Undervalued AI Stocks in 2025: Emerging Infrastructure Enablers of the Next-Gen AI Economy

The artificial intelligence (AI) infrastructure market is undergoing a seismic transformation. By 2025, it has expanded to $71.88 billion, up from $56.98 billion in 2024, driven by surging demand for cloud services, data traffic, and cross-industry AI adoption, according to The Business Research Company report. This growth is not merely a function of technological optimism but a response to structural shifts in how societies and economies operate. Yet, beneath the surface of this boom lie critical enablers—companies that are quietly building the scaffolding for the next generation of AI, often at undervalued price points.

The Infrastructure Imperative

AI's computational demands are outpacing Moore's Law, growing more than twice as fast as traditional semiconductor advancements, according to a
MarketBeat forecast. This has created a dual challenge: meeting insatiable compute needs while addressing energy consumption and geopolitical bottlenecks. For instance, U.S. data centers are projected to allocate 20% of their power demand to AI by 2030, according to a
TS2 analysis, a figure that underscores the urgency of infrastructure innovation.

The market is segmented into hardware, software, training, and inference, with hybrid and multi-cloud deployments emerging as dominant strategies, as noted in the Business Research Company report. Geopolitical tensions, meanwhile, have disrupted supply chains for AI accelerators, inflating costs and delaying cloud expansions, a dynamic the report also highlights. These dynamics create both risks and opportunities for investors.

Undervalued Enablers: A Closer Look

Iren Ltd (IREN) has emerged as a standout. Transitioning from Bitcoin mining to AI infrastructure, the company leveraged its access to high-end GPUs like Nvidia's B-series and AMD's MI350X to diversify into AI cloud services. In Q4 2025, Iren reported a staggering 226% year-over-year revenue increase, with AI cloud segment revenue growing 33% quarter-on-quarter, a strong operational pivot that stood out even amid broader market coverage by CNBC. Its trailing P/E ratio of 107.33 and forward P/E of 38.25 suggest a stock priced for future growth rather than current earnings; those valuation figures are reflected in
StockAnalysis data. Analysts have assigned a “Buy” consensus rating, with an average price target of $39.88, per the StockAnalysis page.

Poet Technologies (POET) is tackling a critical bottleneck: energy efficiency in data centers. By deploying optical interposers, the company reduces energy consumption, a feature increasingly valued as AI's power demands escalate. Despite a current P/E ratio of -7.68, Poet's market cap of $520 million reflects its potential to scale. Analysts project a 17.92% upside from its current price of $5.30, with a consensus “Buy” rating cited in the MarketBeat forecast.

Applied Digital (APLD) is another key player. The company's strategic pivot from crypto-mining to AI and high-performance computing (HPC) has positioned it to capitalize on surging demand for data center capacity. In Q4 2025, Applied Digital secured $11 billion in contracted lease revenue with CoreWeave, a testament to its scalability, as reported in the TS2 analysis. While it operates at a net loss (net margin of -107% in the latest quarter), its market cap of $6–7 billion and a stock price up over 200% year-to-date suggest investor confidence in its long-term potential.

Financial Metrics and Analyst Sentiment

The valuation landscape for AI infrastructure stocks is nuanced. Alphabet (GOOG) exemplifies a mature player leveraging AI across its ecosystem, with a P/E of 19.5 and a $3.05 trillion market cap (coverage noted by CNBC). ASML, the semiconductor lithography leader, trades at a P/E of 33.5, reflecting its monopoly on advanced chip manufacturing, as highlighted in a
Forbes list. Nvidia (NVDA), despite a P/E of 52, remains a bellwether due to its 48.9% EPS growth outlook, also discussed in that Forbes roundup.

For mid-cap and smaller players, the story is different. Iren's 28% quarter-on-quarter net profit growth and Poet's energy-saving innovations highlight their potential to outperform. Applied Digital's debt-to-equity ratio of 1.36 signals a capital-intensive model, but its strategic partnerships and contracted revenue streams mitigate near-term risks.

Challenges and the Road Ahead

The path forward is not without hurdles. Energy demands for AI infrastructure are straining global grids, while geopolitical tensions complicate supply chains. However, algorithmic innovations and efficient supply chain management are emerging as countermeasures, as the Business Research Company report notes. Governments, too, are stepping in: the U.S. Bipartisan Infrastructure Law has allocated $1.2 trillion to support critical infrastructure, including AI-related projects, a development discussed in the TS2 analysis.

Conclusion

The next-gen AI economy will be defined by its infrastructure. While giants like Nvidia and Microsoft dominate headlines, smaller enablers such as Iren, Poet, and Applied Digital offer compelling opportunities for investors willing to look beyond current earnings. Their financial metrics, coupled with analyst optimism, suggest these companies are undervalued relative to their long-term potential. As the AI infrastructure market continues to expand, these emerging players may well become the unsung heroes of the AI revolution.