The AI ROI Crisis: Why 67% of Enterprise AI Projects Miss Financial Targets

SAN ANTONIO, TX – November 12, 2026 — Corporate boards have authorized billions in capital expenditures to capture the generational promise of artificial intelligence. Yet, as the initial wave of enterprise AI matures from isolated pilot programs into full-scale production, a sobering financial reality is taking hold. According to a sweeping new study, the vast majority of these initiatives are failing to deliver their promised returns, bogged down by fractured vendor ecosystems and alarming gaps in corporate governance.

A joint report released today by HFS Research and Rackspace Technology reveals that only 33% of enterprises have met or exceeded the expected business case for their most significant AI deployment. That figure drops even further—to a mere 28%—for organizations attempting to implement the technology across multiple business functions. The study, which surveyed 502 senior AI decision-makers at companies with at least $100 million in annual revenue, exposes a critical inflection point in the market: the transition from experimental AI to operational AI is breaking traditional IT management models.

For years, the technology sector has operated on the assumption that accumulating the best individual components—the most advanced large language models, the fastest compute infrastructure, and the most robust data lakes—would naturally aggregate into compounding business value. The data suggests the exact opposite is happening.

The Accountability Void in the Modern Tech Stack

The central obstacle identified by researchers isn't a lack of computing power or algorithmic sophistication. It is a profound lack of operational accountability. As companies have rushed to build their AI capabilities, they have inadvertently constructed sprawling, multi-vendor architectures where no single entity is responsible for the system's ultimate performance, cost, or security.

According to the findings, 44% of enterprises currently lack a single, consistent framework for running AI models in production. When deployments fail to generate revenue or begin to hallucinate, the resulting finger-pointing between cloud providers, model developers, and internal data teams leaves Chief Information Officers holding the bag.

“Enterprise AI has an accountability problem that nobody in the industry wants to admit,” said Phil Fersht, Founder and CEO of HFS Research. “We’ve built an enormous ecosystem of companies selling models, platforms, infrastructure and consulting, yet when the promised results don’t materialize, everyone points at someone else and the enterprise is left trying to make the pieces work together. Only a third of enterprises have met the business case for their most significant AI program, which should force us to confront how addicted we’ve become to selling AI rather than taking responsibility for making it work.”

This disillusionment aligns with broader macroeconomic indicators. Recent tracking by global research firms like Gartner and McKinsey indicates that while AI remains a top strategic priority, the technology is firmly entrenched in a trough of disillusionment. Top performers capturing significant earnings growth from AI remain a distinct minority, while the broader market struggles with integration complexities and talent shortages.

The 96% Risk: Governance Deficits Freezing Scale

While fractured operations are bleeding ROI, governance failures are posing an existential threat to AI scaling. A staggering 96% of surveyed enterprises reported significant AI governance gaps. Furthermore, 78% of executives now view governance failures as a greater threat to scaling AI than technological limitations.

In my conversations with Chief Information Security Officers across the Fortune 500, the panic surrounding shadow AI is palpable. Business units, eager to boost productivity, are frequently adopting generative AI tools without central IT oversight. This creates massive blind spots regarding data lineage and prompt engineering vulnerabilities. Employees interacting with public or poorly secured internal models can inadvertently leak proprietary source code, financial projections, or protected customer data.

Regulatory pressures are compounding this operational anxiety. The rollout of the European Union’s AI Act has fundamentally altered the risk calculus for global enterprises. The legislation imposes strict requirements for data quality, human oversight, and explainability on high-risk systems, backed by crippling financial penalties for non-compliance. Similarly, the NIST AI Risk Management Framework in the United States demands rigorous structures for fairness, transparency, and security that most decentralized corporate AI deployments simply cannot meet.

When multi-vendor sprawl meets strict regulatory mandates, the result is paralysis. Enterprises cannot confidently scale systems they cannot fully audit, explain, or control.

The Managed Services Pivot and the Premium for Control

Where enterprises see operational friction and regulatory risk, IT vendors see a lucrative commercial opportunity. The HFS Research and Rackspace study highlights a massive shift in buyer appetite: 73% of enterprises stated they would pay a premium over raw infrastructure costs for fully managed AI services that carry explicit governance accountability. Furthermore, 61% prefer a single accountable partner across models, platforms, and infrastructure—a figure that jumps to 74% among firms whose production-stage programs have failed to deliver.

This data neatly validates the strategic pivot underway at infrastructure providers like Rackspace. Long known for multicloud hosting, the company is aggressively repositioning itself to capture this higher-margin enterprise demand. To close the accountability gap, the report advocates for a Services-as-Software Operator—a managed services partner or dedicated internal group solely accountable for the day-to-day performance, economics, and governance of live AI environments.

“Turning AI investment into measurable business value requires clear operational accountability,” said Gajen Kandiah, CEO of Rackspace. “As enterprises move from pilots into production, infrastructure readiness, security, governance and cost have to be managed together. Customers need an operator accountable for integrating, running and continuously improving the environment across infrastructure, applications, data and AI.”

This is not merely an altruistic push for better IT management; it is a calculated land grab. By offering to serve as the single throat to choke for AI performance and compliance, managed service providers can lock in long-term, high-value contracts that far exceed the margins of raw compute reselling. They are betting that corporate boards will gladly pay a premium to offload the legal and operational liabilities of their AI ecosystems.

Efficiency Over Growth: The Current State of AI Value

Until this new operational model matures, the actual value extracted from enterprise AI remains heavily skewed toward back-office optimization rather than top-line growth. The study found that most realized gains are concentrated in efficiency metrics: 58% of respondents cite workforce productivity improvements, and 42% report measurable cost reductions.

In stark contrast, only 28% have seen better customer experiences, and a mere 21% can point to measurable revenue growth directly attributable to their AI investments.

These figures paint a clear picture of the current enterprise AI landscape. The technology is undeniably making workers faster and trimming operational fat, but it is not yet the revenue-generating engine that vendors promised. To cross that chasm, organizations must stop treating AI as a collection of isolated software purchases and start managing it as a unified, tightly governed corporate asset. The era of buying fragmented AI parts and hoping they magically aggregate into business value is decisively over.