Mar 9, 2026, 10:54 AM ETInterview

"Cursor is Dead" is Total BS: Here is Why | Miles Clements

Excel emphasizes a "multi-stage, multi-strategy" approach, avoiding the extremes of only chasing massive late-stage momentum or sitting on the sidelines due to valuation concerns.
The firm's core philosophy, attributed to co-founder Arthur Patterson, is to "focus on hitting singles and doubles and let the home runs take care of themselves," avoiding the over-swinging that leads to failures in early-stage bets.
Myles Clements defines investing as an art and a science: the science lies in proper valuation, while the art lies in knowing when to break the rules.
Excel believes that missing a 100% win rate on deal sourcing and closing is acceptable, estimating a healthy win rate around 80% to remain competitive in difficult negotiations.
The firm conducts rigorous, "global offsite" partnership meetings to audit missed deals and failures, treating them as the most critical internal conversations for future course correction.
Excel explicitly rejects the notion of being a "coverage game" where every deal must be won, instead focusing on high-conviction bets where they can act as the lead investor of record.
The firm prioritizes "non-consensus" investments in smaller, bootstrap companies (e.g., Laravel) alongside consensus plays in breakout leaders to diversify portfolio risk.

Clements evaluates AI companies using a framework of "time to value" versus "durability of value," distinguishing between transient tools and enduring platforms.
High Time to Value / High Durability: Legal and accounting AI (e.g., Basis) require long deployment cycles but offer transformational value once adopted due to high switching costs for professionals.
Low Time to Value / Low Durability: Early "vibe coding" apps offer immediate utility (e.g., weekend projects) but lack long-term retention, leading to rapid market exit.
Coding AI (The Sweet Spot): The coding vertical (Cursor, Cloud Code) wins because it combines short time-to-value (productivity gains in hours) with high durability (compounding team usage and data flywheels).
Cloud Code's recent success is attributed to the underlying model improvements (Opus 4.5/4.6) and market expansion rather than stealing market share from Cursor, bringing new cohorts of non-developers into software creation.
The market is expanding consumption, with significant ARR growth driven by usage intensity (tokens consumed) rather than just new seat additions.

Cursor achieved $2 billion in Annual Recurring Revenue (ARR), having grown from approximately $100 million in ARR at the time of Excel's initial investment to a $27 billion valuation.
Despite public sentiment suggesting Cursor's decline (e.g., "Cursor is dead" memes, user complaints about pricing), internal metrics show 90% of users are daily active on the agent product, which grew 15x last year.
Cloud-based agents now account for 35% of merged PRs in Cursor, indicating a successful shift away from traditional IDE-based workflows.
Cursor employs a multi-model strategy, acknowledging that 50% of developers switch model families daily, allowing Cursor to act as an "index of AI innovation" without being locked into a single provider.
Cursor is building specialized coding models for enterprise tasks rather than general-purpose models, aiming to differentiate through professional utility rather than breadth.
Excel invested in Cursor at a pre-money valuation of roughly $9.5 million, a price point that seemed high at the time but was justified by the company's rapid scaling to $2 billion ARR.
The firm does not rely on rigid revenue budgets or quarterly earnings pressure, viewing financial outputs as a reflection of product-market fit rather than a primary investment constraint.

Rippling: Excel was not an investor in Rippling; reasons included a reputational perception of the founder Parker Conrad at the time, a desire to stick to specific ownership thresholds, and the high valuation required to enter the round.
11 Labs: Excel regrets missing 11 Labs, citing insufficient time spent with the founder and acknowledging it as a "very special" company essential to the modern AI stack.
Abridge: The firm regrets missing the opportunity to invest in Abridge due to a failure to secure a meeting with the founder, Shiv.
Service Titan: Excel missed Service Titan due to rigid valuation rules (capping at 6-8x forward revenue for vertical SaaS), which prevented them from capitalizing on a company that eventually became a $9 billion entity.
Anthropic & OpenAI: Excel participated in Anthropic rounds (including a $180 million round) but is not an investor in OpenAI.
Miro & Canva: While respecting these companies' financial profiles, Excel held their positions at high valuations ($17B for Miro) and did not sell secondary shares, aligning with the founder's strategic vision despite market pressure to take chips off the table.
CrowdStrike: Excel led rounds and held through the IPO, refusing to sell despite opportunities to diversify, a decision credited with capturing the company's eventual $100 billion market cap.

The definition of "Series A" has evolved; Excel notes multiple subcategories exist today, requiring distinct strategies for different growth stages and valuations.
To achieve fund returns, firms must participate in "mega outcomes" ($50B-$100B+), as traditional mid-market outcomes are insufficient for late-stage growth fund size requirements.
Clements believes the era of "triple, triple, double, double" growth is not dead but requires the right valuation metrics and market positioning to be viable.
The "SaaS apocalypse" is viewed as an over-rotation where mediocre growth rates (18-20%) were previously hyper-valued, rather than a fundamental collapse of the sector.
Public markets are currently viewed as a distinct, "casino-like" asset class that VC firms should generally avoid, preferring to stay within their expertise of early-stage technology investing.
Clements admits he was wrong a year ago in believing all generational AI bets were made; he now views the innovation flywheel as just beginning with significant upside remaining.
Excel is optimistic about the younger team's future impact, expressing pride in partners like Christine Esserman, Ben Quazo, and Josh over the next decade.

Remote vs. Rippling: Excel is invested in Remote but not Rippling, viewing them as separate business models despite similar HR tech adjacencies.
Deel: Excel is an investor in Deel, distinguishing its product vision and strategy from competitors like Remote and Rippling.
Atlassian: Clements maintains unwavering confidence in Atlassian founders Mike Cannon-Brooks and Scott Farquhar, stating "Never bet against Mike."
Competition: Excel respects and competes with firms like Meritech, Thrive, and Andreessen (specifically Matt Bornstein and Neil Marter) while acknowledging the breadth of the ecosystem.
Board Dynamics: The best board members combine humility with wisdom, avoiding the "rattling coin box" approach of being overly vocal or prescriptive without added value.
Founder Leadership: Clements prefers founder-led companies, noting that professional CEOs (e.g., Frank Slootman) are effective but that "unmistakably special" founders drive unique, long-term value.
Liquidity Strategy: Secondary sales and tenders are viewed as situational; the decision to sell depends on whether it benefits the company and founder first, citing CrowdStrike as a model for holding through public market volatility.