NEWSLETTER

A company raised its Series A at 80 million post-money on 1.5 million in ARR. Three years later, at 7 million ARR and 95 percent growth, Series B investors priced it at 130 million, well below the 200 million the founder still had in his head.

Andreessen Horowitz put roughly 17 percent of a fund into Skype. Sequoia put 15 percent into WhatsApp. The number that built those returns was never the entry price. It was the willingness to concentrate reserve capital once conviction was earned.

Roughly one in three signed letters of intent never reaches a closing, and quality of earnings discrepancies are one of the leading published reasons. Founder compensation is the line item that most often starts that unraveling.

Horizontal AI agent platforms attract the headlines and the demo day applause. In regulated enterprise B2B, vertical agents are the ones still generating value six months after deployment, and the gap is not close.

A first check is a bet on a person and a market. A follow-on check is a bet on a trajectory, and most funds still make that call on relationship and momentum rather than a repeatable pattern.

Most follow-on decisions get justified with the same metrics used to justify the first check. That is exactly the setup for escalation of commitment. The real question is whether I would back this founder for the first time today, based only on how they behaved over the last twelve months.

Industrial IoT companies exit at nearly double the rate of tech companies generally, on a cumulative fraction of what a single AI mega round raises in one quarter. That gap between fundamentals and capital is the definition of an underwritten category.

Investors negotiate hard over valuation and check size, then treat board composition as a formality near the end of the term sheet. That is backwards, because governance power is sticky and compounds across every round after this one, while capital is spent once.

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.

SLAs are often treated as contract language, but the real value is in whether the business can consistently deliver them without margin leakage. This article explains why investors should view SLA performance as a signal of customer trust, revenue quality, scalability, and 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.

Most companies do not have an AI strategy problem. They have an AI operations problem. Real value comes when AI is governed, supported, measurable, and built to run inside the business.

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