ICT usage in enterprises, 2023
Faster internet connectivity among enterprises using AI
Statistical news from Statistics Sweden 2023-11-30 8.00
During 2023, the speed of enterprise internet connections continued to trend upward. A majority of enterprises reported usage at speeds ranging from 100 to 500 Mbit/s, and the share is even higher among enterprises using AI technologies. This according to new data from the survey on ICT usage in enterprises.
Among employees in Swedish enterprises, 87 percent have access to the internet in their work. Most common was for enterprises to have a fixed connection, with just over 91 percent of enterprises. Among enterprises that do not have a fixed connection, other forms of internet connection are used instead, such as mobile broadband.
The use of faster internet connections has become increasingly more common over time. However, a time series break between the current and previous reference years means that the development over time should be interpreted cautiously. More information about the time series break can be found at the end of this news item. Since 2020, the most common internet connection speed is between 100 Mbit/s and 500 Mbit/s. During 2020, it was more common to have an internet connection speed of 100 Mbit/s or less than one of 500 Mbit/s or more. However, in 2023 the opposite is found. The fastest connections are now more common than the slowest ones among Swedish enterprises.
Enterprises using AI have higher fixed internet connection speeds
The use of AI in Swedish enterprises has increased from 9.9 percent in 2021 to 10.4 percent. Among enterprises using AI-based hardware or software, 98.5 percent also have a fixed internet connection. Using AI technologies is commonly connected with high computing capacity and internet speed, yet for enterprises using AI the most common internet speed is between 100 Mbit/s and 500 Mbit/s with 37.0 percent. A slightly smaller share of enterprises using AI had an internet speed of at least 1 Gbit/s, 33.4 percent. There are also few enterprises using AI that have internet speeds below 100 Mbit/s.
The access to and usability of AI technologies are continually increasing and since the last time questions regarding AI were included in the survey, in 2021, the development has been rapid. Among enterprises using AI, the most common was for enterprises to have purchased commercial software or systems ready to use, 44.5 percent. This was a decrease compared to 2021 with 12.1 percentage points. All ways in which enterprises acquire AI software or systems decreased or were unchanged compared with 2021 except the use of open-source software or systems which increased by 4.4 percentage points.
E-Commerce sales (via web and EDI) in 2022
Turnover from sales of goods and services via the web (the enterprises own website, apps or e-commerce sites) and EDI (Electric Data Interchange) amounted to SEK 2 342 billion, 25.6 percent of total turnover in 2022. Among enterprises conducting e-commerce in 2022, its share of total turnover amounted to 35.0 and 35.7 percent for small and medium sized enterprises respectively. For large enterprises turnover from e-commerce amounted to 39.0 percent of total turnover.
E-commerce sales differ between industries. The most e-commerce intensive industries in 2022 were Manufacture of food products, beverages and tobacco products (SNI 10-12), Repair of computers and communication equipment (SNI 95.1) and Accommodation (SNI 55). However, Wholesale trade, except of motor vehicles and motorcycles (SNI 46) and the ICT sector containing both manufacturing industries and service industries, accounted for the largest sales volumes over the internet.
Changes in the statistical unit
Ahead of the reference period 2023, changes have been made in the statistical unit enterprise. The number of enterprises with multiple legal units increases in the population, from about 30 enterprises to 50 000. This entails that there are fewer, but larger enterprises. This mitigates problems of double counting for some variables, such as turnover, as internal flows are consolidated. It also causes changes in the population structure concerning industry as well as having effects on size classes and regional reporting of the statistics.
More information can be found on Statistics Sweden’s website.
Changes in the statistical unit has also affected the design of the survey which in turn has effects for the statistics produced. More information on the design can be found in the methodological documentation for the survey. A report concerning the time series break can be read here:
Break in Series. Changes in ICT Usage in Enterprises and Innovation in the Enterprise Sector (pdf)
Comparisons over time should be made sparingly and with caution.
Definitions and explanations
The statistical news only reports on enterprises with 10 or more employees. More information on the statistics, as well as results for other study domains, can be obtained from the Statistical Database on Statistics Sweden’s website.
SNI – Swedish Industrial Classification Standard
The survey reports the following operations according to SNI 2007:
Manufacturing industry – SNI 10–33
Energy and waste-disposal – SNI 35–39
Construction industry – SNI 41-43
Trade; repair establishments for motor vehicles– SNI 45–47
Transportation and storage companies – SNI 49–53
Hotels and restaurants– SNI 55–56
Information and communication companies – SNI 58–63
Real estate companies– SNI 68
Other service companies – SNI 69–74, 77–82, 95.1
ICT-sector – SNI 26.1–26.4, 26.8, 46.5, 58.2, 61–62, 63.1, 95.1
For more information, see SNI.
Production and quality of the statistics
The questionnaire on which the data is based, as well as more information on the statistics, can be found in the documentation on the Production of the Statistics (2022). If you have any questions on the quality of the statistics, please refer to the Quality Declaration of the product (2022). These documents are provided here: ICT-usage in enterprises
Feel free to use the facts from this statistical news but remember to state Source: Statistics Sweden.