You have been forwarded this Newsletter? Subscribe here
Last week, SuperReturn International brought more than 6,000 decision-makers from over 80 countries, including 2,000+ LP representing more than $50 trillion in assets, to Berlin. We were there with our partners from TechMiners and the Modular Alpha network, the curated specialist network for value creation we launched last month, with a packed week of conversations (Learn more here). One pattern ran through almost every conversation: How the value of AI will show up in the numbers and to which extent.
What investors were discussing
AI sat at the center of the agenda, debated on main-stage panels by investors. But the tone has shifted. The backdrop this year: LPs pressing managers to return capital, exit values down by around a fifth this year, and a software selloff driven by the fear that AI-native entrants will undercut established business models.
In that environment, AI stops being a story you tell and becomes a number you have to defend. That was our core takeaway from the week: the AI conversation has moved from capability to accountability. Three observations behind it:
Everyone has initiatives but measurability of ROI is behind: Every fund and every company we spoke to is running AI projects. Very few can say what those projects are worth on the P&L, by how much, and over what time.
The fear is not missing the upside. It is mispricing the downside. The software selloff made the question concrete: what is a business worth when a competitor can build the same product at a structurally lower cost base? That question now belongs in every valuation and every budget, not just in tech due diligence.
The mid-market is where the gap is widest. SuperReturn dedicated a full summit day to mid-market strategies, and for good reason. Mid-sized and more traditional businesses face the same cost-curve pressure as software companies, but typically have the least analytical infrastructure to quantify what AI does to their economics.
Two theses that held up all week
Two ideas were confirmed in conversation after conversation on the ground.
Thesis 1: The 95% problem is a strategy gap, not a technology gap. AI investment is at record levels, yet a95% of AI pilots produced no measurable P&L impact so far. A separate survey finds only about 6% qualify as AI high performers, attributing more than 5% of EBIT to AI. Meanwhile, 42% of companies abandoned most of their AI projects in 2025, mostly for unclear business value rather than technical failure. The winners are not the ones with the longest use-case list. They are the ones who quantified, before investing, which levers move the P&L and by how much.
Thesis 2: Value created is not value captured. 74% of AI's economic value accrues to just 20% of organizations. The difference is leakage: to competitors who deploy the same tools, to customers through price competition, and to vendors through infrastructure and usage costs. An initiative built on proprietary data inside a redesigned workflow keeps most of what it creates. The same tool bolted onto an unchanged process keeps almost nothing. Defensibility decides the return.
—> We will release an in-depth paper on this very soon, stay tuned!
The takeaway
The most useful shift is one of sequence. The reflex is to pick a tool, run a pilot, and look for the return afterwards. The few who capture value reverse that order: they decide where value is large and defensible first, and only then spend. This holds whether or not you invest through a fund. Your competitors buy the same models from the same vendors, so access is not the advantage. Judgment about what to keep is.
Three questions to locate where you stand:
Can you name the AI initiative with the largest expected P&L impact in your business, and the number behind it? If not, you are managing activity, not value.
For your most important initiative: how much of the gain survives once competitors deploy the same tools? Savings that everyone can replicate tend to end up in lower prices, not in your margin.
Who owns the P&L for that initiative? If the answer is the technology team, the value will not arrive. Real impact needs an owner in the business unit that captures the benefit.
Weekly Picks
The AI Boom's Next Test: Is the ROI Worth the Cost: Parmy Olson on the moment of reckoning. With OpenAI and Anthropic heading for the public markets at trillion-dollar valuations, the next few months will test whether the returns justify the spend, the same question, one level up, that portfolio companies face on their own AI budgets.
Private markets gather in Berlin with not-so-super returns: The backdrop to the week. LPs pressing for cash back, exits down sharply, and a software selloff driven by the fear that AI-native entrants will undercut incumbent cost structures.
Agentic AI and the Future of Enterprise Software in 2026: A practical conversation for operators. The recurring advice maps onto our own: start with bounded workflows that have measurable outputs, track tokens and tool use against business outcomes, and fund agents by priority and ROI rather than letting every team spend on open-ended experimentation.
Stefan Benndorf & Dr. Philipp Engelhardt
Founding Partners scaleon
scaleon
Value Creation for Digital Leaders. Build for growth.
The questions this newsletter raises are the same ones our clients bring to us. Which AI initiatives deserve a second round of investment? Where does P&L accountability for digital transformation actually sit? How does technology deployment translate into measurable business value?
scaleon works with CEOs and investors of digital businesses on exactly these questions — in growth strategy, operational management, and transaction preparation.
If anything in this edition is worth a conversation, we'd welcome a direct message.