Investing in AI is not the same as investing in technology. AI is a platform shift — comparable in scope to the internet in the 1990s — and the companies that will generate the most value over the next decade are not necessarily the ones generating the most headlines today. Here is what the analysts and researchers are actually saying, stripped of hype and presented with the caveats that responsible coverage requires.
Important disclaimer: Nothing in this article constitutes investment advice. All investments carry risk. Past performance does not guarantee future results. The information below is for educational purposes only. Consult a qualified financial advisor before making any investment decision.
Goldman Sachs Research published a framework in 2023 that has become the standard way analysts think about AI investment opportunities. The framework identifies three layers: infrastructure (the companies building the physical and software foundation for AI), enablers (the companies providing the tools and platforms that allow AI to be built and deployed), and adopters (the companies in traditional industries that are using AI to transform their operations and competitive position).
Each layer has a different risk and return profile. Infrastructure companies — particularly semiconductor manufacturers — have seen the most dramatic near-term gains. Enabler companies are in a more competitive phase. Adopter companies represent the longest-term opportunity but require the most patience.
"We estimate that AI investment could approach $200 billion globally by 2025. The key question for investors is not whether AI will be transformative — it will be — but which companies will capture the most value from that transformation."
— Goldman Sachs Research, "Generative AI: Too Much Spend, Too Little Benefit?" June 2024
NVIDIA remains the dominant company in AI infrastructure. Its H100 and H200 GPU chips are the primary hardware used to train large AI models, and demand has consistently exceeded supply. NVIDIA's data center revenue grew 409% year-over-year in fiscal 2024. The company's competitive moat — its CUDA software ecosystem, which most AI developers have built their workflows around — is considered by most analysts to be the strongest in the industry.
The risk: NVIDIA trades at a significant premium to historical technology valuations. AMD is a credible competitor in AI chips. Custom silicon from Google (TPUs), Amazon (Trainium), and Microsoft (Maia) could reduce dependence on NVIDIA over time. Most analysts maintain a positive outlook but note that much of the near-term growth is already priced in.
Other infrastructure companies analysts are watching: TSMC (the manufacturer of most advanced AI chips), Broadcom (custom AI chip design), and the data center REITs (Equinix, Digital Realty) that provide the physical infrastructure for AI computing.
Microsoft is the most widely cited AI enabler play among institutional analysts. Its $13 billion investment in OpenAI, the integration of AI across its Office 365 and Azure product suites, and its early-mover advantage in enterprise AI adoption have made it the default choice for analysts seeking AI exposure with lower volatility than pure-play AI companies. Microsoft's Copilot is now embedded in products used by hundreds of millions of people.
Alphabet (Google) is the other major platform company with significant AI exposure. Google DeepMind is considered by many researchers to be the leading AI research organization in the world. Google's search business — which generates the majority of Alphabet's revenue — faces genuine disruption risk from AI-powered search alternatives, but the company's own AI capabilities and its cloud business (Google Cloud) represent significant upside.
Amazon Web Services (AWS) is the largest cloud computing provider and a major beneficiary of AI infrastructure spending. Amazon's own AI models (Claude is developed by Anthropic, in which Amazon has invested $4 billion) and its AI services on AWS make it a significant player in the enabler layer.
The adopter layer is where analysts see the most underappreciated long-term opportunity. Companies in healthcare, financial services, legal services, and education that successfully integrate AI into their operations could see dramatic improvements in productivity and competitive position. The challenge is identifying which companies will execute successfully — a judgment that requires deep industry knowledge.
Healthcare AI is a particular area of analyst interest. Companies developing AI-powered diagnostic tools, drug discovery platforms, and clinical workflow automation are attracting significant investment. The FDA approved more AI-enabled medical devices in 2024 than in all previous years combined.
For investors who prefer diversification over individual stock selection, several ETFs provide exposure to the AI economy:
The Global X Robotics & Artificial Intelligence ETF (BOTZ) holds companies involved in the development and use of robotics and AI, including NVIDIA, Intuitive Surgical, and Keyence. The iShares Robotics and Artificial Intelligence Multisector ETF (IRBO) provides broader exposure across the AI value chain. The Roundhill Generative AI & Technology ETF (CHAT) focuses specifically on generative AI companies.
ETFs reduce individual company risk but also cap upside. The choice between ETFs and individual stocks depends on your risk tolerance, time horizon, and investment knowledge.
Not all analysts are bullish. Goldman Sachs published a notable report in June 2024 titled "Generative AI: Too Much Spend, Too Little Benefit?" that questioned whether the $1 trillion in projected AI infrastructure spending would generate sufficient returns. The report cited the difficulty of identifying specific AI use cases that justify the investment, the long time horizon for AI to generate measurable productivity gains, and the historical pattern of technology investment bubbles.
Sequoia Capital published a similar analysis in 2023 estimating that AI companies would need to generate $600 billion in annual revenue just to justify the infrastructure spending — a figure that would require capturing a significant portion of the entire software industry.
These are legitimate concerns. The honest answer is that no one knows which AI companies will generate the most value over the next decade. The analysts who are most credible are those who acknowledge this uncertainty while identifying the structural factors — competitive moats, switching costs, network effects — that make certain companies more likely to capture value than others.
Get the 30-day AI Investor action plan.
Specific daily steps to research, evaluate, and position yourself in the AI economy. Free. Not investment advice.
SOURCES