PJFP.com

Pursuit of Joy, Fulfillment, and Purpose

Tag: anthropic ARR

  • Krishna Rao on Anthropic Going From 9 Billion to 30 Billion ARR in One Quarter and the Compute Strategy Powering Claude

    Krishna Rao, Chief Financial Officer of Anthropic, sat down with Patrick O’Shaughnessy on Invest Like the Best for one of the most detailed public looks yet at the operating engine behind Claude. He covers how Anthropic compounded from $9 billion of run rate revenue at the start of the year to north of $30 billion by the end of Q1, why he spends 30 to 40 percent of his time on compute, the playbook for buying gigawatts of AI infrastructure across Trainium, TPU, and GPU platforms, how Anthropic prices its models, why returns to frontier intelligence keep climbing, and what the Mythos release tells us about the cyber capabilities of the next generation of Claude.

    TLDW

    Anthropic is running the most compute fungible frontier lab in the world, with active deployments across AWS Trainium, Google TPU, and Nvidia GPU, and an internal orchestration layer that lets a chip serve inference in the morning and run reinforcement learning the same evening. Krishna Rao explains the cone of uncertainty that governs gigawatt scale compute procurement, the floor Anthropic refuses to drop below on model development compute, the Jevons paradox unlock from cutting Opus pricing, the 500 percent annualized net dollar retention from enterprise customers, the layer cake of long term deals with Google, Broadcom, Amazon, and the recent xAI Colossus tie up in Memphis, the phased release of the Mythos model in response to spiking cyber capabilities, the internal use of Claude Code to produce statutory financial statements and run a Monthly Financial Review skill, and why the team believes scaling laws are alive and well. The interview also covers fundraising history through Series D and Series E, the $75 billion already raised plus another $50 billion coming, talent density beating talent mass during the Meta poaching wave, and Rao’s belief that biotech and drug discovery represent the most exciting frontier for AI.

    Key Takeaways

    • Anthropic entered the year with about $9 billion of run rate revenue and ended the first quarter with north of $30 billion of run rate revenue, a more than 3x leap driven by model intelligence gains and the products built around them.
    • Compute is described as the lifeblood of the company, the canvas everything else is built on, and the most consequential class of decisions Rao makes. Buy too much and you go bankrupt. Buy too little and you cannot serve customers or stay at the frontier.
    • Rao spends 30 to 40 percent of his time on compute, even today, and the leadership team meets repeatedly on both procurement and ongoing compute allocation.
    • Anthropic is the only frontier language lab actively using all three major chip platforms in production: AWS Trainium, Google TPU, and Nvidia GPU. It is also the only major model available on all three clouds.
    • Flexibility is the central design principle. Anthropic builds flexibility into the deals themselves, into the orchestration layer that maps workloads to chips, and into compilers built from the chip level up.
    • The cone of uncertainty frames procurement. Small differences in weekly or monthly growth compound into wildly different two year outcomes, so the team plans across a range of scenarios rather than a single point estimate, and ranges toward the upper end while protecting downside.
    • Compute allocation across the company sits in three buckets: model development and research, internal employee acceleration, and external customer serving. A non negotiable floor protects model development even when customer demand is tight.
    • Anthropic estimates that if it cut off internal employee use of its own models, the freed compute could serve billions of dollars of additional revenue. It chooses not to, because internal use compounds into better future models.
    • Intelligence is multi dimensional, not a single IQ score. Anthropic measures real world capability through customer feedback, long horizon task performance, tool use, computer use, and speed at agentic tasks, not just leaderboard benchmarks that have largely saturated.
    • Each Opus generation, 4 to 4.5 to 4.6 to 4.7, delivers both capability improvements and an efficiency multiplier on token processing. New models often serve customers at a fraction of the prior cost while doing more.
    • Reinforcement learning is described as inference inside a sandbox with a reward function, so model efficiency gains directly improve internal RL throughput. The flywheel is tightly coupled.
    • Over 90 percent of code at Anthropic is now written by Claude Code, and a large share of Claude Code itself is written by Claude Code.
    • Anthropic shipped roughly 30 distinct product and feature releases in January and the pace has accelerated since.
    • Scaling laws, in Anthropic’s internal data, are alive and well. The team holds itself to a skeptical scientific standard and still does not see them slowing down.
    • Anthropic recently signed a 5 gigawatt deal with Google and Broadcom for TPUs starting in 2027, plus an Amazon Trainium agreement for up to 5 gigawatts, totaling more than $100 billion in commitments. A significant portion lands this year and next year.
    • A new partnership for capacity at the xAI Colossus facility in Memphis was announced just before the interview, aimed at expanding consumer and prosumer capacity.
    • Pricing has been remarkably stable across Haiku, Sonnet, and Opus. The biggest deliberate change was lowering Opus pricing, which produced a textbook Jevons paradox: consumption rose far faster than the price drop, and the new Opus 4.6 and 4.7 slot in at the same price point.
    • Mythos is the first model Anthropic chose to release in a phased way because of a sharp spike in cyber capability. In an open source codebase where a prior model found 22 security vulnerabilities, Mythos found roughly 250.
    • The Mythos release framework focuses on defensive use first, expands access over time, and is presented as a template for future capability spikes.
    • Anthropic now sells to 9 of the Fortune 10 and reports net dollar retention above 500 percent on an annualized basis. These are not pilots. Rao describes signing two double digit million dollar commitments during a 20 minute Uber ride to the studio.
    • The platform strategy is mostly horizontal. Anthropic will go vertical with offerings like Claude for Financial Services, Claude for Life Sciences, and Claude Security where it can demonstrate the model’s capabilities, but expects most application value to accrue to customers building on top.
    • Investors raised over $75 billion in equity since Rao joined, with another $50 billion in commitments tied to the Amazon and Google deals. Capital intensity is real, but the raises fund the upper end of the cone of uncertainty more than they fund current losses.
    • The Series E close coincided with the day the DeepSeek news broke, forcing investors to reassess their AI thesis in real time. Anthropic closed the round anyway.
    • Inside finance, Claude now produces statutory financial statements for every Anthropic legal entity, with a human checker. A library of more than 70 finance specific skills underpins workflows.
    • A custom Monthly Financial Review skill produces a 90 to 95 percent ready monthly close report, so leadership discussion shifts from reconciling numbers to debating implications.
    • An internal real time analytics platform called Anthrop Stats compresses weekly insight cycles from hours to about 30 minutes.
    • The biggest token user inside Anthropic’s finance team is the head of tax, focused on tax policy engines and workflow automation. The most senior people, not the youngest, are leading internal adoption.
    • Talent density beats talent mass. When Meta and others ran aggressive offer waves, Anthropic lost two people while peer labs lost dozens.
    • All seven Anthropic co founders remain at the company, as does most of the first 20 to 30 employees, which Rao credits to a collaborative, transparent, debate friendly culture and a real culture interview that can veto otherwise top tier candidates.
    • Dario Amodei holds an open all hands every two weeks, writes a short prepared document, and takes unscripted questions from anyone at the company.
    • AI safety investments in interpretability and alignment have a commercial side effect. Looking inside the model helps Anthropic build better models, and enterprises selling sensitive workloads want to trust the lab they hand customer data to.
    • Anthropic explicitly identifies as America first in its approach to model development, and engages closely with the US administration on capability releases such as Mythos.
    • The longer term product vision is the virtual collaborator: an agent with organizational context, access to the company’s tools, persistent memory, and the ability to work on ideas, not just tasks, over long horizons.
    • CoWork, Anthropic’s extension of the Claude Code paradigm into general knowledge work, is being adopted faster than Claude Code itself when indexed to the same point in its launch curve.
    • Anthropic’s product teams ship daily, with a fleet of agents working across the company on specific tasks. Everyone effectively becomes a manager of agents.
    • The dominant downside risks to Anthropic’s high end forecast are slower customer diffusion of model capability into real workflows, scaling laws flattening unexpectedly, and Anthropic losing its position at the frontier.
    • Rao is most excited about biotech and healthcare outcomes, especially the prospect that AI could push drug discovery and lab throughput up 10x or 100x, turning currently incurable diagnoses into treatable ones within a patient’s lifetime.

    Detailed Summary

    Compute as Lifeblood and the Cone of Uncertainty

    Rao opens with the claim that compute is the most important resource at Anthropic, and the most consequential decision class in the company. You cannot buy a gigawatt of compute next week. You have to anticipate demand a year or two in advance, and the cost of being wrong in either direction is high. Buy too much and the unit economics collapse. Buy too little and you cannot serve customers or stay at the frontier, which are described as the same failure mode. To navigate this, the team uses a cone of uncertainty rather than point estimates. Small differences in weekly growth compound into vastly different two year outcomes, and Anthropic tries to position itself toward the upper end of that cone while preserving optionality. Rao notes he has had to consciously break a lifetime of linear thinking and force himself into exponential models.

    Three Chip Platforms, One Orchestration Layer

    Anthropic uses Amazon’s Trainium, Google’s TPUs, and Nvidia’s GPUs fungibly. That was not free. Adopting TPUs at scale started around the third TPU generation, when outside observers thought it was a strange choice. Anthropic invested years into compilers and orchestration so workloads can flow across chips by generation and by job type. The team works deeply with Annapurna Labs at AWS to influence Trainium roadmaps because Anthropic stresses these chips harder than almost anyone. The result is what Rao believes is the most efficient utilization of compute across any frontier lab, with a dollar of compute going further inside Anthropic than anywhere else.

    Three Buckets and the Model Development Floor

    Compute gets allocated across model development, internal acceleration of employees, and customer serving. The conversations are collaborative rather than zero sum, but there is a hard floor on model development that the company refuses to cross even if it makes customer demand harder to serve in the short term. The thesis is simple. The returns to frontier intelligence are extremely high, especially in enterprise, so cutting model investment to chase near term revenue is a bad trade. Internal employee use is also explicitly protected. Rao notes that diverting that internal usage to external customers would unlock billions of additional revenue today, but the compounding benefit of accelerating researchers and engineers outweighs that.

    Intelligence Is Multi Dimensional

    Rao pushes back hard on the IQ framing of model progress. Benchmarks saturate quickly, and the real signal comes from how customers actually use the models. Anthropic looks at long horizon task completion, tool use, computer use, and time to result on agentic tasks. Two equally capable agents who differ only in speed produce dramatically different value, because the faster one compounds into more attempts and more outcomes. Frontier model leaps are also fuel efficient. The sedan to sports car analogy breaks down because each Opus generation, 4 to 4.5 to 4.6 to 4.7, delivers a step up in capability and a multiplier on per token efficiency.

    From 9 Billion to 30 Billion ARR in One Quarter

    The headline number for the quarter is a leap from about $9 billion of run rate revenue to over $30 billion, accomplished without onboarding a corresponding step up in compute, because new compute lands on ramps locked in 12 months prior. Rao attributes the leap to model capability gains, products that surface that intelligence in usable form factors, and an enterprise customer base that pulls more workloads onto Claude as each generation unlocks new use cases. Coding started the wave with Sonnet 3.5 and 3.6, and the same pattern is now playing out elsewhere in the economy.

    Recursive Self Improvement and Talent Density

    Over 90 percent of Anthropic’s code is now written by Claude Code, including most of Claude Code itself. Rao describes this as a structural reason to keep allocating internal compute to employees even when external demand is hungry. Recursive self improvement is not happening through models that need no humans. It is happening through researchers who set direction and use frontier models to compress months of work into days. Talent density beats talent mass. When Meta and other labs went after Anthropic researchers with very large packages, Anthropic lost two people while peer labs lost dozens.

    Procurement Strategy and the Layer Cake

    Compute lands as a layer cake. Last month Anthropic signed a 5 gigawatt TPU deal with Google and Broadcom starting in 2027, alongside an Amazon Trainium agreement for up to 5 gigawatts. The total is north of $100 billion in commitments. A new tie up with xAI’s Colossus facility in Memphis was announced just before the interview, intended for nearer term capacity to support consumer and prosumer growth. Anthropic evaluates near term and long term compute deals against the same set of variables: price, duration, location, chip type, and how efficiently the team can run it. The relationships are deeper than procurement. The hyperscalers are also distribution channels for the model.

    Platform First, Selective Vertical Bets

    Rao describes Anthropic as a platform first business, with most expected value accruing to customers building on the platform. The team will only go vertical when it can either demonstrate capabilities that are skating to where the puck is going, like Claude Code did before the models could fully support it, or when it wants to set a template for an industry vertical, as with Claude for Financial Services, Claude for Life Sciences, and Claude Security. He acknowledges that surprise capability jumps make customers anxious about the platform competing with them, and frames Anthropic’s mitigation as deeper partnerships, early access programs, and an emphasis on accelerating customer building rather than disintermediating it.

    Pricing, Jevons Paradox, and Return on Compute

    Pricing across Haiku, Sonnet, and Opus has been stable. The notable exception is Opus, which Anthropic deliberately repriced lower when launching Opus 4.5 because Opus class problems were being squeezed into Sonnet workloads. Efficiency gains made it possible to serve Opus profitably at the new level. The consumption response was a classic Jevons paradox, with usage rising far more than the price reduction would have predicted, and Opus 4.6 then slotted in at the same price with a capability bump. Margins are not framed as a per token markup. Compute is fungible across model development, internal acceleration, and customer serving, so Anthropic measures return on the entire compute envelope rather than software style variable cost per call.

    Fundraising, DeepSeek, and Capital Intensity

    Rao joined while Anthropic was closing its Series D, mid frontier model launch and during the FTX share liquidation. Investors initially questioned whether Anthropic needed a frontier model, whether AI safety and a real business could coexist, and why the sales team was so small. The Series E closed the same day the DeepSeek news broke, with markets violently re pricing AI in real time. Since Rao joined, Anthropic has raised over $75 billion, with another $50 billion tied to the Amazon and Google compute deals. The reason for the size of the raises is the cone of uncertainty, not current losses. Returns on compute today are described as robust.

    Mythos, Cyber Capability, and Phased Releases

    The Mythos release marks the first time Anthropic shipped a model under a deliberately phased rollout because of a specific capability spike. Cyber is the dimension that spiked. Where a prior model found 22 vulnerabilities in an open source codebase, Mythos found roughly 250. The defensive applications, automatically patching massive codebases, are genuinely valuable, but the offensive risk is real enough that Anthropic chose to release to a smaller group first and expand access over time. Rao positions this as a template for future capability spikes, not a permanent restriction. He also describes the relationship with the US administration as cooperative, including the Department of War interaction, with Anthropic supporting a regulatory framework that does not strangle innovation but takes responsibility seriously.

    Claude Inside Finance

    Anthropic’s finance team is one of the strongest internal case studies. Statutory financial statements for every legal entity are produced by Claude, with a human reviewer. A skill library of more than 70 finance specific skills underpins a Monthly Financial Review skill that drafts the monthly close at 90 to 95 percent ready, so leadership meetings shift from explaining the numbers to discussing what to do about them. An internal analytics platform called Anthrop Stats compresses weekly insight cycles from hours to 30 minutes. The biggest internal token user in finance is the head of tax, building policy engines, which Rao highlights as evidence that adoption is driven by the most senior people, not just younger engineers.

    Culture, Co Founders, and the Race to the Top

    Seven co founders should not, on paper, work as a leadership group. Rao argues it works because the culture was set early around collaboration, intellectual honesty, transparency, and humility. The culture interview is a real veto, not a checkbox. Dario Amodei runs an all hands every two weeks with a short written piece followed by unscripted questions, and decisions, once made, get clean alignment rather than residual politics. Anthropic frames its approach as a race to the top, where being a model for how to build the technology responsibly is itself a recruiting and retention advantage.

    The Virtual Collaborator and the Frontier Ahead

    The product vision Rao describes is the virtual collaborator. Not just a smarter chatbot, but an agent with organizational context, access to the company’s tools, memory, and the ability to work on ideas over long horizons. Coding was the first domain to feel this, but CoWork, Anthropic’s extension of the Claude Code pattern into general knowledge work, is being adopted faster than Claude Code was at the same age. Product development inside Anthropic already looks different. Teams ship daily, with fleets of agents working across the company, and individual humans increasingly act as managers of those fleets.

    Downside Risks and What Excites Him Most

    The three risks Rao names if asked to do a premortem on a softer year are slower customer diffusion of model capability into real workflows, scaling laws unexpectedly flattening, and Anthropic losing its frontier position to competitors. None of these are observed today, but he is unwilling to claim them with certainty. On the upside, he is most excited about biotech and healthcare. Lab throughput rising 10x or 100x, paired with AI assisted clinical workflows, could turn currently incurable diagnoses into treatable ones within a patient’s lifetime. That is the outcome he wants the technology to chase.

    Thoughts

    The most consequential structural point in this interview is the framing of compute as a single fungible resource pool measured by return on the entire envelope, not as a variable cost per inference call. That accounting shift, if you accept it, breaks most of the bear cases about AI lab unit economics. The bear argument almost always assumes that a token served to a customer is the only thing the chip did that day. Rao’s version is that the same fleet trains models in the morning, runs reinforcement learning at lunch, serves customers in the afternoon, and accelerates internal engineers in the evening. If even half of that is real, the right comparison is total compute spend versus total enterprise value created by the platform, and on that ratio Anthropic looks structurally strong rather than weak.

    The Jevons paradox on Opus pricing is the most actionable insight for anyone running an AI product. Most teams default to either chasing premium pricing on the newest model or undercutting to chase volume. Anthropic did something more disciplined: it left Sonnet and Haiku alone, dropped Opus when efficiency gains made it serveable, and watched aggregate usage rise faster than the price cut. The lesson is that frontier model pricing is not really a price problem. It is a capability access problem, and elasticity around the right tier is much higher than the standard SaaS playbook implies.

    The Mythos cyber jump deserves more attention than it has gotten. Going from 22 to 250 vulnerabilities found in the same codebase is the kind of capability discontinuity that genuinely changes the regulatory calculus. Anthropic is signaling that it can identify these discontinuities ahead of release and choose a deployment shape that respects them. Whether peer labs adopt similar discipline is the open question. Anthropic’s race to the top framing assumes they will be forced to. The competitive market may say otherwise.

    The hiring data point is the most underrated investor signal. Two departures while peer labs lost dozens, during the most aggressive talent war in tech history, is not a culture poster. It is a structural advantage that compounds every time another lab tries to buy its way to the frontier. Money can be matched. Conviction in the mission, transparent leadership, and a culture interview that can veto otherwise stellar candidates cannot. If you believe scaling laws hold, talent retention at this density is one of the few moats that actually scales with capital.

    Finally, the most interesting personal admission is that Krishna Rao, a finance leader trained at Blackstone and Cedar, is openly telling investors that linear thinking is the failure mode he had to break out of. The companies that pattern match this moment to prior technology waves are mispricing it, in both directions. The cone of uncertainty Anthropic uses internally is the right metaphor for everyone else too. If you are forecasting AI as if it is cloud in 2010, you are almost certainly wrong, and the magnitude of the error is much larger than it would be in any prior era.

    Watch the full conversation with Krishna Rao on Invest Like the Best here.

  • Anthropic’s Growth Strategy Explained: $1B to $19B in 14 Months, Automating Experiments With Claude, and Why Old Playbooks Are Dead (Lenny’s Podcast Recap)

    TLDW (Too Long, Didn’t Watch)

    Amol Avasare, Head of Growth at Anthropic, sat down with Lenny Rachitsky to explain how Anthropic grew from $1 billion to over $19 billion in annual recurring revenue in just 14 months. He breaks down their internal tool called CASH (Claude Accelerates Sustainable Hypergrowth) that automates growth experimentation, why 50 to 70 percent of traditional growth playbooks are now obsolete, why the PM-to-engineer ratio may need to flip, and how Anthropic’s early bet on AI coding created a research flywheel that competitors are only now starting to copy. He also shares how he cold emailed his way into the job, why activation is the single hardest problem in AI products, and how he uses Cowork to detect team misalignment across Slack channels automatically.

    Key Takeaways

    1. Anthropic’s growth trajectory is historically unprecedented. Revenue went from $1 billion at the start of 2025 to over $19 billion ARR by February 2026. That 19x growth in 14 months dwarfs companies like Atlassian, Snowflake, and Palantir, which took 15 to 20 years to reach $4.5 to $6 billion ARR. The number Amol quoted was already outdated by the time the episode aired.

    2. Anthropic is automating growth experimentation with an internal tool called CASH. CASH stands for Claude Accelerates Sustainable Hypergrowth. The growth platform team uses Claude to identify opportunities, build experiments (mostly copy changes and minor UI tweaks so far), test them against quality and brand standards, and analyze results. Amol describes the current win rate as roughly equivalent to a junior PM with two to three years of experience, but notes it was not possible at all before Opus 4.5 and has improved significantly with Opus 4.6. Human review is still in the loop but decreasing week over week.

    3. Activation is the single highest-leverage growth problem in AI. The core challenge is capability overhang: models are improving so fast that users do not know what they can do. By the time you have tested and optimized onboarding for one model’s capabilities, the next model has already shipped with entirely new features that make your learnings obsolete. Anthropic addresses this by adding intentional friction in onboarding to understand who users are and funnel them to the right products and features.

    4. Anthropic indexes 70/30 toward big bets, the opposite of most growth teams. Traditional growth teams spend 60 to 70 percent of effort on small to medium optimizations. Anthropic flips that ratio because they believe the product value delivered two years from now will be 100x to 1,000x what it is today. In that exponential environment, micro-optimizations capture a negligible percentage of future value. Large strategic bets are where the leverage lives.

    5. The PM-to-engineer ratio may need to flip. Engineers are getting 2 to 3x more productive with tools like Claude Code, effectively turning a team of 5 engineers into the equivalent of 15 to 20. But PMs and designers have not seen the same multiplier. The result is that product management and design are “absolutely squeezed.” Anthropic is responding by hiring more PMs and deputizing product-minded engineers to act as mini-PMs on projects under two weeks. The counterintuitive insight: companies may need more PMs, not fewer, as AI accelerates engineering output.

    6. Cold emailing still works if you do it right. Amol got his job by cold emailing Mike Krieger, Anthropic’s Chief Product Officer (and co-founder of Instagram), at a time when no growth role was even listed. Key tactics: use a high-converting subject line you have tested over time, find personal email addresses instead of competing in crowded LinkedIn inboxes, keep the message extremely short, and follow up relentlessly until someone explicitly asks you to stop.

    7. PRDs are largely obsolete at Anthropic. Amol estimates that 60 to 80 percent of what his team ships does not have a formal PRD. For small projects, coordination happens entirely in Slack. For larger initiatives, he will sometimes throw his thoughts into Cowork five minutes before a kickoff meeting to generate a rough document. His default philosophy: if you can skip the doc and jump straight to prototyping or action, do it.

    8. The AI coding bet created a research flywheel. Anthropic’s deep focus on coding was not just a commercial play. A document written by co-founder Ben Mann in 2021, just months after the company was founded, laid out the case for focusing on AI coding because better coding models would accelerate their own researchers, which would produce better models, which would produce better coding tools, in a compounding loop. This is something competitors are only now starting to recognize and copy.

    9. Cowork is being used to detect organizational misalignment. Amol runs a weekly scheduled task in Cowork that uses the Slack MCP to scan conversations across the company and surface areas of potential misalignment. He describes cases where this caught teams about to do overlapping work or spin their wheels on conflicting priorities. He also uses Cowork to simulate coaching sessions with his manager, Ami Vora, by asking Claude to analyze her public writing and internal Slack activity and then deliver feedback from her perspective.

    10. Anthropic’s culture is its most defensible moat. Amol describes a culture where every single person is fully engaged, nobody is checked out, and there is radical transparency through “notebook channels” on Slack where anyone, including leadership, shares their thinking publicly. Employees openly challenge Dario Amodei in these channels after all-hands meetings. These notebook channels also serve a practical purpose: they become training data that helps Claude understand how different teams think and operate.

    Detailed Summary

    The Cold Email That Started It All

    Amol Avasare was not recruited through a job listing, a referral, or a sourcing pipeline. He cold emailed Mike Krieger, Anthropic’s CPO and the co-founder of Instagram, with a short pitch: he loved the product, thought Anthropic badly needed a growth team, and wanted to talk. At the time, Anthropic had no growth roles posted. They were just beginning to think about it internally, and the timing was perfect.

    Amol’s approach to cold email is methodical. He has a subject line formula he has refined over years of founder outreach that produces abnormally high open rates (he declined to share the exact copy). He targets personal email addresses rather than work inboxes or LinkedIn, where competition for attention is fierce. The message itself is brutally short: who he is, why he would be a fit, and a request to chat. His follow-up philosophy is to keep reaching out until someone tells him to stop. Krieger responded on the first attempt.

    What $1B to $19B in 14 Months Actually Feels Like

    From the inside, Anthropic’s growth does not feel like a victory lap. Amol describes it as the hardest job he has ever had, harder than being a founder and harder than investment banking. About 70 percent of his time goes to what the team calls “success disasters,” which are problems created by things going extremely well. All the charts are green and up and to the right, but the underlying infrastructure, processes, and systems are constantly breaking under the strain of hypergrowth.

    The revenue trajectory tells the story: $0 to $100 million in 2023, $100 million to $1 billion in 2024, $1 billion to roughly $10 billion in 2025, and already $19 billion ARR by the end of February 2026. Amol notes that at the end of 2024, Dario Amodei was pushing for growth targets that the team thought were impossible. Those targets were hit and exceeded. The internal culture has adapted accordingly. Linear charts are considered uncool. Everything is presented on a log-linear scale.

    Why Activation Is the Hardest Problem in AI

    The central growth challenge for AI products is not acquisition. It is activation: getting users to understand what the product can actually do for them. Amol frames this as a capability overhang problem. Models are improving so rapidly that even internal teams struggle to keep up with what is newly possible. If Anthropic employees have to carve out dedicated time to explore a new model’s capabilities, the average user is even further behind.

    The danger is that someone signs up for Claude, asks it about the weather, and walks away thinking that is all it does. The product development cycle for onboarding is also under strain: by the time you have run tests, gathered learnings, and shipped an optimized activation flow for one model generation, the next model has shipped with capabilities that make your work irrelevant.

    Anthropic’s approach borrows from Amol’s experience at Mercury and MasterClass. They add deliberate friction to the signup flow, asking users questions about who they are and what they want to accomplish. This allows them to route users to the right products and features. The data also feeds downstream into lifecycle marketing and ad targeting. Amol has seen this pattern work consistently across every company he has worked at: the right friction, applied at the right time, outperforms frictionless flows that dump users into a blank canvas with no guidance.

    The CASH System: Automating Growth Experimentation

    Anthropic’s growth platform team, led by Alexey Komissarouk (who teaches growth engineering at Reforge), has built an internal system called CASH. The name stands for Claude Accelerates Sustainable Hypergrowth.

    CASH operates on a four-stage loop. First, Claude identifies growth opportunities by analyzing trends, metrics, and past experiment results. Second, Claude builds the actual feature or change. Third, Claude tests the output against quality and brand standards. Fourth, Claude analyzes the results and gathers learnings after the experiment ships.

    Currently, CASH handles mostly copy changes and minor UI tweaks. The win rate is comparable to a junior PM with two to three years of experience. A senior PM would still do better. But the trajectory matters: this was not possible at all before Opus 4.5 launched, and results have improved meaningfully with Opus 4.6. Human approval is still required before shipping, but the amount of human time spent reviewing is decreasing week over week.

    The part that Claude still cannot handle well is cross-functional stakeholder management. Getting six people in a room to align on a decision remains a fundamentally human problem. As Amol’s head of design joked: “We will have AGI and it will still be impossible to get six people in a room to get aligned.”

    Why the PM-to-Engineer Ratio Might Flip

    This is one of the most counterintuitive insights from the conversation. The conventional assumption is that AI will reduce the need for PMs. Amol argues the opposite: companies may need more PMs, at least in the near term.

    The math is straightforward. Tools like Claude Code are making engineers 2 to 3x more productive. A team of 5 engineers now produces the output equivalent of 15 to 20 engineers in the pre-AI era. PMs and designers have seen productivity gains, but not at the same multiplier. The result is a bottleneck: one PM managing the equivalent output of 15 to 20 engineers worth of work, while also handling cross-functional coordination, stakeholder alignment, and strategic direction.

    Anthropic’s response is twofold. First, they are actively hiring more PMs. Second, they have formalized a system where product-minded engineers act as mini-PMs on any project that is two engineering weeks or less. The engineer handles everything: talking to legal, talking to security, managing stakeholders. The PM only steps in if things go badly off track.

    For larger projects, the PM remains squarely accountable. But the key insight is about leverage: if you are one PM managing 20 engineers, the highest-value use of your time is not shipping the 21st feature yourself. It is getting 5 percent better at guiding the team on what the right opportunities are and upleveling every engineer’s product thinking.

    The Coding Flywheel That Changed Everything

    Anthropic’s deep bet on coding was not obvious at the time. A document from co-founder Ben Mann, dated 2021, laid out the strategic logic just months after the company was founded. The argument was that investing heavily in AI coding would create a compounding flywheel: better coding models would help Anthropic’s own researchers write code more effectively, which would accelerate model development, which would produce even better coding tools.

    This early focus gave Anthropic a structural advantage that competitors are only now trying to replicate. It also explains why the company went so deep on B2B and enterprise use cases rather than chasing consumer attention. The commercial opportunity of coding was large on its own, but the internal research acceleration made it doubly strategic.

    Amol notes that this focus was partly born from constraint. Anthropic was the smallest, least well-funded player in the space for years. They did not have Meta’s distribution or Google’s cash flow or OpenAI’s first-mover advantage. That constraint forced extreme focus, which is a principle Amol applies broadly. He calls it “freedom through constraints.”

    How Amol Uses AI to Manage His Day

    Amol’s personal AI usage is extensive and worth documenting for anyone looking to see how a power user at the frontier actually operates.

    Every morning, a scheduled Cowork task reviews 20 to 25 charts across Anthropic’s products and sends him a summary of what needs attention. The false positive and false negative rates are improving week over week, giving him increasing confidence in delegating this monitoring.

    He uses Cowork to handle administrative tasks he hates: booking meeting rooms, first-pass email triage, filing expense reports in Brex and reimbursements in Benpass. None of this requires his attention anymore.

    For management, he runs weekly Cowork tasks that review what his direct reports have done, cross-reference their work against team OKRs and meeting transcripts, and surface feedback he should give them. He also runs a parallel task for himself, asking Claude to impersonate his manager Ami Vora based on her public writing and internal Slack activity, and deliver feedback from her perspective.

    Perhaps most powerfully, he runs a weekly misalignment detection task that scans Slack conversations across the company and surfaces areas where teams may be working at cross purposes. He describes cases where this caught potentially expensive coordination failures before they compounded.

    Notebook Channels and the Culture Moat

    Anthropic uses “notebook channels” on Slack, which function like internal Twitter feeds where employees share their thinking, priorities, and provocative ideas publicly. Everyone has one, from researchers to growth PMs to Dario Amodei himself. Employees openly disagree with leadership in these channels, and that is encouraged.

    These channels serve a dual purpose. First, they help scale cultural values and operating principles as the company grows rapidly. When Amol posts about “the importance of being comfortable leaving money on the table,” every new engineer on the growth team absorbs that principle. Second, and perhaps more importantly for the long term, these channels become structured context that Claude can reference. The HR team has even documented which internal documents Claude should reference for specific topics. Amol sees this as something every company will eventually need to do: share thinking in a structured way so that the AI agents running throughout the organization have the context they need.

    AI Safety as Commercial Strategy

    Anthropic is structured as a Public Benefit Corporation (PBC), not a standard Delaware C-Corp. This legally allows the company to optimize for public benefit rather than being bound solely to maximize shareholder value.

    Amol says the company has repeatedly taken significant commercial hits for safety reasons, including delaying product launches when safety risks were identified. He also makes a striking claim: what Anthropic says publicly about AI risk is actually a softer version of what they believe internally. The internal view on the potential downsides of powerful AI is more aggressive than the public messaging.

    From a growth perspective, Amol frames safety as a long-term competitive advantage. Growth teams at most companies try to squeeze every last dollar. Anthropic’s growth team is “very comfortable forgoing metric impact” to protect brand, quality, and safety. He argues this is how all the best products operate, and that as the stakes of AI get higher, Anthropic’s credible commitment to safety will become a moat.

    Advice for Thriving in the AI Era

    Amol’s advice for product managers and growth practitioners boils down to four points. First, stay on top of the tools. Try every new model release. Something that did not work three months ago may work now, and you will not know unless you go back and test it. Second, go deep on your unique spike rather than trying to be well-rounded. The PM who can also design is a unicorn. The engineer who thinks like a PM is a unicorn. Find your interdisciplinary edge and double down. Third, be radically adaptable. Amol estimates that 50 to 70 percent of how you operated in the past is now irrelevant. Clinging to old playbooks creates friction. Fourth, think in exponentials, not linear projections. If you are looking at the AI landscape through a linear lens, you will consistently underestimate how quickly things are moving.

    Thoughts

    This interview is one of the most information-dense conversations about growth strategy in AI that has been published so far. A few things stand out.

    The CASH system is the most concrete example yet of a company using AI to automate its own growth loop. The fact that it currently performs at a junior PM level is almost beside the point. What matters is the trajectory: it went from impossible to functional in a few months. If models continue improving at their current pace, this system will be operating at a senior PM level within a year. Every growth team at every AI company should be building their own version of this right now.

    The PM ratio insight is genuinely surprising and underreported. The default assumption in the tech industry is that AI will reduce headcount across all functions. Amol is making the case that in the near term, the opposite is true for PMs. Engineering output is exploding, and someone needs to direct all that output toward the right problems. That is a fundamentally human, organizational, political job that AI is not close to automating.

    The coding flywheel story is also worth highlighting because it shows the power of strategic focus in a world of unlimited possibilities. Anthropic had a generalist technology that could do almost anything, and they deliberately narrowed their focus to one vertical. That decision, made in 2021 before anyone knew what the market would look like, is arguably the single most important strategic bet in the company’s history.

    Finally, the notebook channels concept deserves more attention. The idea that employees should share their thinking in structured, searchable formats is not just a culture tool. It is an infrastructure investment for an AI-native future where agents need organizational context to be effective. Companies that build this habit early will have a significant advantage when agent-driven workflows become the norm.

    The uncomfortable subtext of this entire conversation is that Anthropic’s growth team, as talented as they clearly are, is riding a wave created almost entirely by the research team. Several YouTube commenters pointed this out, and Amol himself acknowledges it directly. The models are the product. The growth team’s job is to make sure users discover and adopt what the models can do. That is not a small job, especially at this scale, but it is a fundamentally different job than driving growth at a product that does not sell itself.