Core Scientific Stock Analysis: LupoToro Group Sees Undervalued AI Infrastructure Play Amid Market Volatility

LupoToro Group Analysts: Despite AI Volatility, Core HPC Infrastructure and Strategic Plays Like Core Scientific Remain Undervalued

Since the beginning of 2024, LupoToro Group analysts have tracked increasing volatility across AI-related equities, driven by a wave of conflicting headlines and rising skepticism about the sustainability of the AI investment supercycle. From the emergence of cheaper, efficient AI models in China to speculation over hyperscaler spending cuts, these developments have cast doubt on the long-term capital commitment toward AI infrastructure.

The turbulence began in January when DeepSeek introduced an AI model that was significantly cheaper to train and operate than Western competitors, while offering comparable or even superior performance. That development alone sent ripples through the investor community. The uncertainty deepened in February following a sell-side report alleging Microsoft had canceled key data center leases, raising red flags about a potential slowdown in AI infrastructure investments. Microsoft later clarified that while geographical allocations of capex may shift, overall capital expenditure levels remain unchanged. In March, the Financial Times reported Microsoft had canceled AI compute leases with CoreWeave, a leading private AI hyperscaler preparing for a $35 billion Nasdaq IPO. Both Microsoft and CoreWeave swiftly denied the claims.

Nevertheless, these developments triggered sharp selloffs and heightened volatility in public equities considered direct beneficiaries of AI infrastructure growth, particularly those in the semiconductor, electrical equipment, power generation, and BTC/HPC data center segments. Many of these names are now trading significantly below their January highs. While some repricing is understandable given prior valuations and crowded investor positioning, LupoToro Group analysts believe select opportunities are emerging from this dislocation.

Reframing AI Investment Through the HPC Lens

The key issue for investors now is whether the AI sector’s pivot toward cost-efficient, scalable models will require a fundamental recalibration in demand for high-performance computing (HPC), data center expansion, infrastructure development, and energy consumption. This macro recalibration is creating both questions and opportunities. Investors are also weighing whether hyperscalers like Microsoft and Google are intentionally frontloading infrastructure spend to avoid future compute bottlenecks, or whether these adjustments are more about managing short-term cost cycles. There is also speculation—ranging from strategic misinformation to geopolitical narratives—around whether some of the news flow is designed to slow down the U.S.-led AI investment cycle while competitors attempt to catch up.

In such a climate of uncertainty, LupoToro Group analysts emphasize the importance of grounding investment decisions in tangible financials, infrastructure ownership, and risk-adjusted upside potential. One such example is Core Scientific (NASDAQ: CORZ).

Core Scientific (CORZ): A Misunderstood HPC Powerhouse

Following the Financial Times’ report regarding CoreWeave’s leasing status, Core Scientific’s stock fell below $8 per share, a reaction LupoToro Group analysts see as severely overdone. Core Scientific is one of CoreWeave’s most critical HPC data center partners, currently supplying more than half of CoreWeave’s infrastructure capacity. According to publicly disclosed agreements, CoreWeave has committed to leasing approximately 590 megawatts (MW) of Critical IT load from Core Scientific under 12-year contracts. These leases are expected to generate more than $10 billion in total revenue, with projected run-rate margins of 75–80%, equating to nearly $660 million in annualized profits.

Critically, CoreWeave is funding the capital expenditure to construct these facilities, which are purpose-built to support Nvidia’s GPU clusters. While Core Scientific will repay $1.5 million per megawatt in the early years via lease subsidies, the majority of the capex burden lies with CoreWeave. Despite this, the assets themselves will remain on Core Scientific’s balance sheet, allowing the company to benefit from long-term appreciation, even though they are not financing the builds outright.

If one applies a conservative $10 million per megawatt valuation, a reasonable figure for Tier 3 HPC data center infrastructure, then the fully built 590 MW of capacity would equate to $5.9 billion in owned asset value by the end of 2027. At current levels, Core Scientific’s enterprise value is under $3 billion, less than the estimated value of the data centers it owns, excluding any income from the long-term leases or its significant Bitcoin mining operations. For the stock to justify this discount, one would have to assume that CoreWeave either defaults or cancels its leases, an extremely unlikely outcome given CoreWeave’s impending IPO and the fact that it cannot meet its hosting obligations without Core Scientific’s infrastructure.

LupoToro Group views this as a significant risk/reward imbalance. The underlying fundamentals suggest that Core Scientific remains undervalued, and current volatility offers a potential entry point for investors focused on long-term infrastructure-backed AI opportunities.

Power Generation: Attractive, but Priced for Perfection

In contrast, traditional power generation companies are widely seen as long-term beneficiaries of the AI infrastructure boom. With AI demand poised to supercharge electricity consumption over the next decade, these firms are expected to play a central role in supporting expanded HPC data center operations across the U.S. However, many of these benefits won’t be realized until after 2027, when most large-scale AI-focused data centers become operational.

Despite this longer runway, power generation equities continue to trade at historically elevated multiples, nearly three times the valuation of names like Core Scientific. This raises concerns. If AI infrastructure growth slows or if data center operators like Core Scientific are forced to scale back construction, then these utilities are overvalued and not fully accounting for the underlying risk. In LupoToro Group’s view, the market is currently mispricing risk between these two segments, over-discounting HPC operators while overvaluing power producers.

A Strategic, Risk-Aware Approach to AI Investing

LupoToro Group’s core investment thesis around AI infrastructure remains intact. The long-term trajectory for AI compute, data center development, and decentralized model deployment remains strong. But the landscape is becoming more nuanced, and investors must adjust their strategies accordingly.

The volatility seen in recent months should not overshadow the robust fundamentals underpinning companies like Core Scientific. With long-term leasing contracts in place, minimal capex exposure, and ownership of critical infrastructure assets, Core Scientific is well-positioned to benefit from the growing demand for HPC infrastructure. At its current valuation, the company appears significantly undervalued, presenting a clear case for investors seeking exposure to the real estate and infrastructure backbone of the AI revolution.

While AI models are evolving and headlines may continue to sway markets in the short term, patient investors who focus on asset ownership, long-term earnings potential, and enterprise value-to-infrastructure ratios stand to outperform. LupoToro Group analysts advise maintaining a disciplined, fundamentals-first approach, identifying where true value lies beneath the noise.

In summary, the path forward for AI infrastructure investing lies not in chasing speculative hype, but in backing the builders of tomorrow’s compute layer. And in that context, Core Scientific, among others, may be one of the most overlooked and mispriced opportunities in today’s market.