Growing investment in the cloud and artificial intelligence does not solve the fundamental problem facing many businesses: reliable access to services. A Censuswide survey , commissioned by DE-CIX and conducted in July 2026 amongst more than 400 IT decision-makers in the US and the UK, shows that technical teams spend an average of 11.5 hours a week resolving cloud connectivity issues. That is more than one working day a week and nearly 600 hours a year per team.
The most frequently cited challenges include latency between cloud environments, security threats, DDoS attacks and outages. At the same time, 96 per cent of respondents state that their infrastructure is ready for future cloud and AI projects. The discrepancy between this high level of confidence and the time wasted on day-to-day problems may indicate that companies are overestimating the maturity of their networks.
The issue is becoming more significant as the market grows. According to Synergy Research Group, global spending on cloud infrastructure reached $143 billion in the second quarter of 2026, an increase of $43 billion year-on-year. Services related to generative AI saw revenue increase by 165 per cent year-on-year.
Companies are responding by establishing private connections with cloud providers that bypass the public internet. Sixty-one per cent of the companies surveyed use them, but the proportion is significantly lower amongst smaller organisations. This may widen the advantage enjoyed by large organisations, which have the budget and specialists needed to build resilient networks.
The consequence will be an increase in expenditure on connectivity, monitoring and network expertise. Without these investments, the use of AI may increase not productivity, but the number of failures, delays and costly downtime. In Germany, 46 per cent of companies state that a prolonged cloud outage would force them to scale back or halt their operations.
