overhead as resistance rather than support, flipping the mean-reversion bias to the downside.
• A weekly close that holds under ~$12.30 opens the prior leg lower toward the sub-$11 base. Strength is not the setup here the macro driver is carving lower lows and the equity is tracking it down.
• Liquid $12–$15 name, but with no clean breakout and momentum pointed down, this is chop-with-a-downward-drift, not a trend entry.

Catalyst Calendar (next 30 days)

• ~2026-07-03 (est.) monthly June operational update (BTC mined, hashrate, BTC held). Operational color, not a re-rate event; a soft BTC-production figure into a falling BTC price reinforces the bear read.
• BTC spot, daily the real catalyst. A weekly close below $60K is the line that confirms continued miner-economics bleed; reclaim of $72K weekly is the bull re-arm.
• No earnings in window Q1 printed 2026-06-02; Q2 lands ~mid-August 2026, outside 30 days.
• Unscheduled any Starwood site go-live or signed AI/HPC tenant announcement. This is the one event that converts the pivot from promise to fact; it is not on a calendar, so it must be watched for, not anticipated.

What Would Change Our Mind

• A contracted AI/HPC revenue announcement a named tenant, committed MW, and annualized $ figure from a Starwood-converted site. That single print reframes the multiple from BTC-beta to neocloud/digital-infra.
• BTC reclaiming >$72K on a weekly close, which removes the fair-value writedown overhang on the 35,303-BTC treasury and revives miner economics.
• MARA reclaiming $15.30 on expanding volume with BTC stabilizing above $64K restoring the broken May structure and a real higher low.
• Absent those, every rally is suspect and treated as a counter-trend bounce inside a downtrend.

Correlation Notes

• MARA trades as a leveraged BTC proxy it correlates to Bitcoin, not the AI capex cycle, until AI/HPC revenue actually prints. Position sizing across the book should treat it as crypto-beta, not an AI name.
• Do not double-count against MSTR, RIOT, CLSK, IREN all share the same BTC sensitivity; concentrated exposure across them is one bet, not five.
• The 35,303-BTC balance sheet means quarterly GAAP results swing with mark-to-market crypto moves (~$1.0B writedown in Q1 2026), decoupling reported earnings from operational performance.
• Diverges from true GPU/neocloud names (CRWV, NBIS, IREN's compute leg) that trade on AI demand rather than crypto price MARA only joins that cohort if and when a Starwood site contracts compute revenue.

Notes

This is a watch-and-wait, not an entry. The setup the playbook exists to catch an accelerating, cluster-confirmed narrative is absent: the BTC driver is in active decline and the AI replacement thesis is unvalidated. The name becomes interesting only on a contracted AI/HPC announcement or a decisive BTC reclaim, both of which are observable and dated-or-flagged above.

Correlation Notes (addendum)

Crypto-equity cohort risk-off is synchronized: the 2026-06-05 BTC break below $62K hit the entire miner complex, so a MARA breakdown is unlikely to be idiosyncratic it is the group trading with its underlying commodity.

Current Thesis (one-line for dashboard)

Diversification bounce round-tripped; ~$12.32 close (June 5, -11.7%) on a Q1 double-miss with BTC under $62K no accelerating leg, falling knife not a setup.