Marc Andreessen

Andreessen’s core labor call is that AI will expand productive work because it sharply raises each worker’s marginal output. He says leading-edge programmers are already estimated to be 20x more productive than a year earlier, gaining compensation leverage while becoming sleepless, euphoric “AI vampires.” His macro evidence is directional rather than conclusive: despite federal employment falling by estimates of up to 400,000, private-sector gains reportedly kept the latest quarter positive.

AI-attributed layoffs may be clearing years of organizational bloat rather than proving that aggregate employment must collapse. Andreessen has long judged companies to be 2x–4x overstaffed; respondents told him some institutions were closer to 8x. Torenberg cited Twitter as having cut 70%–80% and continued running as well as before; Andreessen said he believed the reduction later went far deeper—“the number has a nine on it,” perhaps the high nineties. He concedes AI reduces headcount for a fixed quantity of code, but argues companies will instead produce far more software and products.

The software org chart could collapse into a broad “builder” role spanning programming, product management, and design. Andreessen sees a “three-way Mexican standoff” in which each discipline believes AI lets it replace the other two—and thinks “they’re all correct.” The coder title might disappear within 10 or 20 years while the number of people building complete products explodes, provided the transition is allowed to happen; Europe, in his view, is running the opposite experiment through a “100% self-inflicted wound.”

Investors evaluating AI adoption should assume that impressions formed even six months ago may already be stale. Andreessen contrasts the entertaining but unreliable GPT-2-to-GPT-4 era with May 2026’s “extraordinary” GPT-5.5, reasoning models, RL post-training, agents, and Codex’s “goal” feature for projects lasting 24 hours or more. His practical benchmark is state-of-the-art access—roughly $200 for a premium package—not a free or bundled model.

Negative AI polling is less informative than revealed behavior, where usage, churn, consumption, and revenue remain the harder signals. Torenberg raised an apparent 30% sentiment or NPS figure, but Andreessen separated general opinion from net promoter score and argued that actual AI products show high NPS, declining churn, and rising recurring consumption. Recalling a better-constructed issue-ranking poll, he said AI placed only 29th: people use it enthusiastically while worrying more about housing, energy, crime, schools, addiction, and health.

The episode’s institutional thesis is that declared missions can conceal incentive loops that manufacture the problem being funded. Andreessen rejects “suicidal empathy” as too charitable when activists gain status, money, and power while attacking opponents; in discussing extraordinary allegations against the SPLC, he repeatedly stressed that they remained allegations and that the organization had not had a chance to present its defense. His due-diligence question is blunt: “What did their donors know?”

The most advantaged labor cohort may be AI-native juniors, not senior workers protected from automation. Andreessen tells graduates to make AI use visible in portfolios and interviews, predicts 14-, 18-, and 24-year-old “super producers,” and rejects the claim that companies will stop hiring juniors. Older workers who adopt the tools can also thrive, but young people inherit both native fluency and a deep skepticism of institutional authority.

AI risk narratives can themselves become part of the system they warn about. Torenberg applied the “golden algorithm”—fear producing precisely the feared result—to Anthropic’s reported blackmail behavior; Andreessen, while admitting he had not read the underlying material, said Anthropic’s thread traced it to AI-doomer literature in the training data. His reductio: if the goal is not to create killer-AI behavior, “don’t train it on all the data” describing exactly that behavior—the “call is coming from inside the house.”

Old-media defense has become a liability: a16z’s operating rule is to “say interesting things and flood the zone.” Ben Horowitz contrasts an early leaked-returns story—so damaging he is “pretty sure Balaji quit because he thought we were dead”—with today’s ability to answer through 30 podcasts with bigger audiences or simply replace the controversy with something more interesting. “The laws of physics are completely different,” so organizations cannot operate half in the old regime and half in the new one.

Founder-led companies have a structural media advantage because audiences now reward visible people, original ideas and full explanations. Marc Andreessen argues that the abstract corporate brand produced executives trained to “say absolutely nothing,” while Ben says professional CEOs are often selected for “what’s not wrong with him” rather than what is right. The emerging ideal is a “Joe Rogan CEO”—someone like Alex Karp or Palmer Luckey who can sustain three hours of interesting, inevitably controversial conversation.

The internet’s dominant unit is the viral post, with roughly 12 hours of ascent, 24 hours of decay and near-erasure by hour 36. Marc updates Marshall McLuhan’s “if it’s on TV, it’s a television show” into “if it’s on the internet, it’s a viral internet post”: emotional enough to make people “flip out,” selected by users, and rapidly displaced by the next “current thing.” Traditional outlets cannot move fast enough and are relegated to covering yesterday’s internet.

For a16z, target-audience density matters more than aggregate reach. TikTok and Instagram reach the global mass market, but Marc and Ben say the leading edge of AI, crypto, technology and policy formation largely lives on X. For a16z, reaching “90% of founders as opposed to 4% of the world” is the superior distribution outcome.

When speed matters, new media can convert decision speed into competitive power through the OODA loop: observe, orient, decide, act. Marc’s example is stark: if one organization completes the loop in one hour while another needs two, every action forces the slower rival to restart before it can decide. Sustained speed lets a company get “inside the other guy’s loop,” producing defensiveness and ultimately “psychological breakdown”; he cites Elon, Anduril and a16z as practitioners.

Format determines cognition, making generic cross-posting a strategic mistake. Marc classifies short written tweets as oral culture because they are bursty and emotional, while three-hour spoken podcasts behave like written culture because they permit abstraction and rigor. A16z therefore assigns platform-native operators to each channel; its Instagram account was growing 35% month over month under an 18-year-old specialist who “knows it like the back of his hand.”

A16z is productizing distribution as a portfolio service, not treating media as corporate communications. Its “launch as a service” package covers messaging, social copy, rollout and custom video, generating “millions and millions of views”; its founder “go-direct” work helped Applied Intuition’s founder earn about 4,000 likes on his first tweet. A new-media fellowship drew 2,000 applications for 65 places and has already produced two a16z hires, building a scarce talent pool that is both “online enough” and professional enough to operate inside companies.

Andreessen’s base case is that AI is only three years into an 80-year revolution, with revenue and demand validating the shift. He ranks it above the internet and alongside the microprocessor, electricity and the steam engine; leading companies are converting demand into banked dollars at an “absolutely unprecedented takeoff rate.” Expect fits, starts and failed economics, but today’s products look primitive versus what users may have in five or 10 years.

AI can reach five billion to six billion internet users without rebuilding the physical distribution network, making adoption unusually fast and monetization unusually visible. “You couldn’t download electricity,” but consumers can download AI onto smartphones that cost as little as $10; some products already offer $200 or $300 monthly tiers. Polls may show panic, yet revealed preferences show people using AI for work, health and relationships—and “they love this technology.”

The enterprise thesis rests on intelligence producing measurable business value while its unit price falls faster than Moore’s Law. Better service, upselling, retention, marketing and AI-native products create direct payoffs; meanwhile, “tokens by the drink” get cheaper, shortages summon new capacity, and hundreds of billions—perhaps trillions—are entering infrastructure. Andreessen expects elasticity to turn collapsing costs into more-than-offsetting demand.

The likely model market is a pyramid, not a winner-take-all endpoint. Frontier “god models” may remain the smartest, while smaller models reproduce their capabilities six to 12 months later and proliferate into local and embedded systems. The sharpest specimen is Kimi: according to early benchmarks, its reasoning approximated GPT-5 and could reportedly run on one or two MacBooks—“another Tuesday, another huge advance.”

Nvidia’s exceptional profits are also “the bat signal of all time” for competing silicon. Andreessen expects AMD, hyperscaler-designed chips, Chinese chipmakers and AI-specific startups to make chips “cheap and plentiful” in roughly five years compared with today. GPUs won by historical happenstance and parallel-processing fit; purpose-built AI architectures could be more economically efficient and pressure costs further.

The strategic AI race is now visibly US–China, but it is more economically entangled than the US–Soviet contest. DeepSeek’s emergence from a quant hedge fund, followed by Qwen, Kimi and other Chinese models, showed that less-resourced entrants can catch up quickly; robotics may favor China because its electromechanical supply chain is already there. Andreessen allows that open-source releases may constitute subsidized “dumping,” but their larger effect is forcing Washington to treat AI leadership as a two-horse race.

The largest near-term US policy risk has moved from federal restriction to roughly 200 state bills spanning red and blue states. California’s vetoed SB 1047 illustrates the stakes: downstream liability could have made an open-source developer responsible for a misuse years later, effectively crushing startup and academic releases. Andreessen expects federal primacy eventually, while conceding that the failed moratorium was too broad to preserve legitimate state authority.

Application startups may capture more value than the dismissive “GPT wrapper” label implies, but the winning pricing and market structures remain “trillion dollar questions, not answers.” Products such as Cursor can orchestrate dozens of models, build their own and substitute open source; vendors can price against labor replaced or productivity created rather than token cost. A company must choose one coherent strategy, while venture can own contradictory bets across big and small, closed and open, foundation and application, consumer and enterprise—creating “multiple ways to win.”

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.

DeepSeek R1 resets AI’s cost curve by making frontier reasoning open, self-hostable, and roughly 30X cheaper. Marc says the full model can run on about $6,000 of hardware, while distilled versions already run on MacBooks and iPhones; V3 and R1 also ship with code and technical papers that become a roadmap for competitors. His categorical forecast: within five years, everyone will have a “superhuman AI lawyer” and “AI doctor” on their phone.

The immediate winners are AI users, while proprietary model labs and NVIDIA face price pressure—but Jevons Paradox could ultimately lift the entire stack. OpenAI and Anthropic face a free, open alternative, while DeepSeek’s efficient use of cheaper chips challenges NVIDIA’s margins of as much as 90%. Yet a 30X price decline might drive 30X, 100X, or 1,000X more usage: reasoning is “fire being brought down from the mountain,” and demand could outrun falling unit costs.

DeepSeek turns U.S.–China competition into a contest over which country can offer the world the more open technology stack. China’s “judo move” is that an ostensibly closed system now offers the best open-source system while American companies called OpenAI and Anthropic remain closed. Because AI will enter healthcare, education, transportation, telecom, energy, and national infrastructure, Marc argues the West must answer by returning to “freedom and democracy and openness,” not by adopting Chinese-style control at home.

Attempts to contain China can force exactly the innovations and competitors U.S. policy seeks to prevent. Chip restrictions helped motivate DeepSeek to optimize cheaper hardware, encourage a parallel Chinese semiconductor industry, and may accelerate the timeline for a Chinese invasion of Taiwan because of its “magic fabs.” Copyright presents the same dilemma: if Chinese models may train on every book while U.S. models are prohibited, Marc warns that “they win and we lose,” though DeepSeek’s actual training corpus remains undisclosed.

AI should commoditize investment analysis, but Marc doubts it will soon replace early-stage venture’s interpersonal “alchemy.” Models such as o1, o3, o7, or R4 will probably read annual reports better than most humans, but that advantage becomes standard once every investor adopts it. Founder evaluation, coaching, and “project picking” remain face-to-face judgments developed over years; Marc’s hedge is refreshingly falsifiable: “If I meet the algorithm that can do that better than I can, I will instantly retire.”

Andreessen Horowitz is being built as an enduring operating company while keeping early-stage venture non-negotiably at its core. The firm pairs timeless principles—respect founders, know the details, avoid “drive-bys”—with rapid adaptation, including a major political operation that did not exist four years earlier. It rejects becoming an asset gatherer harvesting fees: a $5 million Series A can offer as much return opportunity and strategic learning as a $500 million later-stage investment.

Politics has become the fourth—and recently the dominant—constraint on technological returns. Marc’s original framework aligned technology, social readiness, and capital markets; government now overlays all three, and he describes the past four years’ treatment of crypto, fintech, social media, and AI as the largest bottleneck. His broader power thesis is that collapsing institutional trust produces counter-elites: “a small organized elite” always governs a large disorganized mass until a rival group persuades that mass it can represent them better.

China’s phones-to-drones-to-cars-to-robots supply-chain ladder makes physical technology the next geopolitical pressure point. China holds “ninety-nine point something percent” of drones, Unitree sells a $1,500 robot dog against Boston Dynamics’ $50,000–$100,000 equivalent, and humanoids might arrive at $10,000–$20,000—extraordinary abundance with embedded security risk. In defense, one planner told Marc that backpack suicide drones are the biggest innovation since the stirrup: 20 trained soldiers with drones might hold off thousands or tens of thousands of conventional troops.

Andreessen’s upside case is a U.S. “roaring ’20s” driven by energy abundance, imported and domestic talent, AI leadership, and a sharp release of regulatory pressure. While Canada, the UK, and Germany have stalled by his telling, America retained growth, physical security, resources, entrepreneurial culture, and leadership in software, AI, and biotech. His investable call is that removing the “government boot off the neck” could lift productivity, technology adoption, business formation, and national confidence together.

The immediate catalyst is a political and corporate “vibe shift” that Andreessen sees spreading from Trump’s return through Meta, Hollywood, finance, and the wider technology industry. He points to Zuckerberg’s policy changes, Hollywood figures saying “the ice has thawed,” BlackRock retreating from ESG commitments, and the Court of Appeals striking down Nasdaq board-composition rules. His rough preference-falsification model is 20% believers, 60% conformists, and 20% dissenters; the dissent waits until someone safely says, “The emperor is actually wearing no clothes.”

DOGE’s consequential target is not merely waste but the linked machinery of money, personnel, and regulations, many of which Andreessen says agencies created rather than Congress directly legislating. He cites roughly 450-520 federal agencies, about 4 million employees, potentially 20 million contractors, and debt rising by $1 trillion every 100 days—soon, he warns, every 90, 80, then 70 days. Social-media transparency could make previously invisible $30 billion failures politically legible: “If you’re this cavalier about $30 billion, imagine how cavalier you’re about $3 trillion.”

Andreessen argues that censorship evolved from necessary restrictions into a reusable “ring of power” for suppressing legitimate speech, with government coercion crossing into what he calls “flagrant criminality.” Hate-speech rules expanded from prohibiting slurs to policing discomfort; misinformation enforcement culminated in suppressing the lab-leak hypothesis; and informal threats often mattered more than legislation. He cites 18 U.S.C. §§241 and 242, First Amendment and due-process concerns, and debanking as conduct he says can constitute a deprivation of rights, calling for transparency, investigations, and potentially prosecutions.

His revised immigration position keeps the case for exceptional foreign talent but rejects discussing it separately from the exclusion of native-born Americans. O-1 visas often serve breakthrough-company founders, while H-1Bs now concentrate among large technology companies and consulting mills paying roughly $60,000-$100,000; meanwhile, he sees no serious recruiting pipeline from the Midwest, South, rural communities, or overlooked Black, Asian, Jewish, and white Americans. The politically durable answer is “what if we did both?”—welcome extraordinary immigrants while actively scouting and developing domestic talent.

AI coding is Andreessen’s clearest near-term technology conviction: software creation becomes radically cheaper, yet demand expands fast enough that engineering employment may grow rather than collapse. Coding can be validated, reinforced, and improved with synthetic data; AI already writes glue code, explains codebases, accelerates learning, and lets smaller teams create products with fewer traditional production roles. Because code has unusually high elasticity, he predicts “more coding jobs probably by like an order of magnitude 10 years from now,” though the work shifts toward orchestrating agents.

The AI market remains governed by several unresolved “trillion-dollar questions,” so Andreessen refuses to declare a winner among OpenAI, Meta, Google, xAI, Amazon, and others. The fault lines are large versus small models, open versus closed, synthetic data, chain-of-thought and reinforcement learning, hallucination control, chips, financing, censorship, and whether AI-native companies can replace incumbents built around pre-AI economics. His sharper product principle is architectural: build where AI is the first premise, not “the sixth bullet point,” while viewing open models, crypto infrastructure, and contrarian critiques such as Yann LeCun’s as important parts of the landscape.