Enterprise AI spending is climbing rapidly, with boards racing to deploy the technology faster than they can measure whether it works. According to the latest IDC InfoBrief, commissioned by Expereo*, around 70% of organizations are investing in AI, motivated by its potential or by the fear of falling behind the competition, but they lag in disciplined ROI evaluation, and one in five (20%) admit they are investing aggressively in AI with little evaluation, driven by the fear of being left behind.
In Asia Pacific (APAC), that pressure is even more pronounced as 37% of organizations admit investing aggressively with little evaluation – nearly double the global average, and well ahead of the US (10%) and Europe (13%). The pressure is most acute in Australia (45%) and Vietnam (44%), while in Singapore, more than one in three organizations admit the same.
The IDC InfoBrief, based on a survey of 800 technology leaders across APAC, Europe, and the US, found that AI has become one of the most prioritized technology investments globally, with 51% of organizations planning to prioritize AI or machine learning investment over the next 12 months – rising to 61% across APAC. However, returns are failing to keep pace with the hype. Just 19% of global organizations surveyed say their AI implementations have exceeded expectations, and only 5% report they have significantly exceeded them.
Across APAC, 40% say implementations have exceeded or significantly exceeded expectations – ahead of the global average but still leaving the majority falling short. Globally, the most-cited reasons for underperformance are inadequate or poor-quality training data (51%), higher-than-expected costs or ROI not achieved (47%), and AI not performing as well as expected (46%). For APAC specifically, the picture is broadly similar – though costs bite harder: 49% cite poor-quality training data, 54% cite cost overruns or ROI not achieved (rising to 80% in Malaysia), and 46% say AI has simply not performed as expected.
Where organizations have the right foundations in place, the results speak for themselves. Across APAC, 87% report productivity improvements in the business units most affected by AI, and 82% say quality of work has improved.
Underpinning many of these challenges is also a network and infrastructure readiness gap. Globally, 26% of organizations whose AI implementations have failed to meet expectations cite inadequate network or connectivity performance as a contributing factor. Looking ahead, 54% of organizations say they need more flexible and scalable networks to thrive in an AI-driven environment, and 51% need greater resilience and reliability to maximize uptime. In APAC, the gap is acute as only 9% of organizations describe their network infrastructure as fully ready to support new AI, cloud, and digital initiatives, and 37% say it will need upgrading or replacing soon. The need is most acute in Thailand (74%) and Singapore (58%), both of which rank above the regional average on demand for flexible, scalable networks. In Indonesia, nearly half of all organizations (48%) say their infrastructure will need upgrading or replacing soon.
Ben Elms, CEO, Expereo, says: “Every enterprise we speak to is investing in AI, yet the data shows a clear gap opening up between AI ambition and AI outcomes. More often than not, that gap comes down to the network underneath. AI only delivers on its promise when the infrastructure carrying it is built to support it.
Without resilient, scalable, cloud-optimized networks, even the most well-funded AI programs will struggle to deliver ROI. Getting the network right is no longer an IT decision; it is one of the most important conversations happening in the boardroom today to help fulfill AI ambition.”
APAC is also leading on adoption, with 35% of organizations reporting extensive AI use across the business, against a global average of 25%. Yet adoption alone is not enough without the right foundations beneath it.
Eric Wong, President, APAC, Expereo, says: “Asia Pacific is moving aggressively on AI adoption, but many organizations are discovering that scaling AI successfully requires more than just investment in applications and models. The underlying network, cloud connectivity, and operational readiness matter just as much. Across the region, we are seeing enterprises reassess whether their infrastructure is truly ready to support AI at scale, particularly around performance, resilience, governance, and visibility. Organizations that address those foundations early are generally seeing stronger outcomes and faster operational impact from their AI initiatives.”
Boardrooms are also waking up to the longer-term risks of unchecked AI investment. According to the survey, 54% of global tech leaders cite the creation of new security risks as a significant potential future threat for their organization’s use of AI, while 39% globally are concerned about losing track of AI-related costs and ROI once the technology is embedded across the business. In APAC, that concern is sharper still as 41% of technology leaders in the region are worried about losing oversight of AI-related costs and ROI as adoption deepens – a figure that rises to 54% in Malaysia. Digital sovereignty is also moving up the strategic agenda, with 38% of APAC organizations rating it a high or top priority as they look to retain control over data and navigate an increasingly complex regulatory landscape.
For the full IDC InfoBrief, commissioned by Expereo, please visit: [LINK]
ENDS
About Expereo
Expereo is a world-leading Managed Network as a Service provider that connects people, places, and things anywhere. Solutions include Global Internet, SD-WAN/SASE, and Enhanced Internet. With an extensive global reach, Expereo is the trusted partner of 60% of Fortune 500 companies. It powers enterprise and government sites in more than 190 countries, with the ability to connect to any location worldwide, working with over 2,300 partners to help customers improve productivity and empowering their networks and cloud services with the agility, flexibility, and value of the Internet, with optimal network performance.