The international distribution business is certainly lively at the moment. Tariffs are applied and removed on a regular basis around the globe and there are some gloomy economic forecasts but Tim Carroll, DigiKey’s global head of digital business, is in a positive mood.

DigiKey global digital business leader

“Despite what are seen as economic red flags or warning signs, DigiKey is celebrating a record number of website visits and datasheet downloads in October,” he told Electronics Weekly.

It has recorded record highs of 30m website visits and 6m datasheet downloads a month. In addition to key large accounts, which make up 30%-50% of revenues, there are 950,000 direct customers who buy online.

The website-based part of the DigiKey’s business grew 32% [in October] Carroll said. A lot is high mix, low volume orders; customers buying 10-15 different components to try them out, to experiment and to use in prototypes.

“We can see a lot of design activity happening now, and . . . barring economic disruptions or tariff disruptions, we expect to see a really strong year next year,” said Carroll. Sales started peaking in April and May 2025, which will mean that over the next eight to 12 months, the production side of the business will also start to ramp up as those products go into mass market, he explained.

“There are four or five things driving sales right now. Definitely everything in education and maker space: Raspberry Pis, Arduinos, Adafruit, they’re starting to move into more out of the hobbyist space, into more production,” he said. The Raspberry Pi sales were up 100% last month, year on year, he added, confirming dynamic growth for these lower average price orders.

The company has make a large investment in entry-level technology to match its customer base. “Customers like start their like careers with DigiKey, whether they’re in a STEM programme in high school or robotics,” said Carroll. “And we try to build that habit so that people start at that kit hobbyist level, but then as they advance in their careers, they stay”.

Industrial automation has been DigiKey’s top growing segment for the last four years. It is not a strength historically but it is growing with partnerships with companies such as Siemens and Schneider.

Around AI and data centres there is a lot of activity around power, power management, heat management and heat sinks, he reported. “There are more wireless controllers in a data centre than I realised,” he conceded. “A lot of our wireless is going out to customers we know are building data centres around it.”

Carroll is also positive about how AI can be used to help the online business and its customers. It uses AI to analyse new product introductions (NPIs) and a software platform which uses around 80 signals to determine when to restock inventory and by how much. Carroll describes frequency of website visits, datasheet downloads and popularity scores from other sources (e.g. Google) to produce a velocity score. It also looks at lead times. Historically, he said, these can be notoriously inaccurate, so the data can be used to build up inventory as part of a lead time prediction model.

Around 80% of DigiKey product purchases are automated, which Carroll credits with being quick to market and for keeping stock replenished.

There is also a pricing tool which allows the distributor to identify if a lead time on a part changes. The result was, generally, that pricing went down, though, said Carroll. He estimated that 90% of price changes have been downwards “to make sure we’re not too slow to compete on the day-to-day.”

Following China announcing restrictions on rare earth metal imports, sales of Nexperia products went up 1,000%. Human intervention was to limit how much a customer could buy at any one time to prevent shocks to the supply chain as resellers sought to buy up stock. Some adjustments are manual, and sometimes AI is recalibrated. “The more dependent you are on AI, sometimes the more these disruptions, like tariffs or supply chain, break all your models. So you have to be really careful not to let the AI run without guardrails,” advised Carroll.