Published on the 18/12/2025 | Written by Heather Wright
And where AI will create enterprise value in 2026…
What do vibe coding, AI medical scribes, generative engine optimisation and AI personal assistants and wellness companions have in common? Other than the obvious AI link, they’re all, according to private capital market intelligence company PitchBook, the most overheated hyped areas of AI – and riding for a fall.
The company’s 2026 Artificial Intelligence Outlook: The Great Competition Wars Have Begun, says AI is now a core infrastructure layer for the economy, set to drive value creation – and some painful value destruction – across enterprise software, cybersecurity, healthcare and defence.
“The paradigm is shifting from users navigating to a dashboard to insights being delivered conversationally and contextually within their workflows.”
But alongside the upside – and following a year in which AI guzzled 65 percent of total venture capital deal value – PitchBook’s analysts are also flagging a cluster of ‘overheated’, look-alike subsectors where capital may be burned with little defensible differentiation.
(And, to put things in perspective, AI’s 65 percent share dwarfs fintech’s 2022 peak of 17.1 percent, or mobility tech’s 2017 peak of 15.1 percent – despite enthusiasm around self-driving cars and scooters.
Running hot and overheating
The report highlights a number of areas running hot – potentially too hot – with startup sameness and incumbent bundling leading to a high risk of investments plummeting to zero. Leading the list are AI-based code generation startups (aka the much hyped ‘vibe coding’ sector) and autonomous driving software companies. While companies like Anysphere, which developed the Cursor vibe coding tool, has been one of the fastest growing startups around, reaching US$100m in annual recurring revenue in 12 months, its success, and those of some of the leading companies in the sector has attracted other companies to the sector.
That’s prompted Pitchbook senior emerging technology analyst Derek Hernandez, who covers software companies, to declare the market ‘oversaturated with startups that are too often simply thin rappers around foundation models lacking any deep or defensible moats’ (moats, you’ll soon discover if you read the report is a key phrase for Pitchbook). He warns that undifferentiated startups trying to mimic the top leaders are heading toward widespread commoditisation and, very likely, value destruction.
Second-tier AI scribes are also a danger area in what Pitchbook says is a marketplace too overcrowded, while $500m search as a service category, along with the emergent generative engine optimisation/answer engine optimisation visibility platforms risk becoming features in incumbent platforms.
Gaming content tools and parts of the precision medicine market (specifically genomics and biomarkets, where model outputs hinge on data access few startups can actually command, are also highlighted as overheated markets.
The report also warns that many incumbent industries are at risk, with that ‘at risk’ list reading like an enterprise software hall of fame. It includes legacy ERP, accounting and capital markets software (nothing that AI powered real-time analytics, smart forecasting and seamless workflow integration will be sticky), traditional analytics and BI platforms, with AI native competitors operating on natural language, and traditional healthcare administration, with AI ‘excellent’ at processing unstructured data.
AI-native challengers in the ERP space will offer real-time analytics, smart forecasting and workflows that feel embedded, rather than bolted on. “Companies still relying on manual workflows, static rule-based automation or limited data-driven insights will lose ground…”
Of traditional analytics and BI platforms, Pitchbook notes: “The core value proposition of legacy BI, especially manual dashboard creation and static reporting, is being rendered more and more obsolete by agentic AI and natural language querying. The paradigm is shifting from users navigating to a dashboard to insights being delivered conversationally and contextually within their workflows.”
Where AI will win big
On the flip side, it notes plenty of other areas primed for success. In enterprise tech, it flags AI-powered customer service and support as a ‘massive, well-defined market’ with a clear path to demonstrating ROI through automation. PitchBook is forecasting the market to grow from US$27.9 billion this year to $56.2 billion in 2025, as agents increasingly resolve customer interactions with autonomous ticket resolution and workflow automation cutting ticket volumes and handle times for ROI that CFOs can calculate.
Industry-specific specialists will differentiate and create ‘moats’ while others will build sticky systems of record.
Other top picks include foundation models – both incumbent and new model providers who specialise and so establish firmer pricing models – and AI-focused data management as the basis for all model interaction with the data layer. As enterprises wrestle with governance, lineage, retrieval and vector databases at scale, AI data management becomes foundational. PitchBook expects infrastructure SaaS to more than double by 2030, driven by the pivot to predictive, self-healing infrastructure across multicloud estates.
Agentic commerce infrastructure is among the areas primed for success, generating outsized venture returns. “Commerce is an on-ramp to every major platform shift, from the internet to mobile, to cloud and not to generative AI. Core infrastructure across payments, identity, fraud, loyalty and inventory systems will be rebuilt, enabling autonomous transactions in the future.
Rather than general purpose chatbots, value concentrates in domain-specialised models and agentic workflows that transact on behalf of users.
Supply chain planning is ‘the most underappreciated’ subsector within enterprise SaaS today, Hernandez says, but he’s forecasting that to change dramatically next year and into 2027.
“We see significant catalysts across the horizon for this subsector, including escalating geopolitical trade friction and tariff uncertainty, regulatory compliance and rising cost pressures on working capital.”
All of those catalysts are unpredictable and required heavy investment in supply chain resilience across ‘nearly’ every industry.
“We believe AI’s ability to perform predictive demand forecasting and network optimisation is becoming mission-critical,” Hernandez says.
