Good day everyone. September 1, 2025 - A sobering reality check has emerged for the generative artificial intelligence sector, with new research revealing that 95% of organisations are seeing negligible returns from their GenAI investments despite substantial financial commitments. This stark finding comes as cloud infrastructure spending is projected to reach nearly £400 billion in 2025, driven primarily by the enormous computational demands of large language model training and inference.
The profitability concerns have intensified following underwhelming performance from GPT-5, OpenAI's latest model, which failed to meet the ambitious expectations set by CEO Sam Altman. According to analysis by Direct Industry, both OpenAI and Anthropic are expected to report billions in losses for 2025, despite their combined user base exceeding one billion active users. This disconnect between adoption and monetisation has prompted serious questions about the sustainable business models underpinning the GenAI revolution.
The massive capital expenditure surge - from £119 billion in 2020 to a projected £400 billion in 2025 - represents the industry's largest infrastructure bet since the early internet era. However, unlike previous technology waves that demonstrated clear revenue pathways, GenAI applications have struggled to translate impressive capabilities into tangible business value, particularly for enterprise customers seeking measurable productivity gains.
Our view: The current profitability crisis in GenAI reflects the natural maturation of a transformative technology rather than fundamental failure. Whilst the immediate returns may disappoint investors, history suggests that truly revolutionary technologies often require extended development periods before reaching commercial viability. Companies should focus on specific, measurable use cases rather than pursuing GenAI for its own sake, building sustainable value propositions that justify the substantial computational costs involved.
beFirstComment