Softcat just delivered a clean signal that AI demand is now flowing through the wider IT supply chain. The UK infrastructure provider raised its full-year profit guidance to mid-teens growth after a strong third quarter, helped by corporate demand for AI-ready hardware and some order pull-forwards caused by memory shortages.
Softcat's trading update on May 22 gave investors a practical read on where AI spending is moving next. The company reported strong double-digit year-on-year growth in both gross profit and underlying operating profit for the three months ended April 30, with particular strength in the corporate segment. Management lifted its full-year underlying operating profit outlook to mid-teens growth, up from its previous high-single-digit expectation, and Softcat shares rose more than 9 percent in London trading after the announcement.
That matters because Softcat is not Nvidia, Microsoft or Amazon. It is a UK IT infrastructure provider that sells into ordinary corporate technology budgets. When a channel business like this upgrades guidance on AI-enabled infrastructure demand, it suggests the spending cycle is moving beyond a small group of hyperscale buyers and into the companies that have to modernize servers, networks, storage, security and data systems before AI can become useful at scale.
“We have continued to see strong customer demand, with the impact of AI on technology driving investment and innovation across all elements of IT infrastructure,” CEO Graham Charlton said in the trading update. The company's portfolio covers data centers, networking, security, data, automation, hardware, software and services, which puts it close to the practical work of turning AI ambition into deployed systems. For founders building AI applications, this is an important shift. The bottleneck is no longer just model access or product design. It is the physical and operational infrastructure needed to run AI reliably.
What the Softcat signal means for AI infrastructure
Softcat's update is significant because it shows AI demand flowing through the reseller and services channel, not just through chipmakers and cloud platforms. The company serves more than 10,000 customers, ranging from mid-market businesses to large enterprises, giving it a useful view into mainstream corporate IT spending. When that customer base starts buying AI-enabled infrastructure in enough volume to move guidance, the trend becomes harder to dismiss as boardroom experimentation.
The company's broad offer also matters. AI projects do not run on GPUs alone. They need fast networking, better data architecture, stronger security, upgraded storage and service teams that can integrate the pieces without disrupting the existing business. Charlton pointed to Softcat's reach from the data center to the edge, through network, security, data and automation layers. That is where many enterprises are now spending, because AI pilots only become production tools when the underlying infrastructure can carry them.
The memory shortage complication
There is one important caveat. Softcat said growth was supported by “continued pull forward of some orders due to memory shortages.” That means part of the quarter's strength came from customers buying earlier than planned to secure supply. Memory constraints have become a recurring pressure point across the AI hardware market, as demand for servers and high-performance systems competes with limited component availability.
For buyers, the lesson is straightforward. Waiting can become expensive. If a startup expects to need dedicated hardware, upgraded workstations or on-premise AI infrastructure later this year, procurement timing now matters more than it did in a looser supply environment. Longer lead times and rising component prices can quietly reshape a product roadmap, especially for companies that assumed compute would remain a cloud-only question.
Investors also have to separate durable demand from timing effects. Jefferies analyst Charles Brennan noted that the debate is how much of Softcat's current momentum can be extrapolated into fiscal 2027, given the company's tendency to treat pulled-forward orders cautiously when setting future expectations. That is the right question. If customers are ordering early, some demand may have moved from future quarters into the current one.
The broader infrastructure picture
Softcat's update fits a wider pattern across infrastructure providers. Lenovo reported on May 22 that AI-related revenue grew 84 percent year on year in its fiscal fourth quarter and accounted for 38 percent of group revenue in the period. Cisco has also been leaning into AI networking demand, with its February 2026 launch of new Silicon One G300 systems aimed at large AI cluster buildouts and its March expansion of Secure AI Factory with Nvidia. The common thread is that AI spending is spreading into the less glamorous layers of the stack.
Softcat adds a different perspective because it sits closer to enterprise deployment than many hardware vendors. Direct suppliers can show what large buyers are ordering, but a reseller and services provider can show whether mainstream businesses are actually preparing their environments for AI use. That distinction matters. The AI investment thesis becomes stronger when corporate IT teams are upgrading networks, data systems and security controls, not just testing chatbots in isolated pilots.
What to watch next
Softcat's upgrade points to three clear implications. First, the AI hardware cycle is broadening beyond early adopters. Second, channel businesses are capturing value that will not always be obvious from chipmaker earnings alone. Third, supply constraints, especially around memory, may distort near-term demand and make future comparisons harder to read.
The next test is whether memory shortages persist into 2027 or begin to ease. If supply remains tight, customers may keep ordering early and channel demand could stay elevated. If shortages ease, Softcat may have to prove that the underlying appetite for AI infrastructure is strong enough to offset any demand pulled forward into this year. Either way, the third-quarter signal is useful. AI infrastructure spending is no longer just a hyperscaler story. It is showing up in the channel, and that is where many enterprise technology shifts become real.
