a16zMarc Andreessen Reveals His Biggest Wins and Mistakes at a16z
CHAPTERS
Starting a16z in the wreckage of 2008–2009
Andreessen frames company-building as “falling up the stairs,” then recounts launching Andreessen Horowitz at the depths of the financial crisis. He describes the first fundraise, the tiny early LP meeting, and why starting a VC firm in 2009 was a contrarian bet.
- •First fund raised in 2009; first LP meeting in 2010 with “folding chairs” and ~20 LPs
- •2009 was exceptionally hostile for new VC funds (only two raised that year)
- •Starting a firm then required conviction amid market pessimism and capital scarcity
- •Early supporters/LPs remained long-term partners as the firm scaled
Post–dot-com cynicism and why early social tech looked “useless”
He revisits the hangover from the dot-com crash and how any promising tech signal in the mid-2000s was attacked as “bubble 2.0.” The public narrative wasn’t ‘tech is dangerous’ yet—it was that consumer internet products were trivial and couldn’t monetize.
- •Dot-com crash shaped years of negative press and investor skepticism
- •Yahoo’s mid-2000s acquisitions (Flickr/Delicious) triggered “bubble” backlash
- •Early Facebook/social media dismissed as pointless (“what did your cat have for breakfast?”)
- •Monetization doubts dominated discourse before later political/harms narratives emerged
The Yahoo–Facebook near-acquisition and the underestimated upside
Andreessen details how Yahoo nearly bought Facebook for $1B, and how the financial crisis led Yahoo to renegotiate—giving Zuckerberg the opening to walk away. He notes how even bullish internal analyses drastically underestimated Facebook’s eventual growth and business model evolution.
- •Yahoo’s $1B deal for Facebook was close/structured before ad markets collapsed
- •Recession-driven renegotiation attempt helped trigger Facebook’s decision to stay independent
- •Leaked Yahoo deck later showed dramatic underestimation of Facebook’s future scale
- •Early Facebook revenue came from low-CPM remnant ads (pre-targeting era)
Mobile “kills Facebook”… then targeted ads change the math
He recounts the 2012 IPO-era narrative that desktop-to-mobile would crush Facebook because smaller screens meant fewer ad pixels. Reality flipped: mobile increased usage frequency and made targeting more valuable, reinforcing his view that winners survive repeated ‘new things to prove.’
- •Post-IPO skepticism centered on reduced mobile screen real estate and ad inventory
- •Mobile drove higher engagement throughout the day, expanding total opportunity
- •Targeted advertising unlocked outsized monetization on mobile
- •“Falling up the stairs”: even ‘inevitable’ winners face repeated existential narratives
From “ads don’t work” to “mind control”: the politicization whiplash
Andreessen highlights the rapid shift from believing social ads were ineffective to claiming they were powerful enough to manipulate elections. He critiques the internal inconsistency of the Cambridge Analytica/Russian-ad narrative and contrasts it with everyday ad-market realities.
- •Narrative swing: useless tech → democracy-saving (Arab Spring/Obama) → evil mind-control (2016)
- •Claimed Russian spend (~$80K during key period) vs. massive US campaign budgets
- •Skepticism that “psychometric mind control” was real or repeatable (not used today)
- •Questions why alleged effectiveness wasn’t reflected in broader consumer advertising pricing
Counterfactual tech history: the ‘paths not taken’ in major outcomes
The conversation broadens into near-miss deals across tech history, emphasizing how contingent outcomes can be on small decisions. Andreessen argues some forces were inevitable (e.g., smartphones), but specific category leaders often hinge on micro-level choices and timing.
- •Examples: Netscape nearly bought Yahoo; Yahoo nearly bought Google; Netflix nearly sold to Blockbuster
- •“World beaters” often have a pivotal near-acquisition or strategic fork
- •Macro forces create inevitabilities, but company-specific outcomes are fragile
- •Venture outcomes are highly sensitive to decision-making at key moments
Why a16z aimed to be stage-agnostic and scale-aware from day one
Andreessen explains how the classic VC model (A/B + mezzanine then IPO) broke as companies could compound to far larger outcomes. a16z’s founding thesis emphasized the total opportunity across stages—seed through growth—because later rounds could still produce venture-scale returns.
- •Old model: limited private rounds, IPO at ~50M revenue and ~$500M market cap
- •Realization since the ’90s: some companies “rip” far beyond early-stage returns
- •Rise of growth investing (e.g., DST’s Facebook round) reshaped funding arcs
- •a16z pitch: stage matters less than capturing ownership in compounding winners
The shift from generalists to specialized funds (and why vertical depth matters)
He describes how a16z started as a generalist firm (consumer + enterprise) but adapted as tech moved from tools to “full stack” companies embedded in specific industries. Specialization became necessary to distinguish winners in complex domains where you can’t hedge across competitors.
- •Generalist model worked when tech was mostly horizontal tools (databases, routers, OS)
- •2010s wave: Uber/Lyft, Airbnb, Tesla, SpaceX—direct insertion into vertical markets
- •Venture constraint: investing in one company blocks investing in its rivals (conflicts)
- •Domain expertise helps pick the winning team within a chosen space
- •Verticalization occurred in phases (partial ~2013; more complete by ~2017)
Domain expertise vs. the coming ‘AI generalist’: can tools replace depth?
Andreessen argues founders still need deep technical understanding, even when they present as design-led or business-led. He explores whether AI could change that by making expertise and coding accessible—raising the possibility of ‘ideas guys’ managing fleets of AI agents.
- •Successful “design founders” often also strong technologists (e.g., Pinterest, Airbnb)
- •Core debate: can AI provide on-demand depth for generalists when needed?
- •Analogy to “vibe coding” vs. building top-end software companies
- •Thought experiment: non-technical manager supervising 1,000 AI coders vs. technical leader with fewer
- •Open question: AI may shift what ‘domain expertise’ means, but outcome uncertain
Global tech enthusiasm vs. policy drag—especially in Europe
He notes worldwide interest in tech and entrepreneurship, driven by internet-distributed knowledge and aspiration. But he argues many regions, particularly Europe and the UK, suppress innovation with heavy regulation—pushing top founders to relocate to the U.S.
- •Global enthusiasm is rising; talent is widely distributed
- •Europe’s regulatory posture framed as intentional leadership in regulation vs. innovation
- •Regulatory hostility accelerates brain drain toward the U.S.
- •Defense modernization in Europe could be an opportunity—but politics and procurement complicate it
International strategy: where offices matter (go-to-market, investing, policy)
Andreessen outlines three practical international activities for a U.S. venture firm: helping portfolio companies sell globally, selectively investing abroad, and potentially engaging in foreign policy debates. He notes a16z’s general openness outside China, but also the reality that many top founders move to the U.S. anyway.
- •International focus #1: go-to-market and business development for portfolio companies
- •International focus #2: investing abroad (examples include Mistral; past investments in Vietnam, etc.)
- •International focus #3: policy engagement globally—harder due to politics/permission issues
- •U.S.-centric emphasis partly because immigration pulls global founder talent into the U.S.
Little Tech vs. Big Tech: why a16z entered politics (crypto + AI)
He explains how tech became politically salient after 2012 and especially after 2016, with backlash from both left and right. a16z’s ‘Little Tech’ framing emerged in D.C. to distinguish startups from incumbent platforms and to advocate for clear, innovation-friendly rules—particularly around crypto and AI.
- •Pre-2010: tech rarely political beyond antitrust; post-2012: rising politicization
- •2016-era dynamics: left anger (Trump/inequality) + right anger (perceived tech-left alignment)
- •Crypto and AI regulation pressures pushed a16z into active policy engagement
- •Little Tech framing: startups often compete with Big Tech; interests diverge
- •Policy goal: clarity and sensible rules, not zero regulation; support for consumer protection and fair enforcement
Silicon Valley and national security: from Vietnam-era rupture to today’s ‘vibe shift’
Andreessen traces the historical link between Silicon Valley and defense, the cultural break during the Vietnam era, and the modern return via companies like Palantir and Anduril. He frames the debate (e.g., Palantir/ICE controversy) as a serious ethical disagreement but argues tech’s role in defense is now unavoidable.
- •Early Silicon Valley roots in defense R&D (radar, avionics, early computing like SAGE)
- •Vietnam-era protests drove universities/companies to disengage from military work (e.g., SRI separation)
- •Modern resurgence: Palantir/Anduril and renewed willingness to serve government customers
- •Acknowledges moral/philosophical tension while arguing national missions matter
- •Modern geopolitics (e.g., Ukraine) makes tech central to defense and security strategy
