The first week of November delivered a reality check for anyone who assumed the artificial intelligence boom would run uninterrupted. A sharp global equities selloff on Thursday, blamed by several strategists on investors conflating AI's transformational promise with near-term returns, dominated conversation from London to Dubai, while the prospect of blockbuster listings from Anthropic and OpenAI emerged as both a source of hope and a point of nervousness. Beneath the headline volatility, however, capital continued to flow into focused categories: construction robotics has quadrupled its previous funding record this year, and a new mental health venture from experienced founders closed a $6 million pre-seed round. For European operators, the week also brought a reminder that geopolitics is never far from the balance sheet, as the Iran conflict pushed airline fuel forecasts higher and kept oil prices on a knife edge. Funding & Investments The most notable early-stage round of the week came from Siggy, a mental health startup founded by the people behind Nurx and Journey Clinical. According to Crunchbase News, which reported the deal exclusively, Siggy has raised $6 million in pre-seed funding to build a service that uses AI to collect patient histories and track progress between appointments with licensed clinicians. The structure is instructive for European digital health founders: rather than positioning AI as a replacement for therapists, Siggy is using it to fill the gaps between sessions, where clinicians have historically had little visibility into how a patient is doing. That framing sits comfortably with the direction of European regulation, which has tended to favour AI tools that augment rather than displace licensed professionals. Investors have been visibly selective about consumer mental health businesses since the post-pandemic correction, so a $6 million pre-seed for a team with two prior companies behind it says something about where the appetite still lies: experienced operators, clear clinical accountability and a defined use for AI. The bigger funding story of the week, in dollar terms, is sectoral rather than company-specific. Crunchbase News published a snapshot showing that companies spanning construction and robotics have raised roughly $1.6 billion so far in 2026, more than four times the $389 million recorded across the whole of 2025 and comfortably tripling the previous annual peak. The surge was concentrated: three companies accounted for the majority of the capital. That concentration is worth dwelling on. It suggests investors are making a small number of large bets on platforms that can scale across sites rather than spreading capital across many point solutions. For European robotics founders, particularly in Germany, the Nordics and the UK, where construction labour shortages and ambitious housing targets are converging, the message is that the category has moved from curiosity to conviction, but that the bar for a winning round is now set by a handful of very large cheques. Set against those positive datapoints is a more sober macro picture from the investor community. Dan Niles, founder of Niles Investment Management, told Bloomberg Technology that competition from open-source models is weighing on token prices and slowing revenue growth at the leading AI companies. He argued that higher financing costs and postponed public listings mean valuations across the sector need to be scrutinised more carefully, and that the environment now rewards stock picking over broad exposure. The point about token prices deserves attention from European founders building on top of foundation models. Cheaper inference is good news for application-layer startups whose margins depend on model costs, but it is bad news for anyone whose pitch relies on a scarcity premium for AI capability. If Niles is right that open-source pressure is structural, the strategic advantage shifts further toward proprietary data, distribution and workflow integration, which happens to be where many European vertical AI companies have chosen to compete. A note of caution also came from the private credit world. Bob Sahota, managing director and chief investment officer at Revolution Asset Management, told Bloomberg Technology on the sidelines of the Sohn Hearts and Minds conference that his firm is deliberately not betting on data centres or attempting to pick AI winners, preferring cash-flow based lending with conservative underwriting. Sahota's view is one voice from Australia, but it echoes a growing strand of thinking among European institutional allocators: that the infrastructure layer of AI is now a crowded trade where the easy returns have been taken. Exits & M&A No major European exit crossed the wires this week, but the exit conversation was nonetheless the loudest it has been in months, driven by two names that are not European at all. Greg Brockman, president of OpenAI, told a Wall Street Journal technology conference on Wednesday that the company could be ready for an initial public offering in 2027, according to TechCrunch. Brockman used the appearance to emphasise OpenAI's progress in the enterprise market, where rival Anthropic had already made meaningful inroads. The timing of the comments was not accidental. With Anthropic's own listing reportedly imminent, OpenAI appears keen to signal that it will not cede the public-market narrative to its competitor. That Anthropic listing has become, in the eyes of at least one major bank, a pivot point for the entire market. Max Kettner, chief multi-asset strategist at HSBC, told Bloomberg Television that the bank is "in full on P&L protection" mode amid the global stocks rout, and that stabilisation could hinge on whether the Anthropic IPO goes through. "Let's see in the next two weeks if that goes through,"