While we may be in the early stages of artificial intelligence (AI) development and usage, we believe AI has the potential to transform the global economy across various sectors. From increased productivity to software development to content creation, generative artificial intelligence (GenAI) is ushering in a new era of rapid advancement.

In our first installment of Navigating the Boom, where we confront GenAI’s most pressing questions, William Blair equity research analysts explore the economic impact of AI and where the most untapped opportunities are for investors.

Are We in an AI Bubble?

This question, more than any other, has dominated our conversations with investors, especially given massive capex investments tied to AI buildouts, eye-popping valuations for AI companies, and incessant media attention on the potential disruptive impact of AI on the economy and society. While it’s difficult to answer this question with certainty, our view is that we’re still in the early innings of the AI infrastructure buildout, and the potential for immense economic value creation from AI applications is real.

We base this view on the groundbreaking functionality of GenAI models, the fact that successive model releases show material improvements, and the timeline required to build the necessary infrastructure, data centers, and energy production to support market demand. This sustained growth is supported by commentary from both hyperscalers, which are large data centers that provide data management services, and chip vendors. As we enter 2025, we’re only starting to see 100,000+ graphics processing unit (GPU) centers move into production, with the biggest buyers in the world planning their AI buildouts to extend for at least two to three years. Additionally, President Trump’s administration announced its backing of Project Stargate, a new AI venture seeking to secure leadership in AI, create jobs, and protect the security of the U.S. and its allies. That said, we believe investors could play the long game with GenAI, with the potential payoff likely much greater than the invested capital.

If You Build It, Will They Come?

To assess whether we’re destined for a major supply-demand imbalance in the next several years, many wonder if the amount spent on building the models and data centers is worth the investment and revenue potential. A cumulative $1.6 trillion in capex is expected to be spent from 2023-2028 by the largest hyperscalers in the world. William Blair analysts also question whether “AI revenue” is the only framework to think about in terms of return on invested capital. If AI becomes integrated into most—or all—applications over time, what differentiates AI revenue from non-AI revenue? Or, if AI results in material margin expansion for businesses across many sectors due to changing workforces, how would that be accounted for if revenue is the main return framework? The GenAI space feels more like a marathon than a sprint, and returns on current investments will likely take years to play out. While near-term skepticism around the breakneck level of capex is warranted, we think investors should not miss the forest for the trees.

The Most Untapped Opportunities for Investors

In the near-term, William Blair analysts expect the infrastructure/compute/model layers of AI will continue to accrue the most value in the GenAI space, as GPU clusters scale ever larger to support the pretraining of frontier AI models and as post-training and inference become more significant. As AI applications start to take greater hold, we see a few different areas of budding value creation:

  • Compute-adjacent infrastructure components, such as storage and memory
  • Infrastructure software, including databases, data streaming systems, management platforms, and data labeling, governance, and security tools
  • Energy production to power data centers and related equipment

Most of today's value generation comes from chips, hardware systems, and the physical infrastructure required to power data. As we continue to understand the developmental phases of GenAI, William Blair analysts will continue to share updates on critical technological insights on this topic and more.

For more information about William Blair’s technology, media and communications research, please contact us or your William Blair representative.