Вход на сайт

Просмотр новости

Найдите то, что Вас интересует

Troubleshooting cloud connectivity emerges as main barrier to AI adoption

Дата публикации: 31-07-2026 11:34:00

Research finds that despite 96% of respondents claiming their enterprise networks are ready for future cloud and AI, the average IT team still spends more than 11 hours each week resolving cloud connectivity problems

Основное содержимое страницы с новостью.

Seventyfour - stock.adobe.com

Research finds that despite 96% of respondents claiming their enterprise networks are ready for future cloud and AI, the average IT team still spends more than 11 hours each week resolving cloud connectivity problems

Joe O’Halloran

By

Published: 31 Jul 2026 16:34

There are few businesses ignoring artificial intelligence (AI) these days and most deploying it do so in confidence that their networks are ready to support the high-performance connectivity needs of AI. However, research from DE-CIX has found that enterprises are time in troubleshooting cloud connectivity due to increased AI workloads.

The study from the internet, cloud and AI exchange operator was conducted by Censuswide and took the opinion of more than 400 IT and infrastructure decision-makers across the US and UK.

Among the key findings was that companies worldwide will invest almost $7tn in datacentre infrastructure by 2030, with more than 40% of that spending taking place in the US. In the UK, $60bn has been invested in private-sector AI infrastructure since July 2024. Yet the study noted that as governments and businesses continue to spend billions expanding the infrastructure underpinning AI, many enterprises are continuing to overlook one critical component: the network connecting users, applications, data and cloud environments.

It found that while 96% are confident that their networks are ready to support cloud and AI initiatives over the next two years, an average of 11.5 hours per week – more than a full working day – is still spent resolving network connectivity issues.

Other leading concerns included downtime or reliability issues (26%), latency or slow performance (28%), and security vulnerabilities/DDoS attacks (27%). Cost of connectivity, staff expertise and lack of visibility/control over data flows were also cited as major challenges, which DE-CIX said underlines the complexity involved in connecting and managing modern multicloud environments.

As a result, as indicated in the study, many businesses are now turning to private interconnection, which enables enterprises to connect directly to cloud providers over dedicated infrastructure rather than routing traffic across the public Internet.

Designed to deliver lower latency, greater resilience, enhanced security and more predictable performance, private interconnection has become an increasingly important way of supporting modern cloud and AI workloads. Specifically, the data showed that 61% of companies are already using private connectivity to clouds, while another 31% are actively considering it.

In addition, while 71% of enterprises with 1000 or more employees are using private connectivity to clouds, only 31% of medium-sized enterprises with 100 – 249 employees use such solutions. In contrast, only 8.62% of the smaller companies were spending 21 to 40 hours per week dealing with connectivity issues, while just 2.53% of the largest companies in the sample do.

Summing up these findings, DE-CIS said that together they suggest direct interconnection is rapidly becoming a core component of enterprise cloud and AI infrastructure and a competitive advantage for companies, though optimising interconnection strategies clearly remains a pressing challenge for small and medium-sized enterprises.

“AI has understandably brought renewed attention to GPUs, datacentres and compute capacity, but none of those investments deliver their full value without the network that connects them,” said Ivo Ivanov, CEO of DE-CIX. “Every AI application depends on data moving quickly, securely and predictably between users, clouds and AI infrastructure.

“Our research suggests that far too many enterprises are still spending valuable time trying to maintain that kind of connectivity, with more than a third spending  between 11 and 20 hours per week, and just under one in 10 spending between 21 to 40 hours per week. This confirms what we already knew – that that network architecture can make or break AI adoption. For enterprises, the opportunity is not only to improve application performance, but to reduce the operational burden created by complex cloud connectivity.”

Read more on Artificial intelligence, automation and robotics

Схожие новости

#Наименование новостиТональностьИнформативностьДата публикации
1Infrastructure Limitations Push 95% of Enterprises to Delay AI Projects08.6617-08-2026
2Majority of Organizations Agree That Many GRC AI Tools Aren't Ready09.317-07-2026
3Connectivity becomes the weakest link in Britain’s workplaces010.8923-07-2026
4Aligning European Spectrum Policy with AI-Driven Industrial Competitiveness014.5215-06-2026
5Why the next AI race will be won at the inference layer010.3811-08-2026
6AI is saving office workers hours — and stealing much of that time back in 'botsitting'08.4312-06-2026
7Data centre delays expose AI cloud power limits0508-07-2026
8Больше половины топ-менеджеров торопят внедрение ИИ, игнорируя неготовность инфраструктуры018.3327-07-2026
9GPUs And RAM Are In Short Supply, But The Real Bottleneck For AI Is Electricians01027-05-2026

Классификация: Пресс-релизы. Схожих патентов: 0. Схожих новостей: 9. Тональность: 0. Информативность: 12.4. Источник: www.computerweekly.com.