Investing in AI: Assessing the investment opportunities

Artificial intelligence (AI) has moved from a promising idea to one of the largest investment cycles in the global economy. Businesses increasingly use AI to automate work, analyze information, develop products, and improve customer service. Technology companies are investing heavily in the computing capacity, data centers, and power systems that support these applications.

The opportunity reaches beyond companies developing AI models. Semiconductor manufacturers, data centers, utilities, electrical-equipment companies, networking providers, cybersecurity firms, and software developers all participate in the AI ecosystem. Companies in other industries also stand to benefit when they use AI to increase productivity or create new sources of revenue.

The investment case now faces a more demanding test. Record capital spending must produce acceptable financial returns, while election results reflect local opposition to data centers and new chip competition adds uncertainty. These developments reinforce the distinction between the long-term growth of artificial intelligence and the investment results of any one company or technology.

AI investment has reached an industrial scale

The largest technology companies are committing extraordinary amounts of capital to artificial intelligence. Amazon, Alphabet, Meta Platforms, and Microsoft expect their combined spending to increase sharply in 2026 and stay elevated in 2027. Some technology companies are also issuing more debt to finance expansion, increasing the importance of balance-sheet strength and future cash generation.

The scale becomes clearer when compared with another capital-intensive industry. During the trailing four quarters through the third quarter of 2026, the large U.S. technology companies included in our analysis invested approximately $805 billion, more than six times the roughly $126 billion major U.S. oil producers invested at their 2014 peak.¹ AI has transformed part of the technology sector into a major physical-infrastructure investment cycle.

Capital spending also exceeded the technology companies’ combined dividends and stock buybacks during the same trailing four-quarter period.¹ This marks a significant shift for businesses historically known for generating substantial cash without requiring comparable physical investment. Investors now need to assess how quickly spending on data centers, processors, power systems, and networking equipment translates into revenue, earnings, and free cash flow.

What is artificial intelligence? Key terms for investors

Artificial intelligence describes computer systems that perform tasks commonly associated with human intelligence, including recognizing patterns, making predictions, and generating content. Traditional software follows instructions written in advance, while machine learning allows systems to improve as they analyze additional information. Machine learning already supports fraud detection, credit scoring, and recommendations from streaming and shopping platforms.

Deep learning uses more complex models for capabilities such as language translation, facial recognition, and meeting transcription. Generative AI creates new text, computer code, images, audio, or video by identifying patterns in large collections of information. Agentic AI starts with a goal, plans a series of actions, and completes multiple tasks with limited human direction.

These technologies require different combinations of computing power, data, security, software, and infrastructure. Early investment concentrated on advanced semiconductors used to train large models, but adoption now reaches cloud computing, cybersecurity, data management, and business automation. That expansion creates more investment opportunities while making the AI ecosystem increasingly complex.

Strong AI demand faces a higher financial hurdle

AI-focused companies have supported much of their market value with earnings growth rather than promise alone. Companies in the Bloomberg AI Index generated annualized earnings growth of approximately 30% over the six years through November 9, 2026. This record differs from the late-1990s technology boom, when many highly valued companies had limited profits and weaker balance sheets.¹

Recent results continue to show robust demand for cloud computing and AI services. Cloud revenue increased 39% at Amazon Web Services, 48% at Microsoft Azure, and 103% at Google Cloud from a year earlier.¹ Businesses continue to purchase computing capacity, but investors increasingly expect providers to show how that demand translates into acceptable returns on investment.

Strong AI results do not protect a company’s stock when another part of its business disappoints. Advertising, retail, search, and software operations generate much of the cash supporting AI investment at the largest technology companies. Weaker performance in those businesses constrains financial flexibility, increases reliance on borrowing, and reduces the cash available for dividends and stock buybacks.

AI investment raises the earnings bar

A strong investment theme does not guarantee strong performance for every stock tied to it. A company can report rapid cloud growth and still fall short of expectations for revenue, profit margins, or cash flow. Elevated valuations make share prices more sensitive when spending plans assume years of exceptional growth.

Investors need to evaluate the entire business supporting an AI investment program. Established operations must generate enough cash to fund new infrastructure while meeting other corporate commitments. The key question is increasingly whether future AI-related earnings will justify record investment, with valuation risk potentially rising if share prices outpace earnings.

Election results raised data center development risk

The 2026 U.S. midterm elections increased political risk for AI infrastructure. Changes in congressional and gubernatorial control create greater scrutiny of data center tax incentives, electricity use, grid connections, and construction approvals. Local results sent a more direct signal as voters broadly supported measures that restrict development or require developers to address power, water, and community costs.

The results do not point to widespread cancellation of projects already underway. More than 75% of data center capacity planned for 2027 is already under construction, providing greater protection for facilities expected to open in the near term. The larger uncertainty concerns 2028 and later, when more projects still depend on permits, utility connections, tax policy, or local approval.

Restrictions will influence the pace and location of development more than the long-term demand for computing capacity. Delays weaken near-term demand for semiconductors, networking equipment, and electrical systems, while slower capital spending improves free cash flow for some cloud providers. The investment effects vary based on where a company participates in the AI supply chain.

Competition and connected financing raise company risk

Rapid innovation expands the AI market while changing which companies capture the returns. Google recently introduced two specialized versions of its tensor processing units, increasing competition with NVIDIA and giving customers more choices for specific computing needs. Large cloud providers can develop processors for selected tasks while continuing to buy chips from established suppliers, reducing their dependence on one provider and increasing pressure on pricing and profit margins.

NVIDIA also faces an independent board investigation after reports that chief executive Jensen Huang had been the target of a foreign intelligence effort. NVIDIA said it had no reason to believe national security or company secrets were compromised, but the review adds leadership, governance, and regulatory uncertainty. Together, the investigation and new chip competition show how company-specific developments can pressure a market leader even as overall AI demand grows.

Financial connections can extend that pressure across the AI ecosystem. Semiconductor and cloud companies sometimes invest in AI developers that use the funding to purchase their chips or computing services. A sharp decline in a major participant’s stock price can limit its capacity to support customers and partners, pressure connected private-company valuations, and make other investors less willing to provide capital, reinforcing the value of investing across the AI ecosystem rather than relying on one company or technology.

Business adoption will test the AI investment cycle

Business adoption provides the clearest test of whether AI infrastructure spending produces sustainable returns. Data from corporate spending platform Ramp show that the share of U.S. businesses paying for AI tools continued to rise through September 2026. Continued adoption converts infrastructure investment into recurring revenue, while slower adoption lengthens the time suppliers need to recover their spending.

The next phase depends less on companies announcing AI initiatives and more on documented business improvements. Organizations need to show that AI increases productivity, improves customer service, reduces operating costs, or creates revenue. Those outcomes broaden demand from semiconductors and cloud computing toward software, cybersecurity, data management, and workflow automation.

Adoption remains uneven across industries and applications. Coding offers one of the clearest tests because businesses can measure whether AI helps developers complete usable work more quickly. Improvements in research, content creation, and administrative processes also create value, although companies often find those gains harder to separate from other operational changes.

Trusted data can strengthen competitive advantages

The most durable applications combine advanced models with proprietary information, secure systems, and established business processes. Companies with trusted data and direct customer relationships hold advantages over providers offering similar features without differentiated information. Usage, customer retention, pricing, and financial results help investors distinguish lasting adoption from experimentation.

Security and regulatory controls also influence adoption. Businesses in financial services, healthcare, and other regulated industries need to understand how an AI system uses information, protects customer data, and reaches important decisions. Providers that meet these requirements can gain opportunities even without offering the largest or best-known AI model.

Artificial intelligence is reshaping software companies

AI is changing how software companies develop products, serve customers, and maintain competitive advantages. New tools automate coding, contract analysis, document creation, and customer service, helping businesses complete tasks faster and at lower cost. These capabilities improve productivity while replacing features that previously supported premium software prices.

Products embedded in essential business processes retain pricing power when customers value trusted data, security, regulatory controls, and integration with existing systems. Less-differentiated products face pressure when customers can switch providers or reproduce important functions with lower-cost tools. AI will create opportunities for some software companies while weakening the advantages of others.

Investors need to distinguish between businesses that add AI terminology to existing products and those that create durable revenue. Customer retention, usage growth, pricing, profit margins, and computing costs show whether an AI product creates financial value. Companies that document productivity gains or help customers generate revenue hold stronger positions than those relying mainly on long-term forecasts.

AI investment opportunities extend across the stack

Investors do not need to identify one company, processor, or application that will dominate artificial intelligence. Innovation, policy decisions, leadership changes, and financing conditions quickly alter competitive positions. Investing across the AI ecosystem reduces dependence on one perceived winner.

This “across the stack” approach begins with the physical infrastructure needed to develop and operate AI systems. It includes semiconductor designers and manufacturers, data centers, networking equipment, electrical systems, cooling providers, utilities, and other sources of reliable power. It also extends to cloud computing, cybersecurity, data management, software, and companies using AI to improve productivity.

Each layer responds to a different mix of opportunities and risks. Semiconductor companies benefit from demand for computing power but face new processor designs and competition from customers developing chips internally. Utilities and infrastructure businesses benefit from rising electricity demand while managing regulation, financing costs, and construction requirements.

AI opportunities reach beyond technology stocks

Software and cloud providers generate recurring revenue as adoption increases, but they must show that their products improve customer retention, pricing, or profit margins. Companies also create value by using AI more effectively without selling AI products themselves. Broad exposure captures different sources of return without requiring one technology or business model to remain dominant.

U.S. large-cap stocks provide exposure to leading semiconductor, cloud, and technology companies as well as businesses using AI to improve productivity. Emerging-markets investments add semiconductor manufacturers and other suppliers within the global technology chain. Country, currency, geopolitical, and company-specific risks require careful consideration.

Listed infrastructure offers exposure to utilities, communications networks, power systems, and data centers supporting the digital economy. Private markets provide access to AI-native software companies and specialized infrastructure unavailable in public markets. These investments introduce different risks, including regulation, construction costs, limited liquidity, and greater dependence on manager selection.

How to invest in AI while managing risk

Individual-company research remains important within a diversified strategy. Investors need to assess customer demand, competitive advantages, revenue growth, profit margins, free cash flow, and balance-sheet strength. For the largest capital spenders, this assessment includes whether their established businesses generate enough cash to support AI investment without placing undue pressure on borrowing or shareholder distributions.

Valuation also influences future returns. A strong company can produce disappointing investment results if its stock price already assumes years of exceptional growth. Elevated expectations increase sensitivity to spending delays, weaker margins, new competition, and slower customer adoption.

Recent volatility does not invalidate the long-term AI investment case. Cloud demand, business adoption, and infrastructure investment continue to support the theme, but political constraints, competition, and high expectations leave less room for disappointing results. These conditions favor selective participation rather than an all-or-nothing judgment about artificial intelligence.

Diversification can reduce reliance on one AI winner

A diversified portfolio combines established technology companies, businesses supporting the physical buildout, and companies applying AI to improve productivity. This approach recognizes that the economic benefits of artificial intelligence can spread widely even if today’s market leaders change. It also reduces reliance on one company, processor, financing network, or development timeline.

Investors need to align AI exposure with their time horizon, tolerance for market volatility, and existing concentration in technology stocks. Talk with a wealth professional if you have questions about technology sector investments, your personal financial circumstances, or investment portfolio.

FAQs

Why do markets sometimes drop quickly after AI headlines?

When prices are already high, markets can react sharply because investors have less patience for uncertainty. In that environment, new headlines can trigger a “sell first, analyze later” move as people reduce risk quickly and reassess once more details emerge. Elevated valuations can add to that sensitivity, especially when the payoff from large AI spending remains hard to measure in real time.

Does the early 2026 pullback mean AI was a “bubble”?

A bubble typically lifts nearly anything connected to a theme regardless of fundamental merits, and recent market performance has not reflected that phenomenon. Instead, investors have separated stronger business models from riskier ones, and borrowing costs reflect that differentiation for lower-quality technology borrowers. You can interpret this as repricing and sorting, not a blanket rejection of AI’s longer-term potential.

What areas can matter even if the “winners” change over time?

AI adoption depends on dependable building blocks – trusted data, secure systems, and reliable infrastructure – not just one product or one company. Data capture, storage, processing, analytics, security, and electrification support many AI outcomes because they make tools usable at scale. Privacy, governance, and compliance also matter because many organizations require clear guardrails before they expand new technology across sensitive workflows.