Nearly nine in ten enterprises across Asia Pacific have limited or redesigned their AI use cases because of rising token consumption, according to a new report from Cloudera, as rising costs and governance demands reshape how organisations scale enterprise AI.
Key Takeaways
- Cloudera’s State of Enterprise AI 2026 report finds 89% of enterprises in Asia Pacific have limited or redesigned AI use cases due to rising token consumption costs.
- While 58% of APAC respondents increased AI investment over the past year, only 28% say AI currently delivers more value than it costs.
- 79% of APAC respondents say data residency or sovereignty requirements influence where their AI workloads run, as Singapore commits S$500 million to AI chips and compute.
The State of Enterprise AI 2026: The New Rules of Scale report is based on a survey commissioned by Cloudera and fielded by Wakefield Research among 1,500 global IT leaders and decision-makers in September 2026. It finds that rising costs, governance demands and sovereignty requirements are reshaping how and where enterprises run AI, with 84% of global respondents and 79% across APAC saying residency or jurisdiction-specific requirements influence workload placement.
Investment Is Rising Faster Than Returns
Across Asia Pacific, rising AI investment has yet to translate into stronger returns for many organisations. While 58% of regional respondents report increased AI investment over the past year, only 28% say AI currently generates more value than it costs, and 21% say costs now outweigh the value delivered.
Globally, 98% of respondents have pulled back an AI deployment due to failure, hallucination or unexpected behaviour, while 45% have moved AI workloads in response to higher or more volatile costs. In APAC, 26% have gone further, pausing or cancelling an AI use case outright because of cost volatility.
Governance Is Now a Scaling Constraint
Governance and compliance have emerged as a major brake on scale: 75% of global respondents and 73% across APAC cite governance as a constraint, while 55% globally and 56% in APAC report that governance and compliance costs have grown faster than expected.
“Scale is exposing a new reality for enterprise AI: what works today may not make economic or operational sense tomorrow,” said Brian Martin, CTO of Applied AI at Cloudera. “Costs change, models evolve, regulations shift, and the right place to run a workload changes with them.”
Singapore’s S$500 Million Bet Meets a Harder Question
The findings land as governments across the region ramp up AI ambitions, with Singapore committing S$500 million to secure advanced AI chips and computing resources. Cloudera’s data suggests the regional conversation is shifting from simply accelerating adoption to ensuring AI delivers measurable business value.
“Enterprises across Asia Pacific face growing challenges in delivering measurable business value from their AI investments while navigating rising costs and increasingly complex governance and regulatory requirements,” said Remus Lim, Senior Vice President, Asia Pacific and Japan at Cloudera. If given the chance to restart their AI journey, 49% of global respondents say they would strengthen governance from the beginning, while 47% of APAC respondents identify skills and expertise as a constraint on scaling AI further.
