Overview of 20VC: Inside Accel's $4BN Growth Investing Machine (Harry Stebbings with Miles Clements)
This episode is a wide-ranging, candid conversation between Harry Stebbings and Miles Clements (Accel’s growth-investing lead). They cover frameworks for valuing AI companies, the “Cursor is dead” debate, why some AI/coding tools win, how to think about ownership vs. outcome, the state of late-stage SaaS and IPO windows, firm strategy and coverage, board/investor behavior, and practical advice for founders and early-stage investors. Miles is unusually open about misses, wins (e.g., Linear), regrets (e.g., 11Labs, Rippling), and how Accel is adapting to a hyper-competitive AI market.
Key themes and frameworks
- Time-to-value vs. durability of value
- Evaluate AI companies by (1) how fast customers see value and (2) how durable that value is. Coding tools (e.g., Cursor, CloudCode) often score high on both, while some domain-specific AIs (legal, accounting) have slower deployment but durable impact once adopted.
- Growth can hide underlying problems
- Heavy top-line growth should be interrogated with usage/engagement and marginal ease of ARR accumulation (how repeatable and scalable are revenue levers).
- Investing is art + science
- Science = valuation and inputs; Art = knowing when to break rules. Stick to playbook mostly, but selectively bend rules when conviction is unusually high.
Cursor, CloudCode, and the “Cursor is dead” debate
- Context: public debate on which coding AI IDE wins (Cursor vs CloudCode). Cursor reported dramatic ARR growth (claimed $2B ARR in discussion).
- Miles’ defense of Cursor:
- Adoption shift to agents: more users run agents than use old IDE tabs; agents are major driver of activity.
- Multi-model architecture is a strength: developers switch models frequently and want multi-model access—Cursor is positioned as an index/flywheel on model innovation.
- Specialized coding models make sense for professional engineering workflows (not a generalist LLM need).
- Metrics cited (from Cursor posts, per Miles):
- Agent usage > tab usage; 90% of Cursor users are daily active users of agent product in his example.
- Agent product grew ~15x in a year; cloud agents responsible for ~35% of merged PRs soon after launch.
- Cost-model concerns (dependency on Anthropic, etc.): Miles thinks being multi-model and building specialized models mitigates cost/commercial risk.
Valuation, ownership, and portfolio strategy
- Ownership vs. outcome tradeoff:
- Market now rewards mega outcomes; funds with large size must target very large outcomes. But being multi-stage (early + growth + late) allows laddering ownership up over time.
- Miles: there’s room for different strategies—invest in consensus winners, non-consensus, or seek high ownership in small(er) startups. Nuance matters.
- When to break rules:
- Break rules rarely and deliberately. Use early-stage subcategories to decide which kinds of Series A opportunities to participate in.
- Marginal ease of ARR accumulation:
- Focus on whether future growth can be easily driven by repeatable levers (e.g., product-embedded revenue lines, expansion motions) rather than one-off spikes.
Public vs private, IPO window and the $2B–$10B debate
- Many founders are staying private longer because private markets can now supply liquidity, M&A currency, and valuation anchoring.
- Companies in the $2B–$10B range face a tough public market outlook; below ~$5B can be murky for public breakout.
- Miles: expect more M&A / LBO homes for matured SaaS companies (Thoma Bravo, Vista, TA Associates, etc.)—LBO buyers will be attractive buyers in current market.
- SaaS valuation reset:
- The “SaaS-pocalypse” is partly a correction: valuations priced in aggressive growth and terminal multiples; some companies are oversold, some truly impaired.
Firm strategy, sourcing, coverage and internal culture
- Coverage & win rates:
- Accel (as described) aims to cover top opportunities; partners hold themselves accountable in offsites. A healthy firm win rate (term sheets → wins) might be ~80%, per Miles.
- Multi-stage capability:
- Being multi-stage is essential for laddering to target ownership and backing winners across rounds.
- Internal hygiene:
- Regular global offsites to evaluate missed opportunities and improve coverage; treat these as the firm’s most important conversation.
- People: names called out internally as strong sources/pickers (e.g., Christine Esserman, Andrew Braccia, Samir Gandhi).
Founders, boards and investor behavior — practical advice
For founders:
- Seek investors who add value on the big, infrequent decisions (M&A, pivots, hiring execs), not micro-managing all small product choices.
- Look for board members who are humble, thoughtful, and add perspective rather than loud opinions.
- Consider timing of liquidity carefully—do what’s best for the company and founders, not just investor timing.
For investors:
- Embrace nuance rather than extremes; don’t only chase momentum or sit out completely.
- Focus on indicators beyond raw revenue growth: engagement intensity, retention, marginal revenue mechanics.
- Know when to be aggressive on ownership and when to accept lower ownership for access to rare opportunities.
Notable wins and misses (Miles’ callouts)
- Wins:
- Linear: a deeply personal win for Miles; he pursued the founder persistently and felt great about the result.
- Some earlier bets (examples like CrowdStrike referenced as an example of not selling chips off the table).
- Misses/regrets:
- 11Labs: a regret—didn’t spend enough time with the founder and missed the opportunity.
- Rippling: Miles respects Parker Conrad as a generational founder and regrets not participating; firm stuck to rules and missed a big outcome.
- Foundation models: Accel missed some early model investments (Anthropic/OpenAI), acknowledged and corrected course.
Quick practical takeaways
- Evaluate AI startups by time-to-value + durability; prioritize products that are both fast to adopt and sticky.
- Multi-model, agent-focused developer products are powerful—developers switch models often and value tools that integrate many models and workflows.
- Growth is necessary but insufficient—dig into how repeatable and durable revenue is (marginal ease of ARR accumulation).
- Don’t reflexively avoid high valuation momentum rounds—occasionally break rules when conviction and founder quality justify it.
- Public markets are less hospitable for the $2B–$5B bands; expect more late-stage private exits and LBOs.
Notable quotes from the episode
- “Investing is an art and a science. The science is understanding how to properly value a company and the art is understanding when to break the rules.”
- “Focus on hitting singles and doubles and let the home runs take care of themselves.” (Arthur Patterson’s framework — used to advocate consistency and discipline.)
- “Growth can obscure and blind you to a lot of underlying ills in the business.”
- “Optimists make money, pessimists are right.”
Final thoughts
This episode is candid and tactical: it’s valuable for founders trying to understand what top growth-stage investors look for, and for investors refining frameworks in an AI-dominated, high-competition market. Miles emphasizes nuance, process, and the enduring importance of founder quality and repeatable growth mechanics—even as outcome sizes and market dynamics shift dramatically.