Marc Andreessen Reveals His Biggest Wins and Mistakes at a16z
Summary
Andreessen frames a16z’s foundational bet as launching into the 2009 wreckage, when only it and Khosla Ventures raised venture funds, then investing across stages because exceptional technology companies could compound far beyond the old $100 billion ceiling. The firm’s original pitch was explicitly seed, venture and growth; the thesis was that a B, C or D round in a “world-beating” company could still generate venture-scale returns.
Facebook illustrates why the greatest outcomes resemble “a process of falling up the stairs,” with each apparent breakthrough immediately followed by another existential test. Yahoo structured a $1 billion takeout, then tried to lower the price after advertising collapsed, giving Mark Zuckerberg the “cover” to walk away; after Facebook’s 2012 IPO, investors again misread mobile’s smaller screens as an advertising constraint, missing the resulting explosion in usage and the discovery that targeting worked.
The market’s Facebook narrative swung from useless product to impossible superweapon without a stable analytical middle. Torenberg cited claims that roughly $80,000 of Russian election-period advertising—about $140,000 in the relevant period—overpowered Hillary Clinton’s $3 billion campaign. Andreessen’s objection is causal: if the system delivered literal mind control, why did that capability seemingly work for one election and then disappear? He said the Cambridge Analytica psychometric method supposedly used for that mind control never worked.
a16z abandoned pure generalism because full-stack startups made domain-specific selection more valuable than merely spotting a hot market. A generalist can “sense heat,” Andreessen argues, but venture conflicts prevent correcting a wrong company choice within the right category. The firm partially verticalized around 2013 and completed the model around 2017, spanning specialized practices from marketplaces and defense to biotech.
AI might weaken the premium on technical and domain expertise, but Andreessen treats that as an unresolved experiment rather than a settled disruption. o3 Deep Research could give generalists depth on demand, while coding agents might let a nontechnical manager supervise 1,000 AI coders against an expert supervising 100 humans. Today, however, he sees vibe coding as compelling for casual development, not yet a substitute for the strongest technical founders—the possible “era of the ideas guy” remains conditional.
Global startup talent is abundant, but Andreessen sees policy—not human capital—as the binding constraint outside the United States. Europe keeps “shooting itself” across multiple limbs through regulation even as it is supposed to raise defense spending toward 5% of GDP; bright founders consequently relocate, letting a16z capture global talent through U.S. investing. International expansion must serve portfolio-company sales, attractive investments or, potentially, policy engagement.
a16z’s “Little Tech” agenda converts political engagement into startup advocacy as crypto and AI collide with government. The firm distinguishes startups from incumbents with “10,000 lawyers and policy people,” arguing that startups frequently attack Big Tech rather than represent it. Its stated ask is not no regulation but repeated, emphatic clarity—“We want clarity”—plus consumer protection and equal enforcement against noncompliant competitors.
Defense technology marks a generational reversal of Silicon Valley’s post-Vietnam estrangement from government. Andreessen calls the earlier break a “groove shift”; the newer view, embodied by Palantir and leaders such as Alex Karp, Brian Schimpf and Palmer Luckey, treats national-security agencies as legitimate customers and missions involving borders and protection against terror attacks as consequential. He acknowledges the moral seriousness of dissent while concluding that nearly every geopolitical problem now has a technology component investors cannot sensibly ignore.
Deep dive
1. The greatest companies keep falling up the stairs
Andreessen dates a16z’s contrarian origin to March 2009, amid financial devastation and years of reflexive “bubble 2.0” warnings. Only a16z and Khosla Ventures raised venture funds that year. By the first LP meeting in 2010, roughly 20 LPs sat on folding chairs beneath a flat-panel television; many remain with the firm.
The Facebook counterfactual crystallizes how contingent giant outcomes are. Yahoo structured a $1 billion acquisition, then tried to lower the price after advertising collapsed, giving Zuckerberg the “cover” to walk. Andreessen pairs it with Netscape nearly buying Yahoo, Yahoo nearly buying Google, Netflix nearly selling to Blockbuster and Uber nearly merging with Lyft: every world-beater has a “path not taken.”
Facebook’s 2012 IPO created another apparent dead end as usage shifted from desktop to mobile. Investors reasoned that advertising value followed screen pixels, so smaller screens meant lower rates; the stock got hammered out of the gate. What the thesis missed was that phones put Facebook into users’ hands all day, radically increasing usage, while the company learned that targeted advertising worked.
Public interpretation then overshot in the opposite direction—from ads that did nothing to ads that were “literally mind control.” Torenberg cited about $140,000 in Russian ad spend during the relevant period and $80,000 during the actual election run, versus Hillary Clinton’s $3 billion campaign. Andreessen framed the theory as $80,000 of ads swinging an election she had spent $3 billion on. He said the Cambridge Analytica psychometric method supposedly used for that mind control “never worked,” while highlighting the claim that it worked only for that election and was never tried again.
2. Venture scale followed the expanding ceiling on technology outcomes
Mid-1990s venture followed a compact ladder: an A round of roughly $3 million to $7 million, a B of $20 million to $30 million, then a mezzanine C before going public. About $50 million in revenue at a 10× multiple could support a $500 million listing; by Andreessen’s recollection, Amazon went public in 1997 at about a $400 million valuation.
A conversation with John Doerr about Cisco exposed the opportunity in later rounds. If a company could reach $100 billion, investing $50 million at Series C might produce as much absolute profit as investing $5 million at Series A. The eventual realization—that technology’s ceiling was “much, much higher” than $100 billion—made scale and follow-on capital strategic assets.
DST’s Facebook round “kind of redefined” growth investing just as a16z was being designed. The firm’s original pitch was explicitly stage-agnostic: seed, venture and growth were different entry points into one aggregate opportunity. The initial fund was not large enough to support a lot of growth investing, but Andreessen rejected the idea that later rounds automatically surrendered venture-scale returns.
Specialization followed the industry’s move from tools into end markets. Databases, routers, operating systems and boxed software had comparable company structures; Uber, Lyft, Airbnb, Tesla and SpaceX inserted technology directly into end markets. A generalist can “sense heat,” but cannot reliably select the winner—and once a venture conflict prevents investing in alternatives, the mistake cannot be repaired. a16z partially verticalized around 2013 and fully did so around 2017.
3. AI may challenge expertise, but it has not abolished it
Torenberg proposes a new arbitrage around domain experts as software becomes easier and distribution more important. Andreessen’s qualification: successful “design founders” still proved to be formidable technologists, including Ben Silbermann and Brian Chesky. “There’s really no escape from deep domain knowledge,” at least under the operating model that has prevailed so far.
The open question is whether AI makes that depth instantly accessible. If a generalist can summon the equivalent of o3 Deep Research whenever needed, one person might credibly span several specialties. Andreessen is “excited to see people try that,” preserving the uncertainty rather than claiming that an AI interface already substitutes for accumulated judgment.
Coding offers the sharper test. Vibe coding is already exciting for casual development, but the strongest technologists Andreessen knows do not yet believe it can build a top-end software company. Agents nevertheless raise a serious thought experiment: could a nontechnical manager supervising 1,000 AI coders outperform an elite technical leader supervising 100 great human coders? If so, the mocked “ideas guy” might finally contribute something.
4. Talent is global, while investability remains political
Andreessen sees knowledge and ambition spreading globally: thousands of young people attended Ben’s book events regardless of country, while online talks and podcasts made startup knowledge widely accessible. Historians may judge this era harshly for leaving so many capable people unused; startups are one mechanism for matching unusually ambitious young people with consequential work.
The offset is “incredible drag by bad governments and bad policies.” Europe’s posture is captured by a senior politician’s reported line: because it cannot lead global technology innovation, it will lead technology regulation. For a French or German founder, Andreessen imagines the response is to reach the U.S. embassy and apply for a visa.
Defense exposes the contradiction. Europe may increase spending toward 5% of GDP and would benefit from drones and other new systems rather than simply more aircraft carriers, yet debate already centers on whether purchases must remain with indigenous European vendors. Technology procurement is becoming another front in the trade war.
a16z evaluates international activity through three functions: helping portfolio companies sell globally, investing where opportunities justify it, and potentially engaging in policy. It has generally remained open outside China, including recent activity in France and past investments in Vietnam. Practically, U.S. emphasis still captures many of the best foreign founders because many high-caliber people move out of bad systems.
5. Little Tech separates startup policy from incumbent power
Technology enjoyed political “benign neglect” while it mainly produced tools; antitrust actions such as the Microsoft case were exceptional. That changed after 2012 and accelerated around Trump’s 2015–16 nomination and election. Andreessen says the left attacked technology over Trump and inequality, while the right viewed Silicon Valley as uniformly Democratic; crypto then faced a governmental “full-on war,” followed by intense AI politicization around 2021–22.
Crypto and AI convinced a16z that remaining outside politics was untenable: “We’re either going to show up and explain ourselves,” or opponents would define the industry. The resulting cleavage is Little Tech versus Big Tech. Google can deploy roughly 10,000 lawyers, policy people and compliance people; startups need representation precisely because they often exist to disrupt incumbents.
Andreessen calls venture an “escape hatch” for frustrated Big Tech employees who leave to fund competitors to their former employers. That distinction transformed conversations in Washington, where officials otherwise treated startup representatives as representatives of Google. Saying Little Tech attacks Big Tech produced an immediate icebreaker in a city he jokes agrees only on two things: hostility toward Big Tech and China.
The policy ask is “freedom to innovate” in fields lawmakers did not anticipate, but not freedom to commit fraud. Andreessen repeats, “We want clarity,” endorsing sensible rules and aggressive consumer protection. Regulation can benefit portfolio companies when it makes compliance possible and prevents unregulated rivals from gaining an advantage by ignoring obligations that responsible startups honor.
6. Silicon Valley’s defense estrangement is reversing
Andreessen traces Silicon Valley’s origins to 1920s defense work—radar, avionics and missile guidance—followed by projects such as SAGE, a giant mainframe computer for missile early-warning systems. Through the Cold War, universities, companies and defense agencies were assumed to be partners because they had to work together to build systems against the Soviets.
Vietnam produced what he calls the “groove shift.” Antiwar politics pushed universities away from military research, contributing to Stanford Research Institute’s separation from Stanford and wider resistance to activities such as ROTC recruiting. The resulting 1980s–90s culture treated defense and intelligence work as something principled technology companies naturally avoided.
Torenberg presents the Paul Graham–Palantir dispute over work for ICE as emblematic of the transition and says he disagrees with Graham. Andreessen does not question Graham’s ethics: borders and government power involve genuine moral choices. His own conclusion is that agencies are legitimate customers, companies cannot conduct a moral trial for every buyer, and missions involving borders, national security and protection against terror attacks matter.
Andreessen credits Alex Karp, Brian Schimpf and Palmer Luckey with the newer defense model, while pushing back on Karp’s claim that Palantir was first by saying Netscape was doing it 10 years earlier. Ukraine’s battlefield now makes the broader point unavoidable: nearly every topic confronting defense and intelligence leaders has a major technology component. His investment conclusion is direct—venture should be “in the middle of that.”