Most people see a rebrand to PowerCompute and think: "Another miner going AI. Bullish."
The data says otherwise.
LM Funding (NASDAQ: LMFA) owns 26 megawatts of power capacity. That’s enough to run roughly 8,000 ASIC miners. Or, if fully converted, about 2,000 NVIDIA H100 GPUs. CoreWeave operates hundreds of megawatts. Lambda Labs has tens of thousands of GPUs.
Let that sink in.
The market is pricing this as a transformation. I see a distressed miner grasping at the hottest narrative to avoid delisting.
Context: The Halving Hangover
Bitcoin’s April 2024 halving cut miner block rewards in half. Small operators like LM Funding—with a market cap barely above $20 million—were squeezed. Their existing revenue stream (block rewards plus transaction fees) became insufficient to cover rising energy costs and ASIC depreciation.
Rebranding to PowerCompute and shifting stock ticker to POWR is a textbook survival tactic. The company states it will "leverage its existing 26 MW of owned power infrastructure to expand into AI infrastructure." It will retain its Bitcoin holdings as a balance-sheet asset.
Sounds strategic. Feels desperate.
I’ve audited over 12,000 on-chain transactions during DeFi Summer 2020. I learned that narratives without execution are just noise. This is noise with a stock ticker.
Core: The On-Chain Evidence Chain (Off-Chain, Actually)
Let’s break down the numbers.
Power Capacity
- 26 MW total (two facilities).
- For Bitcoin mining, that’s ~8,000 S21 Pros generating ~0.8 EH/s.
- For AI, 26 MW can support roughly 2,000-2,500 H100 GPUs (at ~10.5 kW per GPU server node).
Revenue Potential (AI)
- H100 rental rates: ~$2-3 per GPU-hour on the spot market; long-term contracts at $1.50-2.00.
- Fully utilized 2,000 GPUs at $2/hour = $96,000/day = ~$35M annualized gross revenue before power, cooling, and overhead.
- Reality: they won’t fill 100% immediately. Assume 50% utilization: $17.5M.
Costs
- GPU procurement: 2,000 H100s at $30K each = $60M capital expenditure. LM Funding’s current cash and Bitcoin holdings? Probably under $10M. They need debt or equity.
- Power: At $0.05/kWh, 26 MW full load = $1.14M/month. Plus cooling, networking, staff.
Competition
- CoreWeave: 200+ MW, tens of thousands of GPUs, partnerships with Microsoft.
- Lambda Labs: 100+ MW, direct H100 access.
- Hut 8: 1.1 EH/s Bitcoin mining, 26 MW in AI capacity already operational.
PowerCompute’s 26 MW is a rounding error in this market. They aren’t competing—they’re begging for scraps.
The Bitcoin Balance Sheet Trap
They plan to keep BTC. Good for narrative. Bad for risk management.
If AI revenue lags, they may be forced to sell Bitcoin at a loss. I saw this pattern during Terra’s collapse: companies holding volatile assets as collateral while pivoting to new revenue streams often end up liquidating at the worst moment.
Contrarian: Correlation ≠ Causation
This is the part most analysts miss.
The narrative says: "Miner pivots to AI = instant value creation."
But the correlation between rebranding and actual AI capability is zero.
LM Funding was a tiny miner. Their operational expertise is in ASIC management—a simple compute model that requires low latency, low bandwidth, and minimal cooling. AI inference, let alone training, demands high-bandwidth interconnects (NVLink), liquid cooling, and complex parallel compute stacks.
This isn’t a technical upgrade. It’s a business model shift requiring new talent, supply chain relationships, and customer acquisition. None of which they have yet.
Most investors anchor to CoreWeave’s success and assume a similar trajectory. That’s cognitive bias. Small miners lack the scale to negotiate GPU allocations, the brand to attract enterprise AI clients, and the capital to weather long deployment cycles.
Follow the smart money, not the hype.
Smart money (institutional AI investors) doesn’t buy 26 MW stories. They buy proof of execution: signed multi-year contracts, committed GPU orders, experienced HPC management teams.
PowerCompute has none.
Takeaway: A Signal, Not a Strategy
This rebrand is a signal to watch—but not to buy.
Over the next 90 days, monitor three on-chain (or rather, off-chain) signals:
- GPU Purchase Announcement: Any 8-K filing or press release confirming a major GPU order (H100 or B200). Without it, the pivot remains fiction.
- Client Wins: A named AI company signing a long-term compute contract. Even a small one would prove demand exists.
- Balance Sheet Moves: If they issue stock or sell Bitcoin to fund capex, that’s a red flag for dilution.
Until then, treat PowerCompute as a speculative vehicle riding AI narrative. The underlying asset (26 MW and a few BTC) doesn’t justify a premium valuation.
Exit liquidity is someone else’s entry.
Right now, early entrants betting on this narrative are hoping for later buyers to push prices higher. That’s not investing. That’s gambling on attention span.
Code doesn’t care about your feelings.
Neither does the market. When the next quarter’s earnings show zero AI revenue, the narrative will evaporate faster than a misplaced decimal in a smart contract.
Will PowerCompute prove the skeptics wrong? Possibly. But the burden of proof lies with them. And the data, as it stands, screams: probability low, risk high.
Transparency is the only security.
I’ll be watching the filings. You should too.