Marc Andreessen, co-founder of Andreessen Horowitz (a16z), discusses AI’s transformative impact on software, the venture capital model, crypto investing, the future of Big Tech, and the structural dynamics of innovation in a managerial economy.
He frames venture capital as a modern version of an ancient pattern — “project picking” — where sharp judgment, taste, and financing are directed at risky, path-dependent ventures, much like whaling expeditions in the 1600s.
A central theme is the tension between “bourgeois capitalism” (founder-led, owner-controlled companies that build new things) and “managerial capitalism” (professional managers running large organizations at scale but rarely innovating). Andreessen argues that venture capital exists to keep the bourgeois model alive, funding the 1% of the economy that produces genuine innovation.
AI and the Future of Software
AI is currently the biggest bottleneck and opportunity in software development.
The traditional model of apps — databases, forms, known UI patterns — may be completely upended.
Future applications may be primarily dialogues between humans and machines (text, speech, or other forms), with the human guiding the machine in real time and receiving results iteratively.
This implies an entirely new software stack needs to be built, and every application category could be reinvented in the next five years.
a16z does not have a separate AI fund because AI is viewed as the core of software itself, not a peripheral category.
Regrets and Counterfactuals
Andreessen deliberately avoids revisiting past decisions (e.g., the Netscape and Opsware exits).
He views startups as “path-dependent courses through complex adaptive systems” — small random events lead to wildly different outcomes, making counterfactual analysis unproductive.
His advice to founders: don’t mire yourself in “what ifs”; take your skills and do something new.
Managerial Capitalism and Venture Capital’s Role
Drawing on James Burnham’s The Managerial Revolution, Andreessen distinguishes two forms of capitalism:
Bourgeois capitalism: Owner = builder = CEO. Direct link between ownership and control. This is the model of Ford, Edison, Jobs, Page, Zuckerberg.
Managerial capitalism: Ownership (dispersed shareholders) is separated from control (professional managers). Managers run things at scale but are structurally disinclined to build new things because their incentives favor stability and risk avoidance.
Venture capital’s role is to fund “new bourgeois capitalists” — tech founders — within a system that is otherwise 99% managerial.
Startups are the mechanism by which new things get built; large managerial companies almost never do this.
If venture capital were eliminated, innovation would effectively stop.
a16z itself is run as a bourgeois capitalist entity: Andreessen and co-founder Ben Horowitz control the firm directly, with no board of directors or outside shareholders.
The succession problem is acknowledged: when founder-CEOs hand off to professional managers (which happens almost universally), the company typically stops building new things, which in turn creates the opportunity for the next generation of startups.
100-Year Fund and Basic Research
A 100-year lockup period would not be useful because ventures need regular contact with reality — market feedback, customer validation — to avoid becoming disconnected “bubbles of their own reality.”
Companies that go into long tunnels without market contact generally fail to produce anything valuable.
A more interesting question: what if venture capital could write much larger checks (e.g., $2 billion instead of $20 million)?
Tesla and SpaceX are examples of world-changing companies that required massive capital, though they raised it round by round within the existing system.
On basic research: Bill Janeway’s thesis is that venture capital has only worked in sectors (computing, biotech) where decades of prior federal R&D created a backlog of productizable knowledge.
Andreessen is somewhat more optimistic, arguing that computer science now applies disruptively across many sectors (finance, healthcare, etc.), potentially unlocking innovation even in areas without massive prior federal research investment.
He is open to experimenting with larger-scale venture funding for ambitious projects but sees the number of great entrepreneurs and great ideas as the real bottleneck, not capital.
Crypto Investing
a16z applies the classic venture capital playbook to crypto: backing founders with vision and determination, looking for deep technological/economic change, and focusing on products with real market demand.
They are agnostic between investing in companies, tokenized networks, or hybrids.
They take a long-term buy-and-hold approach (5–20 years), in contrast to many crypto investors who trade tokens frequently based on daily price signals.
On speculation and NFTs:
Andreessen pushes back on the criticism that NFTs (particularly digital art/collectibles) are mere speculation with no underlying value.
He argues that art has always been part of the economy, that aesthetic and cultural value are real forms of value, and that “speculation” in the sense of estimating future value is how all markets work.
He rejects the framing that society must choose between funding utilitarian innovation (flying cars) and funding art — the modern economy has a massive oversupply of capital and not enough viable projects of either type.
The Future of Venture Capital
The core activity of venture capital — “project picking” (a term borrowed from Tyler Cowen) — is ancient and likely permanent.
The pattern: risky, complex, path-dependent ventures led by idiosyncratic people require someone in the background exercising judgment, providing endorsement, and supplying financing.
This is directly analogous to how whaling expeditions were funded in the 1600s (the term “carried interest” literally comes from the share of whale carcings the financier received).
The public/private distinction is already dissolving: private company stocks increasingly trade on secondary exchanges, crypto tokens represent underlying value in new ways, and the regulatory framework around accredited investors is under ongoing debate.
Andreessen expects the boundaries to continue fuzzing out over the coming decades.
Why VCs Don’t Become Entrepreneurs
Building a company is a “full-contact sport” with a hundred decisions a day, requiring a bias to action, comfort with aggression, and tolerance for constant real-world chaos.
Venture capital is more analytical, clinical, and outside-in, with longer decision cycles.
VCs who have become comfortable in their careers are often unwilling to return to the “chewing glass” intensity of startup life.
Scaling to CEO of a Large Company
The critical breakthrough for running a large organization is learning to manage managers (not just individual contributors).
This skill is generalizable and scalable: once you can manage managers, you can theoretically scale to any organizational size.
Of the founders a16z backs, Andreessen estimates 100% have the intelligence to run a large company, about half have the temperament, and about half of those actually want to do it long-term.
The ideal scenario: an entrepreneurial (bourgeois) CEO managing a team of professional managers — combining the ability to build new things with the ability to operate at scale.
a16z Vulnerabilities
Over a 20-year horizon, the most likely cause of mediocre returns would be bad investment decisions (picking the wrong companies) or a broad technological drought (insufficient basic research paying off in investable sectors).
A sector-specific failure (e.g., crypto being regulated away) would not sink the firm because a16z is diversified across six primary investment domains.
The scariest scenario is an exogenous one: innovation being effectively outlawed in major areas (as has happened with nuclear energy in the US, where no new reactor design has been authorized in ~50 years).
Monetizing Twitter
Andreessen was an early angel investor in Twitter and has always believed the public follow graph should be “titanically valuable.”
The platform is where creators, politicians, and brands interact with audiences, yet it has historically failed to capture the economic activity it enables (e.g., concert ticket sales, video monetization, event promotion flowing to other platforms).
He cites Conan O’Brien selling out a 40-city tour in two hours via a single Twitter post — but the transaction happened elsewhere. The question is why Twitter doesn’t capture that value directly.
He notes that Erdoğan once called Twitter “the primary challenge to the survival of any political regime,” which Andreessen took as more credible testimony about social media’s power than the dismissive consensus of Western intellectuals at the time.
He believes we are still very early in the arc of social media’s impact, comparing it to the printing press, which took 200 years to fully play out.
Future of Big Tech
Technology as a share of GDP will only continue to rise as tech infuses every sector.
Most of this growth will come from entrepreneurial (venture-backed) companies because incumbents are structurally poor at building new things.
The opportunity set is shifting: tech is now going after large, less dynamic sectors of GDP — education, healthcare, real estate, finance, law, government — which are harder to disrupt but represent much larger markets.
On education specifically:
Andreessen is deeply critical of the modern university system, describing it as a self-governing cartel that has eliminated intelligence signals (SAT/ACT), become intensely politicized, and produces research that largely fails to replicate.
Cost is soaring while value is declining; he sees a coming educational revolution, though its form and timing are uncertain.
On healthcare:
Despite massive spending, health outcomes have barely moved. He sees technology as the primary path to better outcomes for people who care about the reality of their health.
Is Venture Capital Overstaffed?
Andreessen’s friend Andy Rachleff estimates venture capital is overfunded and overstaffed by roughly a factor of 5.
The sector should probably be 20% of its current size in terms of people, funds, and capital.
The root cause is a global savings glut — too much capital chasing too few productive investment opportunities — combined with the Swensen model of institutional asset allocation, which assigns a percentage to venture capital as an alternative asset class.
Every LP claims to invest only in the top 10 funds, but everyone has a different list.
A structural feature that sustains overfunding: venture capital’s long time horizon means bad investments take years to be judged, giving the sector a “ten-year lease on life.”
On the winner’s curse:
At early stages (seed through Series B), the best companies often raise at below-market prices because the signal of who invests matters more than the absolute valuation.
Entrepreneurs who take the highest bidder often regret it, as those investors may panic at the wrong times or fail to understand the nature of a startup journey — creating a self-correcting equilibrium.
