Enterprise AI capex reaching inflection; software vendors consolidate

Enterprise software and infrastructure vendors are consolidating around AI. SAP announced a strategic investment in n8n, valuing the workflow-automation platform at $5.2 billion, double its previous valuation. SAP plans to embed n8n directly into Joule Studio, its AI assistant platform, signaling that enterprise automation is becoming table-stakes in the AI era. This marks n8n's arrival as enterprise AI infrastructure, a milestone for European technology ecosystem.

Parallel to this, private AI adoption is accelerating. AccountTECH unveiled a major private-AI strategy emphasizing on-premise language models, hybrid development architecture, and protection of client IP. This reflects growing enterprise demand for sovereign AI that avoids reliance on OpenAI or other public LLMs. CME Group and Silicon Data partnered to launch the first Compute Futures, derivative contracts on GPU compute capacity, reflecting financialization of AI infrastructure as a commodity asset class.

Hyperscalers continue to articulate capex intentions: Mitsubishi Heavy notes that global gas-turbine orders will remain strong due to data center buildout. Amazon's negative Q1 free cash flowCash generated after maintenance capex; the actual money the business throws off. of minus $18 billion (following $44 billion in capex) underscores the near-term cash burn, but equities have shrugged this off on the assumption that long-term revenue upside justifies the capex surge. Palantir's US revenue is doubling year-over-year, and Trump's public endorsement of the firm's war-fighting AI capabilities adds political cover to enterprise deals.

The risk is execution: if enterprises deploy private-AI tools and discover poor ROI or data-quality issues, capex cycles could decelerate. Additionally, if OpenAI's revenue-sharing agreement with Microsoft falls apart (as reports of OpenAI discontinuing $38 billion payments suggest), the economics of big-tech AI investments become murky. For now, conviction remains high that AI capex is sticky and multi-year.