Chardan initiates Keel Infrastructure stock with buy on data center potential

Chardan initiated coverage on Keel Infrastructure (NASDAQ:KEEL) with a Buy rating and a $4.50 price target, implying 37% upside from the $3.29 share price. The firm highlighted $520 million in liquidity, North American data center sites suitable for AI/HPC workloads, and a valuation framework tied to Bitcoin mining, Bitcoin treasury holdings, and potential AI tenant wins. Separately, Bitfarms reported a $209 million fiscal 2025 net loss despite 72% revenue growth to $229 million, while its redomiciliation to Delaware and pivot toward HPC/AI infrastructure continue.

Analysis

The real signal is not the re-rating of a single miner-turned-infrastructure story; it is the market beginning to underwrite optionality on stranded power and land as a scarce input to AI buildout. If management can convert even one site into a credible tenant pipeline, the multiple can expand much faster than the underlying cash flow because the market will start valuing the portfolio like a development platform, not a melting-ice-cube miner.

That said, the second-order risk is execution asymmetry: capital intensity rises before contracted revenue does, so liquidity headlines can look strong while equity value leaks through dilution or capex overruns. The path from “HPC-ready assets” to a signed anchor tenant is typically measured in quarters, but the market is repricing this on days, which makes the stock vulnerable to any delay, power interconnect issue, or tenant decision to wait for next-gen hardware cycles.

For NVDA, this is directionally supportive but not enough to move the tape by itself; the incremental benefit is through ecosystem validation, not unit demand. More important is that data-center scarcity reinforces pricing power across the AI supply chain: GPU vendors, networking, cooling, and power infrastructure all gain negotiating leverage when compute capacity is constrained, while smaller miners without credible conversion plans risk being stranded assets and potential source-of-funds sellers.

The contrarian view is that the move may be front-running a story that is already crowded: a lot of low-quality capital has chased the ‘bitcoin-to-AI’ pivot, and only a subset will actually secure tenants. If the first signed deal lands below market-implied economics, the whole cohort could de-rate because the scarcity premium migrates from land banks to proven execution, which would pressure the names that are currently valued on optionality rather than contracted cash flow.

Key Decisions for Investors

• Long BITF on a 3-6 month horizon only if risk-managed with a tight stop below the last major support; thesis is that signed tenant news can re-rate the stock 20-40%, but downside is >25% if execution slips or dilution risk rises.
• Pair trade: long NVDA / short a basket of small-cap bitcoin miners with AI pivot exposure over 1-3 months; the market is likely to overpay for optionality in the miners while NVDA captures the more durable ecosystem benefit.
• Sell upside calls on BITF into strength for event-risk monetization; the implied upside from headline-driven multiple expansion is attractive, but the stock is vulnerable to sharp pullbacks if no contract is announced within 1-2 quarters.
• Use any post-rally weakness in AI infrastructure names with contract-backed assets to add exposure, but avoid names whose valuation assumes tenant wins within 90 days; execution timelines in this space usually slip by 2-3 quarters.
• If looking for a cleaner second-order expression, long power and cooling infrastructure suppliers on the thesis that data-center conversion spending, not just GPU demand, becomes the scarce bottleneck over the next 12 months.