
Tommy Lee Walker - stock.adobe.c
Datacentre outages decreasing in frequency, Uptime Institute Intelligence data shows
Datacentre outages are becoming less common and severe, but power supply issues remain enduring cause of most downtime incidents
Power supply problems remain the most common cause of “impactful” datacentre downtime incidents, but ransomware attacks are the leading cause of major public outages, research from Uptime Institute Intelligence reveals.
The datacentre market watcher’s Annual outage analysis report 2025 tracks the frequency and severity of datacentre outages suffered by respondents to its surveys.
Uptime Intelligence reported an overall decline in the frequency and severity of outages for the fourth consecutive year in a row, with on-site power distribution failures remaining the most common cause of serious outages.
Its data shows that 54% of respondents to its surveys cited power supply issues as the cause of their datacentre’s most recent, impactful outage, with uninterruptible power supply (UPS) issues the most frequent source of problems.
On this point, 42% of respondents cited UPS failure as the cause of the power-related datacentre outages they had experienced, 36% said they had suffered a transfer switch issues, while 28% said they suffered a generator failure. Cooling issues, after power-supply problems, were the most common cause of datacentre outages with 13% of respondents experiencing one.
“This is consistent with previous years and underscores the vulnerability of IT equipment to even brief power disruptions – including voltage sags, switching events and total power loss,” the Uptime report stated. “Cooling-related failures are less frequent than those caused by power issues, but they consistently account for roughly one in eight impactful outages annually.”
The market watcher said there is a risk that cooling-related incidents may become common in the years to come as increases in datacentre rack densities may result in higher failure rates.
The report suggested that the increasing demand for artificial intelligence (AI) may lead to more power and cooling-related datacentre failures in the years to come, while the geopolitical instability may also lead to more outages.
“Soaring demand for AI is straining existing infrastructure designs – especially around power and cooling – while electricity grid limitations and global trade tensions introduce new uncertainty in supply chains and expansion plans,” the report added. “Together, these pressures could eventually affect the stability of current reliability trends.”
Uptime’s outage analysis also includes data about major, high-profile downtime incidents, of which there were 119 that it tracked during 2024. This is the second-highest number of public incidents it has tracked since it began collecting public outage data in 2016, with 2019 being when the most public outages occurred, with 165 taking place that year.
According to the report, cyber attacks – particularly ransomware – were the leading publicly reported cause of outages in 2024: “The rise in cyber attacks reflects multiple factors: increased reliance on digital infrastructure across industries, expanding attack surfaces due to complex hybrid IT environments and remote management tools, and escalating risks from third-party software and supply chains.”
Andy Lawrence, founding member and executive director of Uptime Intelligence, said that while outages remain a part of life for datacentre operators, failure rates are improving and the frequency of incidents has slowed down.
“Datacentre operators are facing a growing number of external risks beyond their control, including power grid constraints, extreme weather, network provider failures and third-party software issues,” he said. “And despite a more volatile risk landscape, improvements are occurring.”
Read more about datacentre outages
- Even if outages in the datacentre are trending down, Uptime data this year traces rising costs per outage. So what can be done?
- A review has been published regarding the circumstances that led to two datacentres, hosting healthcare systems for Guy’s and St Thomas’ NHS Foundation Trust, overheating during the 2022 summer heatwave.