As enterprises embrace AI in the cloud, operational costs and security risks escalate

While public cloud remains the preferred platform for enterprise AI due to its convenience, organisations face mounting challenges in managing costs, security, and operational complexity as AI usage deepens, prompting a shift towards hybrid and on-premise models.

Public cloud has become the default launchpad for enterprise AI because it removes most of the friction that once slowed digital projects. Companies can tap compute, storage, managed services, pretrained models and global infrastructure almost immediately, which means teams can move from idea to pilot without first building a major platform operation. In practice, that convenience has made hyperscale clouds the easiest route for organisations under pressure to show AI progress quickly.

That convenience, however, is colliding with a harder reality: getting AI into production is no longer the main challenge, running it safely and at scale is. Lenovo research cited by TechRadar says many employees are already using AI tools weekly, often outside IT oversight, creating what it described as an execution gap. At the same time, NTT Data found that cloud maturity is holding back AI programmes, with most firms warning that current spending patterns could undermine modernisation work rather than support it.

Costs are becoming a bigger concern as AI usage deepens. TechRadar reported that some businesses are shifting towards hybrid and on-premise models as large language models and AI agents generate persistent API traffic, GPU demand and cloud data-transfer charges. What once looked economical for testing can become far more expensive once workloads move into constant use, especially when organisations are paying for every query, every transfer and every burst of processing.

Security and governance are adding to the pressure. Palo Alto Networks, in research also highlighted by TechRadar, found that AI-powered cloud services are being deployed faster than many firms can control them, with excessive permissions, misconfigured storage and database settings, and growing numbers of non-human identities all increasing risk. Even so, as TechTarget has noted, public cloud remains the main vehicle for many enterprises because it still offers the fastest path to proof of concept, automation and productivity gains. That leaves executives with a familiar dilemma: the cloud may be the easiest place to start AI, but it is increasingly where the operational, financial and security costs become hardest to ignore.

Source Reference Map

Inspired by headline at: ^[1]

Sources by paragraph:

Source: Noah Wire Services

Noah Fact Check Pro

The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.

Freshness check

Score:
10

Notes:
The article was published on May 1, 2026, making it highly current. No evidence of recycled or outdated content was found. The narrative appears original, with no significant discrepancies in figures, dates, or quotes compared to earlier versions.

Quotes check

Score:
10

Notes:
The article does not contain direct quotes. All information is paraphrased or generalised, which is appropriate for an opinion piece. No concerns regarding the use of quotes.

Source reliability

Score:
8

Notes:
The article is published by InfoWorld, a reputable technology news outlet. However, it is an opinion piece authored by David Linthicum, which may introduce subjective bias. The content is not a direct news report but offers analysis and commentary.

Plausibility check

Score:
9

Notes:
The claims made in the article align with known industry trends regarding the costs and challenges of running AI in the cloud. The information is plausible and consistent with other reputable sources. No significant concerns were identified.

Overall assessment

Verdict (FAIL, OPEN, PASS): FAIL

Confidence (LOW, MEDIUM, HIGH): MEDIUM

Summary:
The article is a recent opinion piece from a reputable source, offering analysis on the costs and challenges of running AI in the cloud. However, as an opinion piece, it represents original creative work rather than factual reporting, which carries inherent originality that cannot be fully replicated. Additionally, the reliance on secondary sources without direct links to primary data may limit the ability to independently verify the information. Given these factors, the content does not meet the necessary standards for factual reporting and verification, leading to a FAIL verdict with MEDIUM confidence.