NEWSLETTER

Founder dependency is one of the largest valuation swings in private investing, and almost nobody names it as a discrete line item. It hides inside a discount rate instead of being decomposed, structured, and priced on its own terms.

A trained model and a labeled dataset are a snapshot. The pipeline that keeps that data flowing, licensed, and clean is the actual mechanism, and it is the question most AI diligence never gets around to asking.

Most funds claiming agentic AI exposure are underwriting the same thing, a customer facing agent wrapped around a foundation model. The real category forming beneath it, orchestration, execution, identity, and governance infrastructure, is a different business with a different moat entirely.

A falling gross margin gets read as a red flag in almost every diligence room. That heuristic is increasingly wrong. The question worth asking is not whether margin compressed, but which dollar of compression bought which capability, and whether that trade will repeat.

Market share earned through speed alone is a rented asset once a competitor closes the capability gap. This piece argues that defensible intellectual property, not market share, is the signal applied deep tech investors should diligence first.

In B2B services, indiscriminate growth can hide weak economics for a long time, but deliberate scale reveals real operating discipline. This piece examines why a company that knows exactly why it is the size it is often makes a stronger investment than one chasing a bigger revenue number.

The best exits are rarely the fastest. They are the ones where the business has reached real readiness. This piece lays out the four question framework we use to judge hold versus exit timing across a portfolio.

Family offices face elite level cyberattack exposure with thin internal defenses. That mix of high stakes, recurring need, and thin supply is what makes family office cybersecurity an emerging, underpriced market.

Compliance infrastructure looks like overhead until switching costs kick in. Rising HIPAA complexity and healthcare's highest breach costs of any industry are turning this quiet category into one of healthcare technology's stickiest return generators.

PropTech capital flows to resident apps and leasing tools while the managed IT, security, and infrastructure layer underneath goes largely unpriced. That gap is exactly why property management is an underwritten vertical.

Most investors discount a founder’s heavy spend on people as indiscipline. Read correctly, deliberate over-investment in team development is one of the cleanest, hardest-to-fake signals of operator quality and the asset most likely to compound enterprise value.

A 98% renewal rate can reveal more than customer satisfaction. For venture and private equity investors, strong retention often signals workflow embedding, economic clarity, trust-based stickiness, and real operating quality. In a market full of narrative, retention remains one of the most honest indicators of whether a company has built something durable.

Agentic AI is becoming a real B2B investment category. The winners will be companies that connect autonomous capability to enterprise workflows, governance, defensible IP, and repeatable operational value.

The AI companies worth backing are not the ones with the best models. They are the ones with the best delivery. That distinction is becoming clearer across enterprise markets. AI adoption is broad, but durable value creation is still narrow. McKinsey’s 2025 global survey found that 88% of organizations report using AI in at least one business function, yet only a minority have scaled it in ways that materially change enterprise economics. High performers separate themselves n

The market is rewarding AI exposure. It is not yet rewarding AI durability. That distinction matters. In the last two years, enterprise AI adoption has expanded quickly, but the economic picture remains uneven. McKinsey’s 2025 global survey found that 88% of organizations report regular AI use in at least one business function, yet only about one-third say they have begun to scale AI across the enterprise, and just 39% report EBIT impact at the enterprise level. In other word

Two AI startups. Same starting point. Opposite outcomes. The difference wasn't technology — it was sequencing. This piece breaks down why the most successful portfolio companies build operational foundations before swinging for breakthroughs, and how Azafran uses this pattern to identify investments positioned for premium exits.