Cloud waste hit 29% of estimated cloud spend in 2026, the first increase in five years, per Tech-Insider. This financial drain is pervasive; 97% of enterprise leaders admit to squandering a portion of their public cloud budget. Such inefficiency directly impacts operational budgets and resource allocation.
Yet, Generative AI adoption as a cloud service is rapidly increasing, with 58% of organizations utilizing it, up from 50% the previous year, Tech-Insider.org reports. Generative AI adoption, rapidly increasing with 58% of organizations utilizing it, clashes directly with the struggle against massive public cloud waste, compelling organizations to repatriate workloads even as they embrace new cloud-based AI.
Consequently, companies prioritize cost control and strategic infrastructure ownership. The prioritization of cost control and strategic infrastructure ownership drives a significant shift towards hybrid cloud models and fuels a boom for specialized AI hardware manufacturers, while public cloud providers must justify their value beyond raw compute.
The Mounting Cost Burden
- 29% — Cloud waste reached this percentage of estimated cloud spend in 2026, marking the first increase in five years, according to Tech-Insider.
- 97% — Enterprise leaders admit to wasting a portion of their public cloud budget, as reported by Tech-Insider.org.
- 63% — FinOps adoption has grown to this percentage of enterprises, according to Tech-Insider.org.
- 30% — Globally, over 30% of survey respondents cite cost management as the number one public cloud challenge, according to Broadcom.
- 57% — In APJ, this percentage of organizations experience public cloud waste exceeding a quarter of their entire cloud budget, reports Broadcom.
- 17% — Organizations exceeded their public cloud budgets over the past year, according to Tech-Insider.org.
These figures collectively reveal a systemic challenge: despite growing FinOps adoption (63%), enterprises globally (30% citing cost as top challenge) and particularly in APJ (57% with over a quarter of budget wasted) consistently struggle with public cloud expenditures. The 17% exceeding budgets underscores that current strategies fall short, turning public cloud into a significant financial drain rather than a predictable utility.
The Repatriation Pivot
| Metric | Previous Trend (Implied) | Current State (2026) | Trend |
|---|---|---|---|
| Organizations considering repatriation (APJ) | Not specified | 82% | Strong Consideration |
| Organizations already repatriated (APJ) | Not specified | 54% | Significant Action |
| FinOps Adoption (Enterprises) | Growing | 63% | Increased Adoption |
| Cloud Waste (% of spend) | Lower than 29% (5-year low) | 29% | Rising |
Data derived from Broadcom and Tech-Insider.
The data confirms a decisive shift: 82% of APJ organizations are considering repatriation, with 54% already taking action. The strong movement of 82% of APJ organizations considering repatriation and 54% already taking action, alongside increasing FinOps adoption (63%), indicates a strategic pivot towards greater control and cost optimization. Enterprises are actively disengaging from public cloud for stable workloads, seeking more predictable and manageable infrastructure amid rising costs and persistent waste.
AI's Dual Impact: Driving Demand and Specialization
Generative AI's rapid adoption, now a cloud service for 58% of organizations (up from 50%), drives demand for innovative public cloud offerings. However, this surge simultaneously highlights the need for specialized, high-performance infrastructure. Manufacturers like Sanmina are responding with strategic capital investments in AI racks and printed circuit boards, StockStory reports. Strategic capital investments by manufacturers like Sanmina in AI racks and printed circuit boards indicate a market recognition that compute-intensive AI workloads often find more cost-effective and controllable homes in hybrid or on-premise environments, rather than exclusively in hyperscale public clouds.
Market Winners and Emerging Challenges
Sanmina exemplifies the direct beneficiaries of this market shift. The company reported Q2 CY2026 revenue of $3.46 billion, a 69.7% year-on-year increase, and non-GAAP profit per share of $3.31, exceeding analyst estimates by 18.5%, StockStory confirms. These robust figures validate the strategic pivot towards specialized hybrid cloud AI integration, demonstrating that demand for underlying hardware infrastructure surges as organizations prioritize controllable, cost-effective environments for AI workloads.
Navigating the Hybrid Future
Companies are trading the perceived agility of public cloud for the tangible cost control of private infrastructure, with 82% of APJ organizations considering repatriation, indicating a fundamental re-evaluation of cloud economics beyond initial adoption hype.
- The 82% of APJ organizations considering workload repatriation, per Broadcom, underscores a market-wide reassessment of cloud value.
- The simultaneous rise in public cloud waste to 29% and rapid Generative AI adoption means enterprises effectively subsidize innovation with operational inefficiencies, demanding a strategic approach to workload placement, Tech-Insider.org reports.
- Strategic capital investments by hardware manufacturers like Sanmina in AI infrastructure signal a coming wave of on-premise and hybrid cloud build-outs. Organizations recognize that compute-intensive workloads are often more cost-effective and controllable outside of hyperscale public clouds.
Even for market beneficiaries, Sanmina's Q3 CY2026 revenue guidance of $3.45 billion, 2.1% below analyst estimates, suggests a dynamic and competitive landscape. Continuous adaptation and strategic investment will be crucial for sustained success in the evolving hybrid cloud and AI infrastructure market.
By Q3 2026, companies failing to optimize their cloud environments will likely face continued budget overruns, impacting their ability to fund critical innovation and infrastructure projects.
