The public cloud market continues its explosive growth trajectory, with enterprises rushing to their cloud consoles to allocate more resources, particularly for AI initiatives. Cloud providers are falling over themselves to promote their latest AI capabilities, posting numerous job requisitions (many unfunded “ghost jobs”) and offering generous credits to entice enterprise adoption. However, beneath this veneer of enthusiasm lies a troubling reality that few are willing to discuss openly.
The statistics tell a sobering story: Gartner estimates that 85% of AI implementations fail to meet expectations or aren’t completed. I consistently witness projects begin with great fanfare, only to fade into obscurity quietly. Companies excel at spending money but struggle to build and deploy AI effectively.
How strong is demand for AI really?
There’s a puzzling disconnect in the cloud computing industry today. Cloud providers consistently claim they’re struggling to meet the overwhelming demand for AI computing resources, citing waiting lists for GPU access and the need for massive infrastructure expansion. Yet their quarterly earnings reports often fall short of Wall Street’s expectations, creating a curious paradox.
The providers are simultaneously announcing unprecedented capital expenditures for AI infrastructure. Some are planning 40% or higher increases in their capital budgets even as they seem to struggle to demonstrate proportional revenue growth.
Investors’ fundamental concern is that AI remains an expensive research project, and there’s significant uncertainty about how the global economy will absorb, utilize, and pay for these capabilities at scale. Cloud providers may conflate potential future demand with current market reality, leading to a mismatch between infrastructure investments and immediate revenue generation.
This suggests that although AI’s long-term potential is significant, the short-term market dynamics may be more complex than providers’ public statements indicate.
The ROI conundrum
Data quality is perhaps the most significant barrier to successful AI implementation. As organizations venture into more complex AI applications, particularly generative AI, the demand for tailored, high-quality data sets has exposed serious deficiencies in existing enterprise data infrastructure. Most enterprises knew their data wasn’t perfect, but they didn’t realize just how bad it was until AI projects began failing. For years, they’ve avoided addressing these fundamental data issues, accumulating technical debt that now threatens to derail their AI ambitions.
Leadership hesitation compounds these challenges. Many enterprises are abandoning generative AI initiatives because the data problems are too expensive to fix. CIOs, increasingly concerned about their careers, are reluctant to take on these projects without a clear path to success. This creates a cyclical problem where lack of investment leads to continued failure, further reinforcing leadership’s unwillingness.
Return on investment has been dramatically slower than anticipated, creating a significant gap between AI’s potential and practical implementation. Organizations are being forced to carefully assess the foundational elements necessary for AI success, including robust data governance and strategic planning. Unfortunately, too many enterprises consider these things too expensive or risky.
Sensing this hesitation, cloud providers are responding with increasingly aggressive marketing and incentive programs. Free credits, extended trials, and promises of easy implementation abound. However, these tactics often mask the real issues. Some providers are even creating artificial demand signals by posting numerous AI-related job openings, many of which are unfunded, to create the impression of rapid adoption and success.
Another critical factor slowing adoption is the severe shortage of skilled professionals who can effectively implement and manage AI systems. Enterprises are discovering that traditional IT teams lack the specialized knowledge needed for successful AI deployment. Although cloud providers do offer various tools and platforms, the expertise gap remains a significant barrier.
This situation will likely create a stark divide between AI “haves” and “have-nots.” Organizations that successfully organize their data and effectively implement AI will use generative AI as a strategic differentiator to advance their business. Others will fall behind, creating a competitive gap that may be difficult to close.
A strategic path for adoption
Enterprise leaders must move away from the current pattern of rushed, poorly planned AI implementations. The path to success isn’t chasing every new AI capability or burning through cloud credits. Indeed, it’s through thoughtful, strategic development.
Start by getting your data house in order. Without clean, well-organized data, even the most sophisticated AI tools will fail to deliver value. This means investing in proper data governance and quality control measures before diving into AI projects.
Build expertise from within. Cloud providers offer powerful tools, but your team needs to understand how to apply them effectively to your business challenges. Invest in training your existing staff and strategically hire AI specialists who can bridge the gap between technology and business outcomes.
Begin with small, focused projects that address specific business problems. Prove the value through controlled experiments before scaling up. This approach helps build confidence, develop internal capabilities, and demonstrate tangible ROI.
The road ahead for cloud-based AI
Cloud providers will continue to grow in the coming years, but their market could contract unless they can help their customers develop AI strategies that overcome the current high failure rates. The reasons enterprises struggle with generative AI, agentic AI, and project failures are well understood. This isn’t a mystery to analysts and CTOs. Yet enterprises seem unwilling or unable to invest in solutions.
The gap between AI supply and demand will eventually close, but it will take significantly longer than cloud providers and their marketing teams suggest. Organizations that take a measured approach of thoughtful planning and building proper foundations may move more slowly initially, but will ultimately be more successful in their AI implementations and realize better returns on their investments.
As we move forward, cloud providers and enterprises must align their expectations with reality and focus on building sustainable, practical AI implementations rather than chasing the latest hype cycle. I hope that enterprises and cloud providers both can get what they are looking for; it should be the same thing—right?