Two AI Powerhouses Worth Holding Through 2035: A Deep Dive into Long-Term Investment Strategy

I came across an interesting piece from Zacks Investment Research that got me thinking about the difference between short-term AI speculation and genuine long-term value creation. The article, published on November 14, 2025, makes the case for two AI stocks that could be worth holding for the next decade – a refreshing perspective in a market often obsessed with quarterly earnings beats and flashy product launches.

What caught my attention wasn’t just the stock picks themselves, but the underlying investment thesis. As of November 2025, we’re seeing a clear bifurcation in the AI market between companies riding the current wave and those building the foundational infrastructure that will matter in 2035. The article focuses on Microsoft (MSFT) and what appears to be Constellation Energy (CEG), though the full analysis goes deeper into the competitive dynamics shaping this space.

Microsoft’s position in this landscape is particularly fascinating when you examine the numbers. The Redmond-based tech giant has invested over $13 billion in OpenAI since 2019, but more importantly, it’s successfully monetized that partnership across its entire product ecosystem. Azure’s AI services revenue grew 29% year-over-year in Q4 2025, while Microsoft 365 Copilot has reached 70 million users – a penetration rate that would have seemed impossible just 18 months ago. The company’s ability to embed AI capabilities into existing workflows, rather than creating standalone AI products, represents a fundamentally different approach than pure-play AI companies.

The energy angle, represented by Constellation Energy, reveals something crucial that many investors are missing. Data centers powering AI workloads consume approximately 1-2% of global electricity today, but Goldman Sachs projects this could reach 3-4% by 2030. Constellation Energy, headquartered in Baltimore, Maryland, operates the largest fleet of nuclear power plants in the United States, providing carbon-free electricity that’s becoming increasingly valuable as tech companies face pressure to meet sustainability commitments while scaling AI infrastructure.

The Infrastructure Play Behind AI’s Growth

What’s particularly compelling about this investment thesis is how it acknowledges the massive infrastructure requirements that AI scaling demands. NVIDIA’s H100 chips might grab headlines, but the real constraint isn’t silicon – it’s power and cooling. A single AI training run for a large language model can consume as much electricity as 1,000 homes use in a month. This creates a structural advantage for companies that can provide reliable, clean energy at scale.

Constellation’s recent performance reflects this dynamic. The company’s stock has gained 89% year-to-date through November 2025, driven largely by long-term power purchase agreements with hyperscale data center operators. Amazon’s AWS signed a 20-year agreement for 960 megawatts of nuclear power from Constellation’s Susquehanna facility, while Microsoft has committed to purchasing power from the company’s planned reactor restarts. These aren’t speculative bets – they’re infrastructure investments with decades-long payback periods.

The nuclear angle is particularly interesting when you consider the global context. While the United States debates energy policy, China has 27 nuclear reactors under construction and plans to triple its nuclear capacity by 2035. France generates 70% of its electricity from nuclear power and is positioning itself as Europe’s AI infrastructure hub. Constellation’s domestic nuclear fleet suddenly looks like a strategic national asset, not just an energy company investment.

Microsoft’s competitive positioning becomes clearer when you examine how it’s different from other tech giants. Google’s parent company Alphabet has invested heavily in AI research through DeepMind and Google AI, but struggles to monetize these capabilities beyond search advertising. Meta has poured billions into AI research and the metaverse, but faces regulatory headwinds and user growth challenges. Amazon’s AI efforts are fragmented across AWS, Alexa, and various retail applications, lacking the cohesive strategy that Microsoft has achieved.

The numbers tell the story. Microsoft’s productivity and business processes segment, which includes Office 365 and Teams, generated $20.3 billion in revenue during Q1 2025, with Copilot features driving a 12% increase in average revenue per user. This isn’t just AI adoption – it’s AI monetization at enterprise scale. The company’s ability to charge premium prices for AI-enhanced productivity tools creates a sustainable competitive moat that’s difficult for competitors to replicate.

Market Dynamics and Competitive Landscape

The competitive landscape in AI infrastructure is evolving rapidly, and the winners aren’t necessarily the companies with the flashiest demos. Take Arista Networks (ANET), another company mentioned in the analysis. The Santa Clara-based networking company has become essential infrastructure for AI data centers, with its Ethernet switches optimized for the high-bandwidth, low-latency requirements of distributed AI training. Arista’s revenue grew 20% year-over-year in Q3 2025, driven entirely by demand from cloud service providers building AI capabilities.

What’s remarkable about Arista’s position is how it benefits regardless of which AI models or frameworks become dominant. Whether customers choose NVIDIA’s hardware, AMD’s alternatives, or emerging players like Cerebras, they all need high-performance networking. This “picks and shovels” approach to AI investing has historically generated more consistent returns than betting on specific technology winners.

The market is starting to recognize these infrastructure plays. The Global X Data Center REITs & Digital Infrastructure ETF has outperformed the NASDAQ by 15% year-to-date, while pure-play AI stocks have experienced significant volatility. Investors are beginning to understand that sustainable AI profits flow to companies that control essential resources – power, networking, cloud platforms – rather than just those developing AI algorithms.

Meta’s situation illustrates the challenges facing companies without this infrastructure control. Despite spending $13.7 billion on Reality Labs in 2024 and continuing massive AI research investments, the company’s stock has underperformed Microsoft by 23% over the past 18 months. Meta’s dependence on advertising revenue, regulatory scrutiny, and lack of enterprise AI monetization create structural headwinds that infrastructure plays don’t face.

The timing of these investments is crucial. As of November 2025, we’re seeing the early stages of enterprise AI adoption, but the real opportunity lies in the infrastructure that will support AI workloads through 2035. McKinsey estimates that generative AI could contribute $2.6 to $4.4 trillion annually to global GDP, but realizing this potential requires massive infrastructure investments that are just beginning.

Looking at the global competitive dynamics, the United States maintains advantages in AI research and venture capital, but faces challenges in manufacturing and energy infrastructure. China leads in AI application deployment and has substantial manufacturing capacity, while Europe focuses on AI regulation and ethical frameworks. Companies like Microsoft and Constellation Energy that can navigate these geopolitical complexities while maintaining technological leadership are positioned for sustainable long-term growth.

The investment thesis becomes even more compelling when you consider the capital requirements for AI infrastructure. Building a state-of-the-art AI data center costs $1-2 billion and requires 2-3 years of construction time. This creates natural barriers to entry and favors companies with existing infrastructure assets. Constellation’s nuclear plants, some dating to the 1970s, suddenly become valuable strategic assets rather than aging industrial facilities.

From a risk perspective, these infrastructure plays offer more predictable cash flows than pure-play AI companies. Constellation’s nuclear plants operate under long-term contracts with utilities and corporate customers, providing revenue visibility that software companies can’t match. Microsoft’s enterprise software business generates recurring subscription revenue that’s less volatile than advertising-dependent models or hardware sales cycles.

The regulatory environment also favors these established players. Nuclear power faces complex regulatory oversight, but Constellation has decades of experience navigating these requirements. Microsoft’s enterprise focus means it faces less consumer privacy scrutiny than Meta or Google, while its partnership approach to AI development reduces antitrust risks compared to companies developing proprietary AI models.

As we look toward 2035, the AI landscape will likely be dominated by companies that control essential infrastructure rather than those with the most sophisticated algorithms. The democratization of AI tools means that competitive advantages will shift from model development to deployment capabilities, infrastructure control, and customer relationships. Microsoft’s enterprise relationships, combined with its Azure cloud platform and AI partnerships, create multiple layers of competitive protection.

The energy transition adds another dimension to this investment thesis. As governments worldwide commit to carbon neutrality goals, nuclear power is experiencing a renaissance. The International Energy Agency projects nuclear capacity could double by 2050, driven partly by data center demand but also by broader decarbonization efforts. Constellation Energy sits at the intersection of these two massive trends – AI growth and energy transition – creating what could be a decades-long growth opportunity.

This analysis suggests that successful long-term AI investing requires looking beyond the obvious technology plays to identify companies that will benefit from AI adoption regardless of which specific technologies prevail. The infrastructure providers, energy suppliers, and platform companies that enable AI deployment may generate more sustainable returns than the AI developers themselves. It’s a sobering reminder that in technology investing, the companies building the roads often outperform those manufacturing the cars that drive on them.

This post was written after reading Member Sign In. I’ve added my own analysis and perspective.

Disclaimer: This blog is not a news outlet. The content represents the author’s personal views. Investment decisions are the sole responsibility of the investor, and we assume no liability for any losses incurred based on this content.