The global automotive industry has reached top gear in deploying advanced connectivity systems and creating a huge market for software-defined vehicles (SDVs), and research has found that while data management and systems integration were the leading priorities for manufacturers a year ago, now advanced software capabilities and artificial intelligence (AI) integration are very much in the driving seat, emerging as the key enabling technologies and capabilities of SDVs.
The Sonatus survey, conducted by Wards Intelligence in March 2025, collected data from 576 respondents encompassing manufacturers, Tier 1 suppliers, Tier 2 suppliers, consultants and trade associations. Respondents were based in North America, Europe and Asia. The results are presented in The state of software-defined vehicles: Industry perspectives and market trends report.
Sonatus said the shift in perspective underscored the opinion that the automotive industry sees AI and software-related technology as fundamental to the next generation of vehicles, rather than merely supplementary features. The report noted that while artificial intelligence in vehicles often captures attention for its infotainment capabilities, particularly voice assistants, the most promising AI applications now centre on core vehicle functionalities.
Yet a readiness gap persists. While just over four in five respondents regarded SDVs as being critical to long-term success, only 23% felt “very prepared”. Even though results still demonstrate different SDV priorities across regions and points of the value chain, Sonatus said the variation was considerably lower, showing higher consensus. Looking at regional divergence, automotive manufacturers in Asia-Pacific and Europe showed stronger urgency and alignment than their peers in North America.
Over-the-air updates led the initial adoption of software capabilities by the automotive industry, with most respondents reporting this capability as already deployed. This was followed by the adoption of continuous improvement and continuous integration practices.
While still seeing AI as a key enabling technology for SDVs, respondents in Asia ranked software capabilities ahead of AI, unlike their counterparts in Europe and North America. Respondents anticipated AI deployment in three key phases: first in advanced driver-assistance systems (ADAS) and autonomous vehicle (AV) systems (2026-2027), followed by vehicle applications such as infotainment and comfort features (2028-2029). Organisational processes will incorporate AI from 2028 onwards, with the Asia-Pacific region showing the latest adoption timeline.
Most countries aligned with these timelines, but the survey showed two exceptions were emerging: Japan projects a significantly later implementation of AI in organisational processes, and Germany anticipates ADAS/AV applications deployment in 2028-2029, later than the global projection of 2026-2027. In addition, China demonstrated less urgency in automated software certification, potentially due to lower concerns about product longevity. Respondents are now much more confident that zonal architectures will be the dominant approach for electrical and electronic architecture by 2030, although implementation may take longer in North America.
The results also show a significantly higher appetite for open source solutions in safety-critical domains, especially Linux, compared with last year’s results. While the automotive industry strongly believes in the transformative potential of SDVs, respondents indicate that financial viability remains the primary challenge for manufacturers in scaling this technology.
Although the research suggested that the potential for post-sale monetisation opportunities is evident in luxury and, to some extent, mid-range vehicles, carmakers have not yet identified viable post-sale monetisation opportunities or return-on-investment justification in mass-market vehicles. This is likely because manufacturers continue to evaluate vehicle investments primarily through bill of materials calculations rather than a total cost of ownership evaluation model, overlooking the long-term financial potential of SDVs.
