Panmure Liberum sees artificial intelligence fundamentally reshaping the software sector, accelerating the codification of services and processes while pressuring traditional SaaS business models.
In a wide-ranging note, the firm explores the implications of AI’s rapid development, identifies early commercial winners, and reiterates its top picks in enterprise software.
Panmure’s preferred name remains Alfa Financial Software (Buy, target price 306p), followed by Cerillion (TP 2000p) and Craneware (TP 2850p).
Deep moats
The analysts argue that deep domain expertise, complexity, and usage-based pricing models offer defensible moats amid AI disruption, noting that vertical software players with embedded workflows are better positioned than broader horizontal applications.
The note revisits themes from the firm's October 2023 report on generative AI and software automation, now updated to reflect a market increasingly focused on agentic AI, autonomous systems capable of planning and executing tasks.
The transition from reactive chat tools to multi-agent systems embedded in enterprise platforms is seen as a turning point for how AI integrates with business operations.
Complex workflows
Despite concerns that AI tools may reduce demand for per-seat SaaS licences, Panmure highlights that “not everyone is a businessman,” stressing that complex, regulated workflows still demand tailored solutions and human oversight.
The firm acknowledges that the BVP Emerging Cloud Index has declined 10.4% year-to-date as the market revalues software models under AI pressure.
The report outlines bottlenecks in compute, memory, and power as AI expands. It also flags the hyperscalers’ "prisoner’s dilemma", where massive AI infrastructure spending is necessary to remain competitive despite uncertain short-term returns.
Token costs for leading-edge models have fallen significantly, yet the ROI for many large language model (LLM) investments remains unclear.
Roadmap
The research also maps out the AI landscape into four functional tiers: foundation model labs (e.g., OpenAI, xAI), enterprise platforms (e.g., Palantir, Scale AI), creative and coding tools (e.g., Midjourney, Claude Code), and vertical applications like Harvey.AI for legal services.
The top 23 AI-native businesses tracked generated a combined $41 billion in revenue, with cumulative valuations above $1.6 trillion.
The UK’s role in the AI race is critically examined, with Panmure highlighting challenges in attracting and retaining talent and capital.
The report notes that while the UK has strong academic credentials, high taxes and limited sovereign investment in compute infrastructure risk pushing innovation abroad.
UK breakthroughs, US success
Most UK-founded AI success stories, such as DeepMind and Arm, now generate value primarily for US shareholders.
Panmure concludes that while AI presents significant disruption, it also reinforces the value of well-positioned enterprise software firms.
Companies combining embedded complexity, usage-based pricing, and domain-specific solutions remain attractive amid a rapidly evolving competitive and technological landscape.