Only 20% of In-Car AI Features Profitable as Automakers Face High Costs, Low Returns

Currently, in-vehicle artificial intelligence (AI) technologies have been widely adopted across intelligent cockpits, advanced autonomous driving, and connected vehicle services, becoming a critical tool for automakers to enhance product premium pricing and market competitiveness. The global in-vehicle AI market is projected to grow from $12.8 billion in 2025 to over $50 billion by 2034, with the industry broadly viewing it as a core growth arena over the next decade. However, actual profitability significantly lags behind market expectations. According to a survey of 75 global industry experts, only 20% of current in-vehicle AI features are generating positive returns, while the remaining 80% are either operating at a loss or merely breaking even. Although mainstream automakers have already deployed numerous AI functions at scale—BMW, for instance, offers 22 standalone in-vehicle AI applications covering voice assistants, advanced driver-assistance systems (ADAS), and personalized connected services—monetization remains elusive. The root cause lies in the structural mismatch between the ongoing operational costs of in-vehicle AI and the traditional automotive business model. Conventional car manufacturing relies on one-time hardware sales with relatively low marginal costs, whereas in-vehicle AI demands continuous investment in cloud computing power, data operations, and system updates—each user interaction incurring additional expenses. Moreover, most automakers lack granular cost-benefit analysis for individual AI features, leading to an overabundance of “zombie features” (with daily active usage below 5%) and “useless features” (offering poor user experience and low practicality) that consume resources without delivering commercial returns. To overcome this challenge, industry experts recommend that automakers shift away from a quantity-driven approach and instead prioritize both user experience and commercial value. This involves streamlining existing feature portfolios by eliminating ineffective or detrimental functions, while adopting a hybrid architecture combining on-device processing with cloud-based intelligence to ensure rapid response times and reduce operational costs. Additionally, in-vehicle AI should evolve from a passive service assistant into an active transactional intelligent agent—capable of context-aware recommendations and autonomous payments—to unlock value-added services across the entire mobility ecosystem and explore new monetization models beyond subscriptions. Ultimately, the success of in-vehicle AI will no longer be measured by the number of features offered, but by its ability to establish a sustainable balance among cost control, user experience, and diversified revenue streams.