The $4 Trillion Foundation: Why AI Infrastructure is one of the Decade's Most Enduring Investment Thesis / Blog 031
- Joe
- Oct 21
- 15 min read
The Invisible Backbone of the AI Economy: Chips, Power, and Geopolitics
Artificial Intelligence is no longer a futuristic concept; it is a trillion-dollar reality. Behind every advanced large language model, autonomous vehicle, and sophisticated drug discovery algorithm lies an immense, power-hungry, and capital-intensive physical infrastructure. This infrastructure — the GPUs, high-bandwidth memory, specialized data centers, and the networking gear connecting it all — is quickly emerging as the most crucial and potentially most profitable sector in the global technology landscape.
The core investment thesis is simple: While AI applications may rise and fall with consumer trends, the providers of the architecture sell the indispensable picks and shovels of this new digital gold rush. Investors are recognizing this dynamic, increasingly treating AI infrastructure not as a temporary tech theme, but as a structural, decade-long capital expenditure (CapEx) cycle, echoing the hyperscale cloud build-out of the 2010s.
The Unprecedented Scale of the Investment Cycle
The sheer scale of the AI build-out is unprecedented. Industry leaders and analysts project massive capital outlays over the next five to seven years:
Trillion-Dollar Projections: NVIDIA, the undisputed leader in AI chips, forecasts that global AI infrastructure spending could surge to between $3 trillion and $4 trillion by 2030.
Wider Estimates: McKinsey & Company offers an even more aggressive outlook, suggesting that total spending on AI infrastructure could reach up to $6.7 trillion by the end of the decade, with nearly half of that allocated toward chips and related hardware for AI-equipped data centers.
Immediate Spend: The major hyperscalers — Amazon, Microsoft, Alphabet, and Meta — are collectively projected to spend hundreds of billions of dollars annually, primarily directed toward outfitting and building next-generation AI data centers.
This spending focuses heavily on the Hardware segment, which constituted the largest share of the AI infrastructure market in 2023 (around 60-65%), underscoring the centrality of physical components to the AI revolution.
The New Frontier: Energy, Cooling, and Geopolitics
The demands of large-scale AI processing are creating entirely new bottlenecks, opening up massive investment opportunities in areas previously considered auxiliary: energy and data center management.
The Power Crisis and Utilities
AI workloads are exponentially more power-intensive than traditional cloud computing.
A single, advanced AI-focused data center can consume as much electricity as 100,000 households, and the largest ones under construction are projected to consume up to 20 times that amount.
Analysts predict that global electricity demand from data centers could more than double by 2030, putting immense strain on existing power grids. This power crisis is creating an investment tailwind for:
Power Generators: Utilities and Independent Power Producers (IPPs) are seeing an unexpected boom in long-term demand.
Sustainable Solutions: Companies involved in next-generation, high-density power solutions, including Small Modular Reactors (SMRs), geothermal, and advanced battery storage, are becoming critical infrastructure plays.
Advanced Cooling Technology
The latest AI chips generate so much heat that conventional air conditioning is insufficient. This necessity is driving the immediate adoption of liquid cooling—where cold water is pumped directly into server rooms to manage temperature. This shift creates a new sub-sector within the hardware supply chain, benefiting providers of specialized racks, heat exchangers, and cooling distribution units. Furthermore, climate risks, such as extreme heat and drought, are raising data center operational costs, making efficient cooling solutions an urgent financial imperative.
Geopolitics: The Silicon Curtain and Supply Chain Resilience
The AI infrastructure race is fundamentally intertwined with global strategic competition, primarily between the United States and China, raising both profound risk factors and structural investment opportunities. This confrontation has swiftly transformed the global technology supply chain from an efficiency-driven, open market into a battleground for national security and technological supremacy. This environment has seen the rapid descent of a "Silicon Curtain," effectively bifurcating the technology world into distinct, yet interdependent, spheres.
The core tension stems from stringent export controls implemented by the U.S. government, aimed at curbing China's access to the cutting-edge hardware—specifically the highest-performance Graphics Processing Units (GPUs) and the sophisticated Extreme Ultraviolet (EUV) lithography tools required to manufacture them. These controls directly impact leading chip designers like NVIDIA, forcing them to develop intentionally downgraded, "China-compliant" versions of their flagship chips (such as the A800 or H20). While these controls secure U.S. technological leadership, they also fragment the market, diverting valuable R&D resources and limiting the revenue potential of multinational companies from a massive addressable market.
For investors, this geopolitical reality necessitates a profound shift in focus from pure cost efficiency to supply chain resilience and redundancy. The imperative to move manufacturing capacity—especially for advanced logic and memory chips—out of concentrated regions has become a strategic, state-backed priority. This push for onshoring and friend-shoring provides a long-term, secular tailwind for firms involved in equipment supply, advanced materials, and the construction of new domestic fabrication plants, such as those being built by Taiwan Semiconductor Manufacturing (TSMC) in Arizona and Japan.
The Rise of Techno-Nationalism
The world is witnessing the descent of a "Silicon Curtain", with the U.S. implementing stringent export controls to restrict China’s access to the most advanced AI chips and manufacturing equipment. This political environment is forcing global tech firms to restructure their supply chains, prioritizing resilience and redundancy over pure economic efficiency. This shift benefits:
Diversified Manufacturing: Companies involved in the onshoring or "friend-shoring" of advanced chip fabrication outside of concentrated areas (e.g., TSMC's build-outs in Arizona and Japan).
Custom Chip Designers: As export controls restrict access to cutting-edge U.S.-designed GPUs, major cloud players are increasingly investing in developing their own custom AI accelerators (ASICs and TPUs), opening opportunities for specialized design and testing firms.
The geopolitical landscape has made hardware and energy the definitive chokepoints of the new AI-driven global operating system.
Mapping the Ecosystem: More Than Just Chips and Servers Behind the Most Enduring Investment Thesis
Investing in AI infrastructure isn't limited to a few well-known hardware manufacturers. In fact, a deep, vertically integrated ecosystem lies beneath the surface, offering diverse entry points for investors — from the smallest, yet most critical, component suppliers to large-scale enterprise software platforms.
Just as value creation gradually spread from the physical network layers to the application layers during the rise of the Internet and cloud computing, today, the infrastructure layer offers the most solid foundation in the AI boom. However, the investment opportunities encompass the entire "stack."
This is an overview of the main contributors advancing this ecosystem. While most analysts concentrate on the prominent "giants," the ongoing investment is increasingly benefiting the smaller, strategically crucial "next-tier enablers" as well.
The following list is a subjective selection, and it's important to be aware of the risks associated with individual investments. To mitigate the risk of investing in a single entity, we will also provide some ETFs (Exchange Traded Funds) for better diversification below.
The big players already leading the race
No discussion of the AI revolution would be complete without the titans already shaping its infrastructure. These companies aren’t just beneficiaries of the trend — they’re the ones building its very foundations. From silicon to security, they form the technological backbone of a future increasingly defined by intelligent computation. And while valuations can feel stretched at times, Wall Street’s conviction in their long-term relevance remains strong.
At the center of it all stands NVIDIA (NVDA) — the undisputed leader of the AI hardware ecosystem. Its GPUs have become synonymous with machine learning, powering everything from ChatGPT to autonomous vehicles. Despite bouts of volatility following its explosive multi-year rally, Nvidia continues to deliver on both innovation and revenue growth. The stock trades around $180, while analysts’ 12-month consensus target remains close to $220, signaling confidence that demand for AI accelerators will persist well beyond the initial boom. Many portfolio managers still view Nvidia not just as a semiconductor company, but as the “core holding” of the entire AI era — a linchpin around which the rest of the industry revolves.
Alongside Nvidia, Micron Technology (MU) plays a quieter but equally crucial role. AI workloads require vast quantities of high-bandwidth memory, and Micron sits at the forefront of that supply. The company’s HBM3E chips have become essential in data centers optimized for training and inference tasks. Trading near $200, Micron’s consensus price targets in the low-to-mid $200s reflect analysts’ belief that the AI wave will keep memory demand strong and pricing power intact. In many ways, Micron’s story is a reminder that every AI model — no matter how sophisticated — ultimately depends on efficient access to memory.
Further down the silicon stack, Marvell Technology (MRVL) has emerged as a critical enabler of AI networking. Its semiconductors facilitate the high-speed data movement required inside hyperscale data centers, where GPUs must communicate seamlessly across racks and nodes. Marvell’s stock currently trades in the $70–80 range, with a consensus target near $90. Analysts increasingly describe the company as a “next-tier infrastructure play” — not as headline-grabbing as Nvidia, but deeply embedded in the AI supply chain. As data throughput demands surge, Marvell’s role in optimizing the connections between chips may become even more indispensable.
Finally, as artificial intelligence moves from lab environments into enterprise-scale deployment, Fortinet (FTNT) stands to benefit from a parallel surge in cybersecurity needs. Each AI model integrated into production systems expands the surface area for potential digital threats. Fortinet’s suite of AI-enhanced network protection tools has positioned it as a leading name in secure infrastructure — a theme few investors can afford to ignore. With shares hovering in the $60–65 range and analyst targets averaging around $75–78, the company is widely viewed as a steady, long-term participant in the broader AI transition. While its valuation often fluctuates with quarterly results, the strategic importance of cybersecurity in an AI-driven world keeps Fortinet firmly in Wall Street’s “must-watch” category.
Together, these giants form the architecture of artificial intelligence itself — the chips, memory, networks, and protections that make the technology both possible and scalable.
Unlike speculative names, these firms combine proven earnings power with exposure to one of the most transformative trends in decades. For many investors, they represent the foundation stones of the modern AI economy.
The Rising Stars: Smaller Players Powering the AI Ecosystem
Beyond the mega-caps dominating headlines, a new generation of smaller, high-growth companies is steadily earning the attention of analysts and institutional investors. These firms might not yet have the scale of Nvidia or Micron, but each sits at a critical junction in the AI value chain — from data connectivity to cybersecurity and enterprise AI software. Their valuations are volatile, but their stories are increasingly compelling.
C3.ai (AI) remains one of the few publicly listed companies whose core business is entirely built around enterprise AI applications. The company develops modular AI tools for corporations seeking to integrate predictive analytics and automation into their operations. After a volatile year, its shares trade around $17.70, while the average Wall Street 12-month price target sits near $22. Analysts remain divided — some flagging weak near-term profitability, others pointing to the company’s early positioning as enterprises begin adopting production-grade AI solutions at scale. If C3.ai manages to translate proof-of-concept deployments into recurring contracts, it could evolve into a meaningful software player in the AI economy. For now, investor sentiment remains cautious but watchful — a classic “early innings” story.
Another company quietly benefiting from the AI expansion is SentinelOne (S), a cybersecurity specialist that uses machine learning to detect and neutralize threats autonomously. As AI systems themselves become targets for new types of cyberattacks, demand for AI-powered security has surged. SentinelOne’s stock currently trades near $16.70, with an average analyst target of about $24, implying healthy upside potential. Analysts tend to highlight the company’s strong technology but note that execution and profitability metrics will remain the key catalysts for sustained re-rating. In an environment where digital infrastructure is multiplying, SentinelOne’s approach to scalable, automated defense could make it a structural winner in the long term.
The more technical side of the infrastructure stack brings us to Credo Technology Group (CRDO), a lesser-known but increasingly essential component supplier. Credo’s high-speed interconnect solutions link GPUs and chips within AI servers — the invisible circuitry that allows AI clusters to communicate efficiently. The company’s shares have climbed into the $140 range, and analysts’ consensus price targets currently hover between $106 and $133, though recent upgrades from major banks like JP Morgan and TD Cowen stretch as high as $150–165. That dispersion reflects both opportunity and risk: Credo’s technology is vital, but the market for such specialized components is still developing. Nevertheless, analysts’ tone has turned notably bullish in recent months, as hyperscale data-center spending continues to accelerate.
Meanwhile, Celestica (CLS) has evolved from a low-margin electronics manufacturer into one of the most interesting AI hardware plays of 2025. Acting as a contract manufacturing and design partner for hyperscalers, Celestica now helps build the servers and racks underpinning next-generation data centers. Its stock trades around $270–280, and while consensus price targets vary by source — averaging roughly $229, with some bullish calls reaching $340 — most analysts have upgraded the stock over the past quarter. Momentum has been strong since several banks reclassified Celestica as an “AI infrastructure supplier,” recognizing its deeper integration into high-value data-center solutions. For investors looking beyond semiconductors, Celestica represents the manufacturing side of the AI supply chain — less glamorous, but equally vital.
Finally, an addition to this group worth mentioning is Snowflake (SNOW). While not a hardware company, Snowflake provides the cloud-based data architecture that enables AI systems to train, deploy, and scale effectively. Its platform allows enterprises to centralize, clean, and share massive datasets — a prerequisite for meaningful AI outcomes. The stock recently traded near $240, with analyst consensus targets in the $251–257 range, and several bullish projections stretching past $300. Analysts generally view Snowflake as a key beneficiary of the data-driven AI boom: as models grow larger and enterprises invest in data readiness, Snowflake stands to capture a growing share of that workflow.
Collectively, these smaller players illustrate the diversity of the AI ecosystem.
From enterprise software and autonomous cybersecurity to networking and manufacturing, each represents a distinct piece of the puzzle that powers artificial intelligence at scale. While they don’t carry the same market heft as Nvidia or Micron, they operate in high-growth niches with meaningful long-term potential — and that’s exactly where early-stage opportunities often emerge.
AI's Global Horizon: The ETFs Spreading Your Bet Beyond Silicon Valley
The rise of Artificial Intelligence is one of the most compelling investment themes of the decade, but navigating the volatile world of individual tech stocks can be daunting. To capitalize on the AI revolution while protecting your portfolio from the extreme swings of single company risks, the best approach is to embrace globally diversified ETFs.
These funds act as a basket, giving you exposure to the entire AI ecosystem, from the chip makers building the hardware to the software companies creating the applications.
And crucially, they look beyond the dominant U.S. market to capture innovation wherever it occurs around the globe.
Here is a narrative breakdown of the top international and globally-minded AI ETFs, designed to be the bedrock of your AI investment thesis.
The Global Generalist: Global X Artificial Intelligence & Technology ETF (AIQ)
Imagine a single portfolio that travels the world to find the most promising AI developers. That is the philosophy behind the Global X Artificial Intelligence & Technology ETF (AIQ). This fund - with an AUM of USD 6.5 billion as of October 2025 - provides a broad, unconstrained, and globally diversified path to the AI theme by tracking an index of companies involved in AI development and Big Data.
While it naturally holds major U.S. tech giants, its strength lies in its intentional, significant allocation to international AI leaders. When you invest in AIQ, you are explicitly gaining exposure to key players outside the U.S. market, such as:
Taiwan Semiconductor Manufacturing (TSM): The world's largest contract chipmaker, whose advanced chips are essential infrastructure for AI data centers globally.
Samsung Electronics (005930 KS): A South Korean giant active in AI research, chip manufacturing, and consumer electronics.
Alibaba Group (BABA): One of China's largest technology conglomerates, deeply integrated into the country's cloud and AI landscape.
This built-in global perspective makes AIQ an excellent choice for a core AI holding that intrinsically diversifies away U.S.-specific risks.
The Full Value Chain Tracker: iShares Future AI & Tech ETF (ARTY)
BlackRock's iShares Future AI & Tech ETF (ARTY) is designed for investors who want exposure to the entire AI value chain. This means it doesn't just focus on the end-user applications but also the foundational infrastructure.
ARTY ETF is explicitly constructed to track an index of both U.S. and non-U.S. companies across four critical AI verticals:
Generative AI: The cutting-edge companies building the models (like large language models).
AI Data & Infrastructure: The hardware and cloud services that power AI (chips, data centers).
AI Software: The enterprise and consumer software products that utilize AI.
AI Services: The consulting and deployment firms helping businesses integrate AI.
By casting a wide net across these global sub-sectors, ARTY minimizes single-point failure risk, ensuring your portfolio benefits from the AI growth trend regardless of whether a hardware or software company is leading the charge at any given time.
European-Domiciled Global Exposure (UCITS Funds)
For many international investors, particularly those in Europe, using UCITS ETFs (funds domiciled in Ireland or Luxembourg) offers benefits such as regulatory oversight and potential tax efficiency. These funds provide the global exposure you seek, but under a European fund structure (identifiable by the ISIN starting with IE or LU).
These UCITS options are typically accumulating (re-invest dividends), offering a straightforward long-term path to capital appreciation while providing the desired geographic diversification. Crucially, they operate under the strict European UCITS regulatory framework, which provides investors with a high level of consumer protection regarding fund structure and liquidity.
The Applied AI Specialist: Global X Robotics & Artificial Intelligence ETF (BOTZ)
If your investment thesis is particularly strong on the application of AI in physical systems, then the Global X Robotics & Artificial Intelligence ETF (BOTZ) offers a more specialized angle.
While its name highlights robotics, the fund focuses on the convergence of robotics and AI in fields like industrial automation, healthcare (e.g., surgical robots), and autonomous vehicles. This thematic focus means it tends to have a high concentration of companies from nations where industrial technology is strong, such as Japan and Germany, in addition to US leaders like Nvidia.
BOTZ ETF is a good choice to complement a broader AI fund, providing targeted exposure to the manufacturing and industrial side of the AI trend that other software-heavy funds might miss.
By spreading your capital across multiple, globally-focused AI ETFs, you effectively diversify your bet, allowing you to participate in one of the world's most transformative technological shifts while greatly reducing the inherent risk of this high-growth sector.
Your Move: Playing the AI Investment Game
The Artificial Intelligence revolution isn't a distant science fiction scenario; it is the definitive investment theme of our time, the most enduring investment thesis of the decade. rewriting the industrial and economic landscape in real-time. From autonomous cars to generative software that writes code, AI is evolving from a single sector into the foundational layer upon which all modern business will be built.
As we've explored, the core decision isn't if you should invest, but how. Do you choose to pursue the single-stock path — the high-stakes game of picking the next NVIDIA or the next OpenAI darling — with its potential for colossal gains but its matching risk of colossal losses? Or do you opt for the more diversified, steady approach offered by global ETFs?
These internationally-focused funds (like AIQ, ARTY, or the UCITS structures like the above mentioned X-Trackers or Wisdome Tree funds) are designed to be a durable foundation, capturing the entire global value chain, from Asian chip manufacturers to European robotics firms. They offer a mechanism to profit from AI's inevitable march without being vulnerable to the missteps of a single CEO, a single product launch, or a single national market.
Ultimately, your strategy is a reflection of your own risk tolerance and conviction. The AI story is not a race but a decades-long transformation. So, the question remains: How will you play the game?
Will you be the high-conviction sniper, betting heavily on one or two companies you believe will define the future? Or will you be the patient builder, deploying a globally diversified strategy to ensure you capture the growth of AI, no matter where in the world the next great innovation is born?
The future of technology, and perhaps your portfolio, awaits your move.
Your Fortune Blog 031
Disclaimer: Important Information about Your Fortune Blog
Please note that the information provided on this website about investments is for general informational purposes only. We do not intend to sell any product or recommend any particular financial institution or investment product. The content provided is not intended to constitute investment advice or to offer any investment products or services whatsoever. It is important to note that investing involves risk, and the reader must take full responsibility for their own investment decisions. We strongly advise seeking professional financial advice before making any investment decisions, particularly if you are subject to cross-border restrictions or tax liabilities.
Please be aware that the information we provide may not be suitable for individuals subject to cross-border restrictions or tax liabilities in their country of residence or jurisdiction. It is your responsibility to understand and comply with any applicable laws and regulations, and we do not accept any liability for any consequences resulting from non-compliance.
The views expressed in this blog are solely those of the authors and do not necessarily reflect the views of any financial institution, organization, or individual. We make no representations or warranties of any kind, express or implied, about the completeness, accuracy, reliability, suitability or availability with respect to the content provided in this blog. Any reliance you place on such information is therefore strictly at your own risk. We shall not be held liable for any loss or damage whatsoever arising from the use of this educational blog, including without limitation any direct, indirect, incidental, consequential, or punitive damages.