GLXY is up 12% on the news. That’s not a signal—it’s a rounding error. The market just woke up to the fact that Galaxy Digital now controls 1.63 gigawatts of approved power capacity in Texas. Compare that to CoreWeave’s entire portfolio. Or Google’s. The difference? Galaxy got it for the price of a Bitcoin miner’s land.
I don’t build models on hope. I build them on signed leases. And this lease—15 years with CoreWeave—is the most important piece of financial engineering in crypto infrastructure since Bitmain’s ASIC monopoly. Let me walk you through why.
Context: The Old Galaxy vs. The New Galaxy
Galaxy Digital started as a crypto merchant bank. Mike Novogratz’s brainchild. By 2020, it was a mix of trading, asset management, and mining. The mining arm? A cost center. Helios, its flagship site in West Texas, was built to mine Bitcoin. But by 2023, Bitcoin mining margins were shrinking. Halving loomed. Energy costs crept up. The playbook for every miner: either get efficient or get creative.
Galaxy got creative. It quietly acquired two more land parcels—one near McGregor (74MW initial, 2028 energization) and another unnamed plot—while simultaneously pushing Helios through the ERCOT approval gate. Result: 1.63GW of fully approved capacity. Not hypothetical. Not conditional. Approved.
Then it signed CoreWeave as anchor tenant. CoreWeave is not a retail cloud provider. It’s the hyperscaler that powers some of the most demanding AI workloads on the planet. If you’ve ever run a stable diffusion model on a rented cluster, you’ve touched CoreWeave’s infrastructure. The 15-year term? Locked in.
The Core: Order Flow Analysis from an Infrastructure Lens
Let’s kill the narrative that Galaxy is still a crypto company. That’s like calling Amazon a bookstore in 2015. The real asset is electricity—specifically, the right to draw 1.63GW from the ERCOT grid. How do you value that?
Step one: What does 1.63GW of hyperscale data center capacity earn? Industry benchmarks: A well-designed AI data center can generate $25–$40 per kW per month in EBITDA. At 1.63GW, that’s $40 million to $65 million per month. Call it $500 million annually in EBITDA from Helios alone. The new land adds another 74MW initially (McGregor), with potential to scale to 500MW over time. Even conservative math puts Galaxy’s infrastructure EBITDA above $600 million within three years.
Step two: What multiple does the market assign? Pure-play data center REITs trade at 15–20x EBITDA. Bitcoin miners trade at 5–8x. Galaxy currently trades at about 8x, because the market still sees it as a miner. The moment the analyst community re-rates it as an infrastructure play, the stock should double, minimum.
But here’s where my personal experience kicks in. Back in 2017, I audited an ICO smart contract that promised AI-driven arbitrage. The code had a reentrancy bug that would have drained $4 million. I flagged it. The team “fixed” it with a band-aid. A month later, they got hacked anyway. What I learned: technical integrity eats social capital for breakfast. Galaxy’s move to lock down 1.63GW with a 15-year tenant is the equivalent of auditing the smart contract twice. They created a structural moat that can’t be copied in a month.
Contrarian: What Retail Misses About the Risk
The market isn’t dumb—it just has a short attention span. Right now, retail sees Galaxy as a Bitcoin proxy. “If BTC goes to $200k, GLXY moons. If BTC crashes, GLXY crashes.” That’s the lazy thesis. But look at the balance sheet: the Helios lease alone will throw off enough cash to cover operating costs even if Bitcoin mining goes to zero. Galaxy can afford to mothball its ASICs and still collect rent.
The real risk isn’t Bitcoin volatility. It’s single-tenant concentration. CoreWeave is a fantastic partner today, but if their AI workload growth stalls—or if they decide to build their own Texas campus (they won’t, because ERCOT approvals take years)—Galaxy could face capacity overhang. But that’s a 2028 problem, not a 2024 one.
Another blind spot: ERCOT grid reliability. I lived through the 2021 Texas freeze. The grid failed. People died. Today, ERCOT has improved, but adding 1.63GW of continuous load on a system designed for peak summer AC is a bet on infrastructure upgrades. If the grid can’t handle it, Galaxy needs backup generators—likely natural gas turbines. That’s an additional capital expense not in the current models.
Still, the smart money is piling in. Institutions like D1 Capital and Tiger Global already hold GLXY. They see what I see: a company that shifted from a volatile revenue stream (mining rewards) to an annuity (AI computing lease). The market doesn’t care about your thesis until the P&L shows it. This one will, soon.
Takeaway: Actionable Price Levels
I don’t give price targets. I give levels. GLXY currently sits at around $30 (post-news pop). If it breaks $35 on volume, that signals the re-rating is real. If it backtests $25, that’s a chance to add. The real catalyst? Next earnings call. Watch the split between mining revenue and AI/hosting revenue. If the latter crosses 50%, the street will have no choice but to reclassify.
Risk management is the only alpha that lasts. So here’s your risk rule: close half your position if GLXY drops below $22. That’s the level where the AI narrative breaks and it becomes just another miner.
Final thought: When I was trapped in the Terra collapse in 2022, I survived because I refused to hold concentrated positions. Galaxy is building the opposite: a concentrated bet on AI demand. But the difference is they locked in a tenant before the buildout. That’s discipline. That’s battle-tested.
Charts don’t lie, but narratives do. The narrative that Galaxy is a crypto miner is dying. The chart is already showing the transition. Are you fast enough to read it?