Gartner’s 2024 Technology Hype Cycle—which outlines the dynamics of expectations surrounding technological innovations across five distinct phases—indicates that generative AI has surpassed the peak of inflated expectations. Nevertheless, the associated hype remains persistent (see figure below), and the technology continues to hold the potential to become truly transformative, revolutionising content discovery and creation, automating human tasks, and reshaping both customer and employee experiences (Jaffri 2024).
Although the Hype Cycle captures the evolution of expectations and public attention, it is closely linked to the S-curve of technological adoption (diffusion model), as both offer complementary perspectives on the development and spread of innovation (Shi & Herniman 2023). While the Hype Cycle focuses primarily on the trajectory of expectations and media visibility, the S-curve illustrates technological maturity and the actual pace of adoption—depicting how a given technology becomes widespread and productive over time (Dedehayir & Steinert 2016, 2). The relationship between these two models is crucial: the peak of inflated expectations and the subsequent trough of disillusionment observed in the Hype Cycle often serve as precursors to the slow, then accelerating adoption characteristic of the early stages of the S-curve.
Generative AI is currently situated in the trough of disillusionment phase of the Gartner Hype Cycle (Gartner 2024) and in the early adopter stage of the S-curve, where practical implementations remain limited, yet the technology’s value is increasingly apparent. As expectations consolidate and productivity benefits become more tangible, the technology may soon enter the steep growth phase of the S-curve, facilitating broader adoption and generating wide-ranging socio-economic impacts.
Today, scarcely a day passes without some form of technical breakthrough—whether in AI chips, large language models, or video generation—accompanied by a steady stream of announcements from promising AI startups. However, since the burst of the dot-com bubble in the early 2000s (Goodnight & Green 2010), investors have approached emerging technologies with far greater caution—and rightly so, given that most such innovations have failed to stand the test of time (Maheshwari 2024). The excitement surrounding generative AI has also gripped the corporate sector. Yet beneath the surface, the actual results remain modest. Most companies are experimenting with small, exploratory teams and launching proof-of-concept projects, while larger firms enlist waves of consultants to develop first-generation applications. Only a small fraction of these pilot initiatives ever make it to full-scale deployment, prompting a critical question: is the enthusiasm surrounding generative AI merely another instance of hype, or does it reflect genuine, lasting value (Maheshwari 2024).
The contrast between the current generative AI boom and the early 2000s dot-com bubble is well illustrated by the examples of Cisco and NVIDIA (Hodges 2024). While the dot-com era was driven largely by speculation—with Cisco’s price-to-earnings (P/E) ratio peaking at 132, far exceeding its previous five-year average of 37—the present AI surge appears to rest on more stable foundations. In the case of NVIDIA, the current P/E ratio stands at 39, closely aligned with its 2019–2023 average of 40, suggesting a more measured and cautious investment approach (Coatue 2024). This relative balance indicates that the AI boom is less reliant on inflated expectations and places greater emphasis on practical applications and tangible benefits of the technology (see figure below).
According to Gartner’s projections (Gartner 2024), by the end of 2025 nearly 30% of generative AI projects are expected to stall after the proof-of-concept phase, primarily due to poor data quality, high implementation costs, and unclear business value (Aaron 2024). Ensuring the long-term success of the technology will therefore depend on the strategic prioritisation of use cases and the effective management of associated risks. Although early adopters have already realised notable business benefits—such as increased revenues and cost reductions—the true potential of generative AI can only become sustainable once it moves beyond the trough of disillusionment within the Hype Cycle.
References:
1. Coatue. 2024. ‘8th Annual East Meets West (EMW) Conference’. https://www.coatue.com/blog/company-update/coatues-2024-emw-conference – ^ Vissza
2. Dedehayir, Ozgur, and Martin Steinert. 2016. ‘The Hype Cycle Model: A Review and Future Directions’. Technological Forecasting and Social Change 108 (July): 28–41. doi:10.1016/j.techfore.2016.04.005 – ^ Vissza
3. Gartner. 2024. ‘Hype Cycle for Artificial Intelligence, 2024’. https://www.gartner.com/en/documents/5505695 – ^ Vissza
4. Goodnight, G. Thomas, and Sandy Green. 2010. ‘Rhetoric, Risk, and Markets: The Dot-Com Bubble’. Quarterly Journal of Speech 96 (2): 115–40. doi:10.1080/00335631003796669 – ^ Vissza
5. Hodges, Paul. 2024. ‘Stock Market Bubbles Follow the Same Pattern, as Nvidia and Cisco Confirm’. https://www.icis.com/chemicals-and-the-economy/2024/06/stock-market-bubbles-follow-the-same-pattern-as-nvidia-and-cisco-confirm/ – ^ Vissza
6. Jaffri, Afraz. 2024. ‘Explore Beyond GenAI on the 2024 Hype Cycle for Artificial Intelligence’. Gartner. https://www.gartner.com/en/articles/hype-cycle-for-artificial-intelligence – ^ Vissza
7. Maheshwari, Archit. 2024. ‘Is GenAI Hype Dying? – Why Now Is the Best Time to Build’. Artefact. https://www.artefact.com/blog/is-genai-hype-dying-why-now-is-the-best-time-to-build/ – ^ Vissza
8. Shi, Yuwei, and John Herniman. 2023. ‘The Role of Expectation in Innovation Evolution: Exploring Hype Cycles’. Technovation 119 (102459): 1–10. doi:10.1016/j.technovation.2022.102459 – ^ Vissza
9. Tan, Aaron. 2024. ‘Nearly a Third of GenAI Projects to Be Dropped after PoC’. https://www.computerweekly.com/news/366599232/Nearly-a-third-of-GenAI-projects-to-be-dropped-after-PoC – ^ Vissza
