Correction added 16 October 2026: The original article turned projects still in pilot or development into an eventual failure rate and a per-dollar loss probability. Those inferences are withdrawn. Respondent-level readiness is also kept separate from workforce readiness, with the original survey population and timing clarified.
Riverbed's manufacturing results released on 4 March 2026 report that 62% of AI projects were in pilot or development and 37% of manufacturing respondents felt fully prepared to operationalise AI at scale. These are survey findings, not a record of eventual project outcomes. The underlying research was conducted in July 2025 across multiple industries; the published release does not identify the manufacturing subsample size.
The findings raise questions about data infrastructure, governance and organisational readiness alongside spending. They do not establish which of those factors caused a particular project to remain in development.
The Paradox: More Money, Less Readiness
Of the manufacturing respondents, 90% considered better data quality critical to AI success, while 91% regarded data movement and sharing as important to their AI strategy. Those answers describe priorities. They do not establish that 91% faced a data-sharing obstacle, or that respondents were knowingly perpetuating one.
A project still in development may later succeed, change scope or stop. The 62% snapshot cannot be converted into an eventual failure rate, and a count of projects cannot establish a probability of losing each dollar invested. The original failure and per-dollar return inferences are withdrawn.
Riverbed also reported that 87% of manufacturing respondents said their AIOps returns met or exceeded expectations. Satisfaction with particular initiatives and readiness for enterprise-wide deployment are different measures. Neither establishes that enterprise AI generally fails, nor that every pilot delivers a return.
Why Capital Alone Cannot Close This Gap
The historical pattern is instructive. Manufacturers' digital transformation journey over the past decade consumed billions in capital while delivering mixed results. MIT research documents that manufacturers have completed less than 50% of their digitalization roadmap after ten years of investment, a timeline that suggests the obstacle is not money but orchestration.
The pattern repeats because the allocation pattern repeats. Capital flows toward what is visible and measurable: new platforms, licensing, compute infrastructure. It does not flow toward what is structural: data governance frameworks, cross-functional change management, or workforce reskilling. Gartner's 2026 forecast warns that 60% of agentic AI projects will fail specifically due to lack of AI-ready data, yet data readiness remains subordinate to technology acquisition in most manufacturing budgets.
The arithmetic of investment itself reveals the issue. Research from MIT Sloan shows that manufacturers allocate only 1.8% of net revenue to digital transformation on average. Organizations investing 3% or above achieve 2.5 times better ROI. Rather than increasing the allocation to digital infrastructure, manufacturers are maintaining the baseline while expanding scope, adding AI initiatives without proportional support for the foundational layer that makes those initiatives deployable.
The Three Gaps That Double Spending Cannot Close
Three areas warrant examination. The first is data infrastructure. The survey's 91% figure measures the reported importance of data movement and sharing, not the prevalence of a particular architecture. Manufacturers need to assess their own data quality, access and integration before deciding whether a data mesh or another approach fits their operations.
The second is organizational readiness. Workforce skills shortages outpace AI adoption velocity. A Deloitte and NVIDIA analysis found that 45% of G2000 original equipment manufacturers will connect field and engineering data via AI-assisted platforms by 2026, but only if workforce capacity exists to implement and sustain those connections. Training and hiring velocity has not matched AI ambition.
The third is measurement. Track project stage, operational use and financial returns separately. A readiness answer does not quantify realised returns, and satisfaction with one initiative does not establish that an enterprise is ready to scale others.
What This Means for Manufacturing Leadership
The immediate decision is not whether to invest in AI; that decision is made and doubled. The decision is where to redirect capital next. Audit enterprise AI deployment readiness before funding the next platform. Map the state of data governance across the organization. Measure not the number of AI projects initiated but the percentage of AI initiatives that have moved from pilot into standard operational practice.
Three moves follow. First, treat data infrastructure as a capital priority equal to AI technology. A data mesh architecture (distributed, federated, governed) takes 18-24 months to mature and requires organizational commitment. That timeline means the decision point is now, not after the next pilot cycle.
Second, distinguish results from individual AIOps deployments from evidence of enterprise readiness. Continue improving systems with demonstrated operational value, while assessing the architecture, governance and change management needed to extend them.
Third, assess workforce capabilities directly. The 37% readiness figure concerns manufacturing respondents' views of their organisations; it does not mean that 37% of one workforce is ready. Align hiring, reskilling and organisational design with identified deployment needs.
The manufacturers doubling investment and achieving high pilot ROI are positioned correctly in the near term. The manufacturers moving pilots into enterprise operations without addressing governance and infrastructure will discover that doubling capital solved the wrong problem.
Source: Study: Manufacturing Organizations Doubled AI Investment, Yet Only 37% Fully Prepared to Deploy AI Enterprise-Wide Category: Big Data & Analytics | Impact: high_impact | Pillar: Industry 4.0