Skip to content
a16za16z

Is Non-Consensus Investing Overrated?

Is non-consensus investing overrated—or the secret to venture returns? a16z General Partner Erik Torenberg is joined by Martín Casado (General Partner, a16z) and Leo Polovets (General Partner, Humba Ventures) to unpack the debate that lit up venture Twitter/X: should founders and VCs chase consensus, or run from it? They explore what “consensus” really means in practice, how market efficiency shapes venture outcomes, why most companies fail from indigestion, not starvation, and the risks founders face when they’re too far outside consensus. Timecodes: 0:00 Introduction 0:27 Consensus vs. Non-Consensus: Investor Perspectives 3:14 Market Efficiency and Company Valuations 6:06 Anecdotes and Data: Hot Rounds and Outcomes 11:10 The Role of Founders and Fundraising Dynamics 15:21 Risks and Rewards: Indigestion vs. Starvation 17:50 Market Cycles, Efficiency, and the AI Craze 19:54 Personal Startup Stories & Lessons Learned 32:46 Fund Size, Ownership, and Mega Outcomes 38:50 What's the New Norm? 43:39 Venture Identity vs. Market Reality 45:00 The Future of Venture: Efficiency, Competition, and Growth 54:00 Seed vs. Multi-Stage Funds: Who Wins? 55:24 Closing Thoughts Resources: Find Leo on X: https://x.com/lpolovets Find Martin on X: https://x.com/martin_casado Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on Twitter: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details, please see a16z.com/disclosures.

Martín CasadoguestLeo PolovetsguestErik Torenberghost
Sep 4, 202555mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 2:28

    Why “non-consensus investing” can be dangerous (and what the tweet really meant)

    Martín explains the viral tweet: being proudly non-consensus can be a red flag because you might be missing something, and because startups depend on future capital. He clarifies he’s not advocating “consensus investing,” but arguing that ignoring market consensus is risky—especially given his belief that early markets are more efficient than people assume.

    • Non-consensus can mean you’re wrong or missing key information
    • Startups often need follow-on funding; markets can veto survival
    • Martín’s academic peer-review analogy: great work still needs acceptance
    • Core clarification: awareness of consensus matters; chasing consensus doesn’t
    • Claim: early-stage markets are “pretty darn efficient”
  2. 2:28 – 3:36

    Seed vs. Series A viewpoints: non-consensus early, consensus later

    Leo largely agrees that companies eventually need consensus, but argues many of his best seed investments began as non-consensus because proof points were scarce. The upside of getting in early is that once a company becomes obviously good, valuations can move so fast that later investors capture smaller multiples.

    • You eventually need broader investor buy-in to keep companies funded
    • Many top seed deals look non-consensus before clear proof exists
    • Valuations can re-rate extremely fast once traction appears
    • Non-consensus isn’t about genius insight; often it’s about timing/proof
    • Stage matters: pre-seed/seed has more uncertainty and more upside
  3. 3:36 – 6:04

    Anecdotes vs. real consensus: “hard round” isn’t the same as “market disagreement”

    The group challenges common narratives that big winners were non-consensus simply because fundraising was difficult at some point. Martín argues many cited examples had elite founders, known spaces, YC signaling, and consistently high pricing—suggesting they were not truly non-consensus in the market-wide sense.

    • Defining “consensus” is slippery; people talk past each other
    • A tough raise doesn’t necessarily imply the market disagreed on value
    • Examples like Anduril/Scale: elite founder + known space ≠ non-consensus
    • High median pricing across rounds can contradict ‘non-consensus’ claims
    • Focus should be on company quality, not getting a ‘deal’ vs other investors
  4. 6:04 – 11:06

    Hot rounds, follow-on momentum, and how to measure market efficiency

    They explore whether hot rounds predict future success, and whether investor demand is itself a useful signal. Martín and Leo discuss doing correlation and basket analyses (e.g., number of term sheets, speed of follow-ons) to test whether “hotness” reflects underlying quality or just herd behavior.

    • Peter Thiel quote: fast/higher up rounds may be a positive signal
    • Hypothesis: the best predictor of a hot round is that the prior round was hot
    • If round momentum persists, that suggests inductive market efficiency
    • Basket analysis beats cherry-picked anecdotes (compare cohorts over time)
    • Investor perception can create value even when business execution lags
  5. 11:06 – 13:08

    Founder fundraising dynamics: the hidden cost of being labeled non-consensus

    Erik and Martín emphasize that founders feel real pressure: they must appear fundable to VCs to ensure runway and follow-ons. Martín notes founders’ DMs strongly agreed with the tweet, describing the tension between being non-consensus in product and consensus in fundraising narratives.

    • Founders must raise again within ~18–24 months; perception matters
    • Being publicly ‘passed on’ can hurt future fundraising
    • Founders: need product-market alpha, but must sell a consensus story
    • Investor pattern-matching drives founder frustration and anxiety
    • Different audiences interpreted the tweet through identity/status lenses
  6. 13:08 – 17:26

    Indigestion vs. starvation: when consensus capital becomes harmful

    Leo argues non-consensus fundraising can force healthy frugality, while consensus rounds can lead to weak diligence and reckless spending. Martín agrees, adding his belief that many companies fail from ‘indigestion’—raising too much too easily—citing the 2021 era as likely a major capital wipeout cohort.

    • Hard-to-raise companies often become cash-efficient by necessity
    • Consensus/hot markets can create “house of cards” fragility
    • Weak diligence failure mode: ‘brand-led’ markups with little verification
    • Martín’s maxim: companies often fail from indigestion, not starvation
    • 2021 mega-rounds (e.g., huge Series Bs) likely produced heavy losses
  7. 17:26 – 19:54

    Cycles and the AI craze: efficiency on average, bubbles at the edges

    They reconcile two truths: markets can be broadly efficient over time, but still show bubbles and neglected sectors during cycles. Martín points to AI attracting speculative capital while strong non-AI infra struggles; Leo references dot-com vs. 2010-era vintages as evidence that sentiment shifts strongly affect returns.

    • Two persistent failure modes: bubbly consensus and undue pessimism
    • AI cycle: speculative raises in unclear models; good ‘off-theme’ companies starve
    • Despite hype, leaders like OpenAI/Anthropic show real demand signals
    • Venture vintage data: dot-com median funds bad; 2010 era rewarded contrarians
    • Market efficiency may improve on average even as extremes persist
  8. 19:54 – 23:19

    Martín’s startup story: exuberance, downturn, then consensus returns

    Martín recounts his own company’s fundraising journey: early term-sheet frenzy, a 2008 crash where funding vanished, then renewed interest as signs emerged, culminating in a major acquisition that later proved highly valuable inside the acquirer. The story illustrates how market sentiment, timing, and emerging signals interact—and how early exuberance can look irrational but still align with eventual outcomes.

    • 2007: hot seed at a high price (10M post) before clear direction
    • 2008 recession: couldn’t raise; near-bankruptcy; harsh investor feedback
    • Recovery: hot round returns as signs of life appear; later rounds become very hot
    • Acquisition looked expensive on current metrics but proved strategic long-term
    • Interpretation ambiguity: luck/overexuberance vs. market sensing latent potential
  9. 23:19 – 26:19

    How non-consensus becomes consensus: milestones, follow-on narratives, and deep tech

    Martín presses Leo on what bridges the gap from a lonely seed to broad validation. Leo explains he underwrites milestone-based de-risking: early checks fund progress that makes a later, larger round plausible—especially in deep tech where the ‘asset’ often won’t be fully working by Series A.

    • Key question: do you bet on business working soon, or on milestones that attract capital?
    • Deep tech: product often still being built at Series A/B; milestones matter most
    • Fundraising plans must match realistic next-round expectations (10M vs 50–100M)
    • Harder bet: needing a ‘top 5%’ mega Series A after a small seed
    • Investors must model what future investors will require to say yes
  10. 26:19 – 28:20

    AI vs. deep tech/humanoids: growth speed, weaker moats, and valuation hype

    Leo notes AI companies can grow faster than the old ‘triple-triple-double’ playbook, but durability and moats feel more uncertain. They discuss hype cycles in defense, bio, and robotics (especially humanoids), where valuations can surge without corresponding fundamental changes, affecting opportunity cost and investor discipline.

    • AI: unprecedented growth speed, but outcomes may be less durable due to weak moats
    • Deep tech cycles: defense valuations jumped post-conflict without fundamental shift
    • Humanoids: extreme hype and capital concentration makes new entrants harder
    • Opportunity cost framing: invest in hyped sector at 40 or quieter sector at 15?
    • Consensus areas often get implicitly avoided due to pricing and crowding
  11. 28:20 – 33:38

    TAM traps and unit economics: why “infinite markets” distort venture thinking

    Martín and Leo critique the tendency to justify any valuation with a massive TAM (e.g., ‘human labor is trillions’). Martín argues unit economics and standalone business viability must anchor investment theses, citing autonomous vehicles as an example where enormous spend hasn’t yielded venture-grade economics for most startups.

    • Huge TAM can make ‘any price’ seem rational, warping decision-making
    • Unit economics as the discipline that prevents hype-driven investing
    • Autonomous vehicles: $100B invested, yet economics may not beat Uber-like benchmarks
    • Two styles: invest for standalone scale vs. invest for likely acquisition outcomes
    • Picks-and-shovels can be a better way to play uncertain unit economics
  12. 33:38 – 43:07

    Outcome expansion, fund size mechanics, and why prices may still be ‘too low’

    Erik and Martín argue that because outcomes are now orders of magnitude larger, investing at higher prices can still deliver venture-like returns, and being in the winner matters more than bargain pricing. This leads to a debate about whether prices are constrained less by intrinsic value and more by fund mechanics and LP capital access—echoing SoftBank/Tiger-era experiments.

    • Bigger outcomes can justify paying ‘seed-like’ multiples at Series A/B prices
    • Martín: top returns imply some winners were underpriced even at high valuations
    • Constraint may be fund mechanics/LP capital, not lack of opportunity
    • SoftBank/Tiger tested ‘bigger checks/higher prices’; results were mixed for many reasons
    • Key claim: for top outcomes, price and ownership matter less than access to the winner
  13. 43:07 – 50:25

    Venture identity vs. market reality: efficiency, cost of capital, and why venture isn’t dead

    Erik reframes the debate as ‘competitive vs. non-competitive rounds’ and ‘working vs. not working’ rather than identity-laden consensus language. Leo warns a fully consensus world becomes a cost-of-capital game where the lowest required return wins, while Martín argues more venture capital is broadly positive because it funds growth and creative destruction over incumbent preservation.

    • If markets get more efficient, firms that can’t win deals will struggle
    • Consensus language is tied to VC identity; better frames: competitiveness and traction
    • Fully consensus markets risk devolving into cost-of-capital competition
    • Martín: public markets prioritize predictability; venture uniquely funds innovation
    • Ecosystem lens: more capital + competition can accelerate breakthroughs (e.g., healthcare)
  14. 50:25 – 55:51

    What they’ll test next: cohort pricing vs. outcomes, and stage-based differences

    Martín outlines a planned data analysis: compare winners vs. non-winners by relative pricing versus stage medians, and determine whether most returns come from companies priced above median. They close by acknowledging that check size and stage shape the feasibility of non-consensus investing, and briefly discuss whether multi-stage firms have structural advantages in seed for certain founder profiles.

    • Planned metrics: (1) were winners priced above/below median by stage? (2) where did returns concentrate?
    • Shared takeaway: don’t over-index on price arbitrage; missing great companies is costly
    • Stage and check size matter: large-check investing makes non-consensus harder
    • Multi-stage advantage at seed is strongest for proven founders and obvious spaces
    • Agreement to revisit once numbers are computed

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.