By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel's net takeaway is that Keel Infrastructure's 2.2 GW pipeline and grid interconnections are promising, but the company's heavy reliance on volatile Bitcoin mining revenue, potential stranded assets due to regulatory hurdles, and lack of proven non-crypto revenue streams pose significant risks that outweigh the opportunities.
Risk: Volatility and regulatory risk associated with Bitcoin mining revenue, and potential stranded assets due to non-transferable grid interconnection agreements.
Opportunity: The scarcity of grid-interconnected power, which is the primary bottleneck for hyperscale AI.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
We just covered From Fired Researcher to $13.7 Billion King: How Leopold Aschenbrenner Broke the Hedge Fund World and Keel Infrastructure Corp. (NASDAQ:KEEL) ranks 20th on this list.
Keel Infrastructure Corp. (NASDAQ:KEEL) first appeared in the 13F portfolio of Situational Awareness LP. in the fourth quarter of 2025. Back then, this position comprised close to 7 million shares. Filings for the first quarter of 2026 show that the fund owns close to 20 million shares in the firm, up around 188% compared to filings for the previous quarter. The firm operates digital and energy infrastructure with focus on high-performance computing (HPC) and artificial intelligence workloads in North America, Canada, and the United States. It primarily owns and operates data centers housing computers to validate transactions on the bitcoin blockchain, as well as sells computational power used for hashing calculations for cryptocurrency mining.
READ MORE: Billionaire Tom Steyer’s 10 Stock Picks with Huge Upside Potential.
Leopold Aschenbrenner is bullish on Keel Infrastructure Corp. (NASDAQ:KEEL). In the current AI land grab, the rarest commodity is not the GPU. It is a secured connection to the power grid. A new grid interconnection request can take five to seven years to clear utility queues. Keel bypasses this queue entirely. The firm owns a development pipeline controlling 2.2 GW of power capacity with established grid interconnections already locked in across prime, low-latency, and energy-dense regions including Pennsylvania, Washington, and Québec. At 2.2 GW of capacity, bulls point out that the market is valuing this infrastructure at less than a fraction of what a ground-up data center build costs to clear utility hurdles.
While we acknowledge the potential of KEEL as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: Growth Stock Portfolio: 12 Stock Picks by Carl C. Icahn and Chris Rokos Stock Portfolio: Top 10 Stock Picks.
Disclosure: None. Follow Insider Monkey on Google News.
Four leading AI models discuss this article
"KEEL's core Bitcoin-mining operations contradict the AI power-infrastructure thesis the article promotes."
The article frames Keel Infrastructure as an AI power play with 2.2 GW of pre-cleared grid capacity, yet explicitly states its primary business is Bitcoin blockchain validation and cryptocurrency mining. Aschenbrenner's stake jump from 7 million to 20 million shares occurred in Q4 2025-Q1 2026 filings, a period when crypto mining economics improved but AI data-center demand was still nascent. This raises the possibility the position reflects hash-rate exposure rather than HPC readiness. Crypto miners face halving cycles, energy-price swings, and shifting regulatory treatment that pure AI infrastructure does not, making the grid-queue narrative potentially misleading.
Even a crypto-centric asset with locked-in power could be repurposed for AI workloads if economics shift, and Aschenbrenner's prior record makes it risky to dismiss the position outright without seeing the full 13F rationale.
"Power scarcity is real, but KEEL's 2.2 GW pipeline valuation relative to execution risk and competitive positioning remains unproven by this article."
Aschenbrenner's 188% Q4-to-Q1 accumulation of KEEL (7M to 20M shares) signals conviction, but the article conflates two separate theses: (1) power scarcity is real and valuable, (2) KEEL's 2.2 GW pipeline is worth the valuation. The first is defensible; the second requires scrutiny. 2.2 GW sounds large until you realize hyperscalers like Microsoft and Meta are each securing 5-10+ GW annually. KEEL's 'established grid interconnections' claim needs verification—interconnection ≠ capacity delivery. Timeline risk is buried: even locked-in requests face permitting delays. The article also omits KEEL's current profitability, debt load, and execution track record. Aschenbrenner's track record is strong, but this is a 13F filing, not a public endorsement with skin in the game beyond portfolio holdings.
If 2.2 GW is genuinely scarce and pre-interconnected, why hasn't KEEL's stock price already reflected this? Either the market knows something Aschenbrenner doesn't, or the interconnections are less 'locked in' than the bull case implies.
"KEEL’s valuation is currently driven by the scarcity of grid-interconnected power, but its reliance on Bitcoin mining revenue creates a dangerous disconnect between its infrastructure value and its cash flow stability."
KEEL’s valuation hinges on the scarcity of grid-interconnected power, which is the primary bottleneck for hyperscale AI. While the market often treats data center operators as REITs, KEEL is effectively a power-arbitrage play masquerading as infrastructure. Aschenbrenner’s 188% stake increase suggests he is betting that the 'utility queue' premium will compress as AI demand outstrips supply. However, the reliance on Bitcoin mining revenue is a massive red flag; it introduces high volatility and regulatory risk that could decouple the stock from the stable, long-term contracts typical of AI infrastructure. If crypto prices crash, KEEL’s cash flow profile deteriorates rapidly, leaving investors holding expensive, stranded assets.
The company’s heavy exposure to Bitcoin hashing makes it a proxy for crypto volatility rather than a reliable AI infrastructure play, potentially leading to a sharp valuation collapse if mining profitability wanes.
"The bull thesis hinges on an unverified 2.2 GW pipeline and interconnections; without proven monetization, financing and execution risk, and crypto/energy volatility could overwhelm any upside."
The article leans bullish on KEEL's claimed 2.2 GW development pipeline and 'locked-in' grid interconnections, implying a near-term uplift from data-center/HPC and crypto-mining infra. Yet the path to value is opaque: enormous capex, potential equity dilution, and lengthy regulatory timelines could erode returns. Crypto exposure and energy-price sensitivity add cyclicality and ESG/regulatory risk that can crush margins. The notion of bypassing utility queues may be overstated or misunderstood. Real monetization requires timing precision and favorable power costs—areas where KEEL has yet to demonstrate sustainable profitability.
Strong counterpoint: the 2.2 GW pipeline and 'bypassed' interconnections read like marketing; without proven execution, financing, or regulatory clearance, the upside may be illusory and could reverse if buildouts lag or crypto demand softens.
"Unstable crypto cash flows cannot reliably fund the capex required for any AI pivot."
Gemini's volatility warning connects directly to ChatGPT's capex point: KEEL's Bitcoin cash flows are too unstable to fund the grid upgrades needed for HPC conversion, raising dilution risk if crypto margins compress further in 2026. Aschenbrenner's stake may simply be riding the post-halving recovery rather than underwriting any AI timeline.
"The dilution and volatility risk hinge on whether KEEL's capex is actually funded by mining cash flow or by pre-signed HPC contracts—a distinction the article and panel have both skipped."
Grok and Gemini both assume crypto cash flows fund capex, but neither checked KEEL's actual balance sheet. If the company already has committed financing or off-take contracts locking in HPC revenue independent of mining cycles, the dilution thesis collapses. The article omits this entirely. Before accepting the volatility-kills-the-thesis argument, we need to know: does KEEL have non-crypto revenue locked in, or is it genuinely 100% mining-dependent? That's the crux.
"Grid interconnection agreements for crypto mining are frequently non-transferable to HPC/data center use, creating a 'stranded asset' risk that invalidates the AI pivot thesis."
Claude is right to demand balance sheet transparency, but misses the regulatory trap. Even if KEEL secures non-crypto off-take, grid interconnection agreements for mining are often non-transferable to data centers. Utilities view high-uptime HPC loads differently than interruptible mining loads. If KEEL attempts to pivot, they may lose their 'locked-in' status entirely, forcing a re-queue. The market isn't just pricing in volatility; it is pricing in the high probability that these assets are legally stranded.
"Monetization depends on transferable, long-term HPC revenue secured via off-takes, not on crypto cash flows or the mere existence of a 2.2 GW queue."
Gemini flags crypto volatility as the red flag; I push on execution risk: even with 2.2 GW in the queue, interconnection rights may be non-transferable and permitting delays inevitable. Utilities treat HPC commitments differently, so KEEL could be stranded or forced to re-queue. Until KEEL proves long-term, non-crypto revenue streams or transferrable, firm off-takes, the upside hinges on a monetization path that may never materialize.
The panel's net takeaway is that Keel Infrastructure's 2.2 GW pipeline and grid interconnections are promising, but the company's heavy reliance on volatile Bitcoin mining revenue, potential stranded assets due to regulatory hurdles, and lack of proven non-crypto revenue streams pose significant risks that outweigh the opportunities.
The scarcity of grid-interconnected power, which is the primary bottleneck for hyperscale AI.
Volatility and regulatory risk associated with Bitcoin mining revenue, and potential stranded assets due to non-transferable grid interconnection agreements.