turbomotion046 - stock.adobe.com
UK falls behind on supercomputing amid slow investment, NAO warns
While UKRI has improved its oversight of research and innovation, funding remains fragmented and has been too slow to replace supercomputers
UK Research and Innovation (UKRI) spent £1.17bn on on research infrastructure in 2024-25, but fragmented funding and “too optimistic” business cases have led to key projects suffering delays, according to a report by the National Audit Office (NAO).
The report on the UKRI highlights that while there has been significant improvement in oversight with a more consistent approach to funding, several areas still require improvement.
The UKRI is sponsored by the Department for Science, Innovation and Technology (DSIT), and the department has began to take a more active role in its work and identified three broad priorities for the research and development (R&D) system, according to the NAO.
Commenting on the report, NAO head Gareth Davies said DSIT and UKRI have been “making headway on the oversight and management of research infrastructure to support strategic government priorities. But a number of key risks remain to government’s ambition for the UK to be one of the top three places in the world to create, invest in and scale-up a fast-growing technology business.”
One of the key issues, according to the auditor, is that UKRI is not managing its research infrastructure as an integrated portfolio. “Managing projects and other activities as a portfolio allows organisations to maximise the likelihood that they achieve their intended outcomes regardless of the performance of individual components of the portfolio,” the report said.
It added that while the organisation is trying to apply portfolio management techniques to the 30 projects part of the Infrastructure Fund, it is struggling to do so due to the “inflexibility of the funding arrangements”, which meant there limitations on moving money between projects or prioritising the most strategically important investments.
Between November 2023 and July 2025, projects were not allowed to bring spending forward, despite knowing other projects were developing slower than planned, which resulted in an £18m underspend.
The 30 projects have a combined expected cost of £2.04bn, and £451m had been spent by March 2025.
The report also highlighted concerns regarding the UK’s supercomputing abilities, warning that the country has fallen behind international competitors due to delays in replacing the ageing Archer2 system and procuring a new supercomputer. It added that there was no overarching replacement strategy until recently.
“The UK’s supercomputers have lagged behind those available to researchers in other countries in recent years. The cost of building and operating supercomputers has increased significantly in recent years – the Next National Supercomputer is expected to cost up to £750m to build and operate for five years,” the report said.
“Archer2’s declining performance relative to other supercomputers means it is less capable of supporting world-leading research. Additionally, it is currently expected to close 13 months before its replacement can be switched on, although UKRI believes it will be able to limit the consequences of this for researchers.”
Investment challenges
The report also underlined issues with UKRI’s business cases, particularly in the early stages. It said that in several cases, the business cases had underestimated the complexity and cost of delivering new research infrastructure.
One project, aiming to upgrade a research aircraft, spend £46m of its £49m budget before it was decommissioned as it would not deliver all planned benefits.
“Our review of business cases found that UKRI did not do enough initially to ensure that the projects’ delivery risk assessments were realistic, and it fixed budgets too early before costs could have been well-understood,” the report said.
“UKRI did not initially put in place appropriate safeguards to ensure it challenged overoptimism from individual research councils – who were naturally keen to ensure that their priorities were funded.”
Despite these issues, the report found that overall, research infrastructure across the UK is already delivering significant benefits, including medical research, biotechnology and climate science.
However, ageing facilities and underinvestment in maintenance continue to pose a growing risk to research and innovation in the UK. Universities would require around £5.6bn to fully restore their own infrastructure, while the Science and Technology Facilities Council estimated that £360m would be needed to restore its research estate to an acceptable condition.
A previous NAO report on UKRI’s research and innovation funding, published in May 2025, also sighted a lack of coordination in how the government expects UKRI to support the delivery of a range of objectives.
Due to the broad nature of UKRI’s activity, government departments indicate their policy priorities to UKRI through a variety of means, including ad hoc and routine meetings, government strategies and mission statements, and spending review budgets. However, the NAO reported that these are not consolidated or ranked.
The previous report called on UKRI to establish a strong approach to understand how its work is providing a return on investment for taxpayers.
Read more about UKRI and innovation
- Government funding to help businesses discover AI value: The government is betting the bank on the power of artificial intelligence to fix the public sector, mend roads and boost the UK economy.
- The UKRI strategic framework for 2031 lays out the steps the UK needs to take to drive forward innovation and academic research in artificial intelligence.
- OpenUK and UKRI collaboration aims to boost awareness of open source initiatives, community development and the procurement process.
