georgeclerk/E+/Getty Images
Follow ZDNET:
Add us as a preferred source on Google.

ZDNET’s key takeaways
Boards of directors are beginning to scrutinize artificial intelligence spending more rigorously than ever before. These inquiries into the fiscal prudence and measurable outcomes of AI initiatives present a timely chance for strategic recalibration. The moment calls for renewed focus on building organizational capacity, nurturing robust relationships with technology partners, and pursuing co-development approaches that yield sustainable value.

Despite the growing skepticism, capital flowing into AI remains immense. The financial commitment from enterprises shows no indication of slowing down. According to analyst firm Gartner, the worldwide expenditure on AI is projected to reach an astonishing $2.52 trillion by 2026 — representing a striking 44% annual increase. Yet, this surge in investment is unfolding against a backdrop of mounting doubt. In Gartner’s Hype Cycle for Emerging Technologies, AI has begun to descend into what analysts call the Trough of Disillusionment — a period where lofty expectations meet economic reality. This shift has prompted board members to demand justification for every dollar invested, while expecting business leaders to turn theoretical promise into quantifiable, bottom-line impact.

As ZDNET previously noted, numerous segments of AI have already entered this phase, where enthusiasm wanes after repeated failures to achieve expected returns. Generative AI, once the darling of technological conversations, is now emblematic of this decline. Its early promise of transformative change has begun to collide with the sobering truth of incomplete integrations and elusive profitability. Although it may appear that the optimism surrounding generative AI has evaporated and the bubble may be on the verge of bursting, some experts view this moment less as a crisis and more as an inflection point.

John-David Lovelock, Gartner’s chief forecaster and distinguished vice president analyst, views this downturn as a vital opportunity rather than a setback. In an interview with ZDNET, he emphasized that the so-called ‘trough’ provides a necessary pause—a moment in which enterprises can reconsider their investment strategies, prioritize realism over spectacle, and assess which AI applications truly support their mission-critical objectives. In his words, companies should actually welcome this descent because it restores rational expectations. For two years, AI has often been entangled with overly ambitious, “moonshot” ventures that promised revolutionary results but lacked solid grounding. With research from MIT indicating that approximately 95% of generative AI projects fail to yield tangible business value, Lovelock advocates for a methodical recalibration aimed at focusing resources where they can truly make an impact.

To guide organizations through this period of reflection and renewal, he outlines three key priorities that should shape AI investment strategies through 2026.

1. **Focus on capacity building**
One of the most pressing imperatives is to strengthen the fundamental infrastructure underpinning AI operations. Gartner’s analysis forecasts that expanding AI infrastructure will be a defining characteristic of emerging technology investments during this period. Merely constructing the essential AI foundations is expected to drive a 49% uptick in spending on AI-optimized servers — already comprising 17% of total AI expenditure this year. Cumulatively, investments in AI infrastructure are projected to reach $401 billion by 2026 as technology providers race to develop platforms capable of training and running complex models at scale.

Lovelock stresses that such foundational capacity building remains crucial, even as the market enters the Trough of Disillusionment. The hyperscalers, cloud giants, and software vendors currently investing in advanced data centers are laying the groundwork for the next generation of AI functions, including the training and deployment of powerful models and intelligent agents. For example, a financial institution seeking to automate credit card approvals must determine where and how to allocate its compute capacity—whether by running its own infrastructure, engaging a major provider like AWS, Microsoft, or Google, leveraging a managed platform, or connecting via an API to a specialist service such as OpenAI. The essential decision lies in aligning infrastructure investments with organizational strategy—assessing how much technology to build, buy, or co-own, and how deeply to embed it into proprietary systems.

2. **Create strong partnerships**
Successfully navigating these choices depends on cultivating trusted partnerships. Lovelock emphasizes that meaningful collaborations between enterprises and technology providers will serve as the backbone of AI success in the coming years. For most organizations, innovation will stem not from internal moonshot projects but from leveraging the capabilities of established technology partners. Visionary companies may push boundaries with in-house development, but the vast majority should focus on applying proven technologies through reliable collaborations.

As AI stabilizes within the Trough of Disillusionment, it will increasingly be distributed through incumbent software ecosystems rather than through speculative experimental ventures. This dynamic means that many businesses will achieve their most productive outcomes by strengthening their relationships across their digital and data stacks rather than starting from scratch. Lovelock reinforces this idea by stating that identifying and aligning with the right technology allies—those who can guide an organization along its chosen path—is central to bridging the gap between aspiration and execution, whether the goal is simply to enhance existing capabilities or to eventually operate as a fully autonomous enterprise.

3. **Avoid random explorations**
Finally, Lovelock cautions against undirected experimentation. In a period when enthusiasm outpaces clear direction, it becomes easy to scatter investments across too many initiatives without strategic cohesion. Gartner advises digital leaders to refrain from impulsive or generalized explorations into emerging technology and instead ensure that only their most promising, carefully governed projects progress toward meaningful outcomes.

To transform pilot projects into sources of measurable value, Lovelock recommends focusing on three interconnected pillars: reliable partners, high-quality data, and well-defined processes. Equally vital is stakeholder alignment; success depends on how effectively executives, managers, and business units collaborate to define shared outcomes and sustain commitment from concept through to implementation. He further notes that organizations thrive when their providers share in that commitment. Structuring relationships around value-based or outcome-based pricing models helps ensure that vendors are directly invested in their client’s success. Co-development represents the deepest and most mutually beneficial form of partnership, though it also requires strong governance and sustained trust.

While such alignment takes time and organizational effort, Lovelock concludes that when executed successfully, these partnerships yield enduring mutual rewards. By tying a provider’s compensation directly to the client’s success, both parties move beyond transactional interactions toward genuine collaboration capable of transforming “moonshot” experiments into practical, high-impact production systems. In essence, the current downturn in AI enthusiasm is not a warning of failure—but an invitation to build maturity, sharpen focus, and turn visionary rhetoric into measurable returns.

Sourse: https://www.zdnet.com/article/how-to-target-your-ai-investments-gartner/