Growth in 2026 looks less like a playbook and more like a brawl. The most successful companies in the United States—Anthropic, Lovable, HubSpot, Meta, Amplitude, Duolingo—are growing at speeds that defy historical precedent. Anthropic went from $1 billion to $19 billion ARR in 14 months. Lovable crossed $200 million ARR in under a year with fewer than 100 people. These are not incremental wins. They are logarithmic.
Yet behind the ARR charts and funding announcements, the operators driving this growth fundamentally disagree on what works. Some have scrapped 60% of their prior playbooks. Others claim the old tactics never worked to begin with. A few are automating entire growth functions with AI. One VP of growth spends 95% of her time innovating, not optimizing—a complete inversion of the discipline as it existed five years ago.
This essay synthesizes conversations with over 40 U.S. growth leaders to surface where consensus lives, where it fractures, and what the fractures reveal about the terrain ahead.
The Automation Divide: Who Owns Growth When AI Does the Work?
At Anthropic, Amol Avasare leads what might be the hardest job in growth today. His team built a platform called CASH—Cloud Accelerates Sustainable Hypergrowth—to automate growth experimentation using Claude. The results are real. The speed is unprecedented. Yet the broader question remains unsettled: when AI does the work, who owns the function?
Leading growth inside of Anthropic is the hardest job I've had in my life. You come into Anthropic, you need to understand that 50, 60, 70% of how you operate in the past, just throw it out the door.
— Amol Avasare, Head of Growth at Anthropic
Kieran Flanagan at HubSpot goes further. He believes traditional growth teams are becoming redundant. His argument: growth historically focused on touchless onboarding, activation, upgrade flows—the product-led motion. But when a multimodal AI agent can onboard, upsell, support, and guide a customer across the entire lifecycle, the handoff between growth, sales, customer success, and support dissolves. Who owns the agent? HubSpot's answer is an "AI innovation pod" that spans go-to-market, not just product-led growth.
What happens when your entire onboarding experience, your buying experience is done through an AI multimodal agent? Who owns that experience? To me, an AI innovation pod should be the pod that's trying to integrate that experience across your go-to-market.
— Kieran Flanagan, CMO (fmr VP Growth) at HubSpot
Elena Verna at Lovable offers a contrarian take. She's thrown out 60–70% of her playbook, but not because of automation. Her team spends 95% of their time inventing new growth loops, not optimizing existing ones. In a market where everyone is launching a vibe-coding product, optimization is table stakes. Differentiation demands reinvention. Lovable gives away massive amounts of free product, builds in public, and ships things worth talking about. The trick isn't efficiency—it's creating moments that blow people's socks off.
I feel like only 30 to 40% of what I've learned in the last 15 to 20 years of being in growth transfers here because we just need to invest in such bigger bets and innovate and create new growth loops here. Right now I'm spending 95% innovating on growth and only 5% on optimization.
— Elena Verna, Head of Growth at Lovable
The divide is clean: some leaders are automating the mechanics of growth. Others are abandoning mechanics altogether in favor of narrative, velocity, and structural innovation. Both camps are winning. Neither thinks the other is wrong. That's the dissonance.
Product-Market Fit Is Dead (or at Least Dying Every 90 Days)
The concept of product-market fit as a destination—a state you achieve and then scale—is collapsing in AI-native companies. Elena Verna describes a world where Lovable must recapture product-market fit every three months. The rate of foundational model improvement, competitor launches, and user expectation shifts means what worked in Q1 is stale by Q2.
This is not hyperbole. Anthropic's product value will be 1,000x what it is today in two years, according to Amol Avasare. Linear charts, he notes, are "just not cool" inside the company. Everything is log-linear. The implication: growth operators must think in exponentials, not increments. The mental model of "find PMF, then scale" assumes a static target. In 2026, the target moves faster than most teams can instrument.
Chris Miller at HubSpot saw this early. When HubSpot launched its free CRM in 2015, it was disruptive but formless. There was no clear path from free to enterprise value. Miller's growth team took ownership of self-service revenue—a tiny percentage of total ARR that no one else wanted to touch. They "immediately blew it up." The lesson: growth teams that wait for explicit mandates miss the most asymmetric opportunities. Triangulating what the business should ask you to solve, then solving it before they ask, builds credibility and territory.
We approached the team who owned it and we were like, are y'all working on this? And they were like, nah, we're working on a bunch of other stuff. We were like, can we take this? And they were like, sure, if you want it. And so we took it and immediately blew it up.
— Christopher Miller, VP Product, Growth & AI at HubSpot
Bengali Kaba, now at YouTube, frames this as "understand work" versus "identify, justify, execute." Most teams start with a solution, then pull data to justify it. Bengali inverts the sequence: first, understand from first principles what is actually happening. Then identify the intervention. The discipline of understanding before acting is what separates growth teams that chase vanity metrics from those that rewire distribution.
First, you have to really understand from first principles what is actually going on. So understand, identify, execute.
— Bangaly Kaba, Director of Product (fmr Head of Growth) at YouTube / Instagram
The Tactics That Never Worked (But Everyone Keeps Trying)
Elena Verna compiled a list of growth tactics that never work, yet appear on roadmaps everywhere. The most provocative: hiring a growth team before you have product-market fit. Founders outsource distribution before they've proven the thing distributes. Verna is blunt—you cannot delegate finding PMF. It's founder work.
Laura Schaffer at Amplitude echoes this with onboarding friction. At one company, her team ran a rogue experiment—adding questions to the signup flow on a Friday night without approval. The expectation: conversion would drop. The result: conversion improved 5%. Why? The questions gave users a moment to articulate intent, which increased commitment. The lesson isn't "add friction." It's "most assumptions about friction are untested cargo cult behavior."
I'm fully expecting, okay, this is gonna hurt our numbers, but maybe it won't be so bad. And I totally got out, thinking of it like written, started to write a framework for how I wanted to surface this. And we start to get the data for this thing. I'm like getting an improved conversion.
— Laura Schaffer, VP of Growth at Amplitude
Verna's second never-work tactic: rebrands and website redesigns sold as growth initiatives. She's never seen one produce meaningful performance lift. New CMOs arrive, redesign the site to reflect personal taste, promise acquisition gains, and deliver nothing. The root issue: these are branding exercises disguised as growth projects. Growth is about loops, incentives, and distribution mechanics—not aesthetics.
Her third: treating every initiative as an experiment. When every roadmap item is an A/B test, paralysis sets in. Some bets are too strategic to test. Some require conviction, not statistical significance. The over-indexing on experimentation culture—once a corrective to HiPPO-driven product orgs—has become its own religion, and religions don't question their rituals.
If every single one of your initiatives that you're doing on growth is an experiment, that's a problem. It's almost like a disease, like a paralyzing disease.
— Elena Verna, Growth Advisor at Amplitude / Miro / Dropbox
Cold Emails, Canonical Docs, and the Vanishing Ladder
Amol Avasare didn't apply to Anthropic. He sent Mike Krieger a cold email. Subject line: secret. Body: brief, non-LinkedIn, personal email address. Open rate: high. Outcome: head of growth at the fastest-growing company in history. Avasare has "perfected" cold email over years as a founder. His tactic: reach people where they aren't getting outreach. Everyone emails work accounts and LinkedIn. Find the personal inbox.
This is emblematic of a broader pattern among the leaders interviewed: they carved their own paths. Bengali Kaba wrote a legendary post, "How to Choose Where to Work and What to Work On," built on the idea that impact is a function of environment variables and skill variables. You don't wait for opportunities. You construct them by choosing the right intersection of leverage and capability.
Naomi Gleit at Meta, employee 29, has been at the company for 19 years. Her superpower: taking gnarly, complex projects and simplifying them. One mechanism: canonical docs. She demands one source of truth per project. If you ask five people and get five answers, the project is broken. The canonical doc contains everything, links to everything, and everyone knows where it lives. This sounds trivial until you've been on a team where onboarding takes three weeks because documentation is archaeology.
I really believe in frameworks for things. That helps drive extreme clarity. A lot of times I'm ramping up mid-project. I'm like, where can I learn what I need to learn about this project? I ask 5 different people, get 5 different answers. That is unacceptable.
— Naomi Gleit, Head of Product (fmr Growth) at Meta
Chris Miller at HubSpot describes early growth teams as "aggressive" in the best sense. They adopted "radical accountability" and an "ownership mentality." Every problem was their problem. This posture unlocked opportunities the business hadn't explicitly prioritized. When you look hungry, people keep feeding you.
The implication: ladders are disappearing. The companies growing fastest don't wait for org charts to formalize roles. They give ownership to people who take it.
Speed as Strategy, or Why Duolingo Ships 10x Faster
Cem Kansu, CPO at Duolingo, runs a product org that ships consumer delight at scale. Ninety million monthly active learners. Pixel-perfect attention to detail. Kansu's mantra: "Consumer products live and die in the pixels." The details aren't the details—they are the product.
But Duolingo's edge isn't just craft. It's velocity. The team uses AI to build 10x faster. Kansu describes a shift from "mini CEO" product managers to operators who understand that execution is strategy. The CEO is the CEO of the product. The PM is the one who makes the pixels perfect, fast.
Consumer products live and die in the pixels. The product manager has to understand how to make your change pixel perfect for the user.
— Cem Kansu, CPO at Duolingo
This velocity thesis extends beyond product. Raaz Herzberg at Wiz helped scale the company to a $30 billion valuation in record time. Wiz's go-to-market was founder-led and bottoms-up. They sold millions of ARR per hour before hiring salespeople. The product delivered value in 15 minutes—an enterprise security tool that breaks every assumption about deployment timelines. Time to value is value. Speed is not a nice-to-have. It's the entire moat.
We always believed that time to value is value. You connect this to your environment and it's an enterprise security product. Those things are expected to take months. But in Wiz, literally, you get value in 15 minutes.
— Raaz Herzberg, CMO at Wiz
The Contrarian Consensus: Build in Public, Give Away Product, Remove Barriers
Lovable's growth strategy centers on building in public, employee and founder-led social, and giving away enormous amounts of free product. When a user offers to run a hackathon at their company, Lovable doesn't gate credits. They ask, "How much do you need?" The logic: why prevent someone who wants to do your marketing from using your product?
This is the opposite of traditional PLG, which optimizes conversion funnels and gates value behind paywalls. Lovable's bet: the constraint is not pricing—it's distribution. Get more people to try it. The word-of-mouth loop only activates if you blow their socks off. You can't blow socks off behind a paywall.
HubSpot ran a similar play in 2015 with the free CRM. At the time, the bridge from free to paid was a chasm. Chris Miller's team turned that chasm into a staircase by layering in a self-serve paid plan. The metaphor is literal: moving from $0 to five figures is a jump most people won't make. Introduce a mid-tier step, and the funnel flows.
Laura Schaffer describes the Plus plan launch at Amplitude as a "symphony" of pricing and packaging. The challenge: people are terrible at predicting what they'll pay for, and opinions proliferate. Her solution: hypothesis and counter-hypothesis frameworks. Where teams disagree, codify the disagreement. Ship the primary hypothesis. Track the top 10 counter-hypotheses in soft launch. This turns philosophical debates into empirical questions.
Where we disagree, instead of sitting in meetings over and over trying to get everyone to align, just recognize we're not gonna agree. What's the hypothesis we're gonna go with? What's the strongest counter-hypothesis? Let's pay attention to that.
— Laura Schaffer, VP of Growth at Amplitude
Memory Imports, Multimodal Agents, and the Clever Moves No One Notices
One of Anthropic's cleverest growth tactics: letting users import memory from ChatGPT. Activation in AI is brutal. People try a chatbot, get mediocre results, churn. By importing context from a competitor, Anthropic lowered the activation energy and increased the likelihood of an aha moment. The user didn't have to start from scratch. They started from their existing workflow.
This is a microcosm of a larger shift. Growth in 2026 is less about acquisition funnels and more about eliminating friction at every handoff. At HubSpot, Kieran Flanagan envisions a multimodal agent that onboards you, upsells you, supports you—across text, audio, screen sharing. The agent sees your screen, guides you, stays with you. The traditional org chart—growth owns activation, sales owns expansion, support owns tickets—becomes incoherent. The agent owns the relationship.
At Wiz, product-market fit was so strong that deployment happened in minutes, not months. Customers connected their cloud environment and immediately saw value. The wedge wasn't a free tier or a viral loop. It was solving a massive, urgent problem (cloud security) with a product that respected the user's time. Raz Herzberg describes a "before Wiz and after Wiz moment" for security teams. That clarity of value is what allowed founder-led sales to close millions without a sales team.
What Comes Next: Conviction, Not Consensus
The operators interviewed for this essay do not agree on whether growth teams will exist in three years. They do not agree on whether experimentation culture has gone too far or not far enough. They do not agree on whether PLG is dead, reborn, or just rebranded. What they share is a willingness to operate from conviction, not consensus.
Amol Avasare uses log-linear charts because exponential growth makes linear scales useless. Elena Verna spends 95% of her time innovating because optimization is commoditized. Bengali Kaba insists on understanding before identifying because most roadmaps are solutions in search of problems. Naomi Gleit demands canonical docs because clarity compounds. Laura Schaffer runs rogue experiments on Friday nights because sometimes forgiveness beats permission.
The through-line is not a tactic. It's a posture. The fastest-growing companies in the U.S. are not executing playbooks. They are inventing them, testing them, discarding them, and moving on before the rest of the market catches up. They treat speed as a feature, conviction as a moat, and disagreement as a signal that the terrain is still worth exploring.
Growth in 2026 is not a discipline converging on best practices. It is a discipline fracturing into schools of thought, each optimizing for a different future. The leaders who win are the ones who pick a future, build toward it, and don't wait for permission to be right.