PJFP.com

Pursuit of Joy, Fulfillment, and Purpose

Tag: openrouter

  • BlackRock’s Machine-Native Economy Paper Is the Most Bullish Institutional Case Yet for Bitcoin and Crypto: AI Agents, Stablecoins, x402, and the Coming Market for Tokenized Compute

    BlackRock, the largest asset manager on the planet, just published an 11-page research paper arguing that artificial intelligence may be the most underappreciated demand driver the crypto economy has ever had. It is called The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute, and it comes from Head of Digital Assets Robert Mitchnick, Head of Digital Assets Research Will Su, U.S. Head of Equity ETFs Jay Jacobs, and Head of U.S. iShares Product Innovation William Helm. The language is careful, the way institutional research always is. The conclusion is not. Read plainly, BlackRock is saying that AI is machine-native intelligence, crypto is machine-native money, and the two were built for each other. For anyone who has been bullish on bitcoin and digital assets, this is the thesis, now written on BlackRock letterhead.

    TLDR

    BlackRock argues that AI and digital assets are converging into a single machine-native economy and that broad AI adoption is an underappreciated source of demand for crypto. The paper makes three cases. First, LLMs and blockchains share an analogous tokenization architecture, turning language and value into standardized units machines can process natively, which gives AI agents a more direct interface with on-chain assets than with fragmented legacy systems. Second, agentic commerce needs machine-native payment rails: card networks and ACH carry human onboarding, merchant fees, and settlement delays that make always-on sub-cent machine payments uneconomic, while stablecoins and protocols like Coinbase’s x402 settle around the clock with no human in the loop. Adjusted stablecoin volume topped $11 trillion in 2025, in the same range as Visa and Mastercard, growing at an 80% CAGR versus roughly 8.5% for ACH. Third, compute is becoming a trillion-dollar commodity (hyperscaler cloud revenue is projected near $1.1 trillion by 2030), and standardized, tokenized claims on compute could become a major new digital asset market. The paper also cites Bitcoin Policy Institute research in which AI models favored stablecoins for everyday payments and bitcoin for long-term value preservation, sketching an AI-native monetary architecture with bitcoin as the reserve asset.

    Thoughts

    Start with the most important sentence in the paper, the one tucked into the tokenization section on page four. Citing the Bitcoin Policy Institute, BlackRock describes “a potential AI-native monetary architecture in which stablecoins serve as transaction money and bitcoin as a store of value.” Think about what that means. When you ask the machines themselves which money they would choose, they reach for dollars on-chain to spend and bitcoin to save. That is exactly the division of labor bitcoiners have described for a decade: stablecoins as the checking account, bitcoin as the treasury. Every stablecoin transaction an agent makes is denominated in a currency its issuer can inflate. An agent optimizing for long-horizon value, with no nostalgia, no home bias, and no attachment to any particular central bank, has one obvious answer for savings: a fixed supply of 21 million, final settlement, no counterparty, and no one who can change the rules. BlackRock rightly notes these are simulated model responses, not observed behavior. But simulations are where agent behavior starts, and the direction is not ambiguous.

    The payment rails argument is where the bull case becomes mechanical rather than philosophical. BlackRock lists the problems with legacy rails bluntly: account setup that needs a human, merchant fees that make tiny transactions uneconomic, settlement that takes a business day or longer, and scalability limits as machine volume grows. An AI agent cannot walk into a bank branch. It cannot pass a credit check. It will want to pay a fraction of a cent for an API call, thousands of times an hour, at 3 a.m. on a Sunday. The only rails that do this natively are blockchains. x402, which revives the long-dormant HTTP 402 “Payment Required” status code, lets an agent pay for a resource inside the web request itself. It is worth remembering that the bitcoin world got here first: Lightning Labs’ L402 protocol paired the same 402 status code with Lightning payments years ago. The idea that the internet finally gets a native payment layer, and that the layer is crypto, is no longer a cypherpunk dream. It is in a BlackRock paper, next to Stripe, Visa, Google, and OpenAI protocols.

    Then look at the numbers in the middle of the paper, because they are staggering. Adjusted stablecoin transaction volume exceeded $11 trillion in 2025, putting it in the same broad range as Visa and Mastercard. Circulating stablecoin supply is north of $300 billion. From 2020 to 2025, adjusted stablecoin volume compounded at 80% a year, against roughly 8.5% for ACH. And this happened before agentic commerce showed up in any meaningful volume. Humans alone, with clunky wallets and regulatory fog, built a payment network that rivals the card giants in five years. Now add the GENIUS Act in the U.S., MiCA in Europe, and licensing regimes in Hong Kong and Singapore. Then add millions, eventually billions, of autonomous agents that pay far more often than any human ever will. BlackRock also makes the second-order point clearly: settlement on permissionless networks drives demand for blockspace and validator services, a direct transmission channel to native cryptoassets. The stablecoin boom is not a rival to crypto. It is fuel for the chains underneath it.

    The compute section in the back half is the part most coverage will skip, and it is the most important long-term idea in the paper. BlackRock argues that compute is becoming a distinct, investable commodity: cumulative AI capex above $5 trillion through 2030, hyperscaler cloud revenue near $1.1 trillion by 2030, and inference on track to be the largest AI workload. Commodities get financial markets, and BlackRock expects standardized, tokenized compute contracts that can be “represented, transferred, pledged as collateral, and settled through programmable infrastructure.” Here is the bitcoin angle the paper leaves implicit: AI and bitcoin run on the same scarce input, energy. AI turns electricity into intelligence. Bitcoin proof of work turns electricity into the hardest money ever created. Bitcoin miners are already among the best-positioned owners of powered land and grid interconnects on earth, and many have been signing AI hosting deals. The machine economy’s two most important commodities, compute and sound money, are both energy commodities, and bitcoin is the only monetary asset whose issuance is anchored in physical energy expenditure.

    Finally, weigh the conclusion and the signal in the byline. BlackRock ends by saying digital assets “could become increasingly integral to AI’s economic infrastructure,” across stablecoins, tokenized real-world assets, and “native cryptoassets that support blockchain settlement.” It points to Stripe’s August 2026 agreement to acquire OpenRouter as evidence that compute procurement, usage billing, and programmable settlement are merging. And two of the four authors run BlackRock’s ETF and iShares product businesses. This is not an academic exercise. It is the firm that runs the iShares Bitcoin Trust telling its clients the demand story for crypto is about to get a second engine. The first engine was institutions discovering bitcoin as a macro asset. The second is the machines. The paper admits that agentic payments are nascent and that compute-market liquidity is thin. That is the bullish part. You do not get a paper like this when the trade is crowded. You get it when the smartest money in the room can see the curve but the market has not priced it yet.

    Key Takeaways

    • BlackRock calls AI “the defining technology theme of this era” and digital assets a concurrent theme with major implications for financial infrastructure, and says the two are now converging.
    • The paper’s central framing: AI is machine-native intelligence and digital assets are machine-native money.
    • BlackRock says broad AI adoption “may represent an underappreciated source of demand, utility, and application growth across the digital asset economy.”
    • Agentic AI, systems that plan and execute multistep tasks with limited human intervention, pushes AI from generating content to taking real-world action, including purchases and financial transactions.
    • LLMs split text into tokens, map them to numeric IDs, and embed them as vectors. Blockchains represent value and ownership claims as standardized tokens recorded on a ledger. Different functions, same idea: convert real-world inputs into machine-native formats.
    • Because both systems use structured, machine-readable data, AI agents can interface with blockchain data more directly than with fragmented legacy databases.
    • Tokenizing more asset classes reduces bespoke integrations and lets agents orchestrate complex multi-asset workflows: checking balances and rules, executing authorized transactions, and verifying settlement.
    • Compliance (AML, KYC, and the new “know-your-agent” or KYA checks) generally happens off-chain, with verified results passed on-chain to determine eligibility.
    • Bitcoin Policy Institute research found AI models in controlled simulations generally favored stablecoins for everyday payments and bitcoin for long-term value preservation.
    • BlackRock frames that result as a potential AI-native monetary architecture: stablecoins as transaction money, bitcoin as the store of value.
    • Crypto rails are “particularly well suited” to high-frequency, sub-cent, around-the-clock machine-to-machine transactions like API calls, on-demand data, and consumption-based compute.
    • Legacy rails struggle with agents because of human-dependent onboarding, merchant fees that kill micropayments, slow settlement and dispute finality, and scaling limits.
    • Modified traditional rails will still matter for business-to-machine and consumer-to-machine commerce, where agents deal with human-run businesses.
    • Agentic payments sit on foundational standards: Anthropic’s Model Context Protocol (MCP, November 2024) for tool and data access, and Google’s Agent2Agent (A2A, April 2025) for agent coordination.
    • Coinbase’s x402 uses the HTTP 402 “Payment Required” status code to let machines pay inside web requests. It is blockchain-agnostic, with USDC as an early primary use case.
    • x402 offers 24/7, near-real-time, verifiable settlement, which reduces counterparty exposure for providers and lets them release data or services the moment payment confirms.
    • More x402 usage on permissionless networks could increase demand for blockspace and validator services, a transmission channel to native cryptoassets.
    • Other protocols in the stack include Stripe and Tempo’s Machine Payments Protocol (MPP), Stripe and OpenAI’s Agentic Commerce Protocol (ACP), Google’s AP2 with cryptographic mandates, and Visa’s Trusted Agent Protocol (TAP).
    • BlackRock’s example workflow: a user asks an agent to book a trip under $2,500, the agent delegates to a travel sub-agent via A2A, the sub-agent pays for fare data via x402 settled on-chain, and the primary agent books through ACP.
    • Stablecoins are likely to lead transactional use because price stability gives agents a reliable unit of account.
    • Stablecoins are the largest category of tokenized real-world assets, with more than $300 billion in circulation as of September 2026.
    • Adjusted stablecoin transaction volume exceeded $11 trillion in 2025, in the same broad range as Visa and Mastercard’s annual payment volumes.
    • Stablecoin volume grew at an 80% CAGR from 2020 to 2025, versus about 8.5% for ACH, which still moved $93 trillion in 2025.
    • Regulatory clarity, including the GENIUS Act, MiCA, Hong Kong’s licensing regime, and Singapore’s framework, should support continued stablecoin growth.
    • Stablecoin growth spills over to the chains that settle them. On networks like Ethereum, native assets such as ETH pay for consensus, validators, and fees, so more activity can mean more value capture.
    • Purpose-built stablecoin chains like Circle’s Arc, where USDC is the native gas asset, offer a complementary model.
    • Some estimates put cumulative AI capital spending above $5 trillion between 2025 and 2030, and BlackRock says ongoing operating spend deserves equal attention.
    • Consensus estimates for AWS, Microsoft Intelligent Cloud, and Google Cloud imply about $1.1 trillion in combined revenue by 2030, a 29% CAGR from 2025.
    • Compute is becoming a distinct, large, investable economic resource that could support a new class of digital assets.
    • Inference is expected to be the largest AI workload by 2030, and its user base is far larger and more fragmented than the concentrated training market.
    • Challenges remain, including chip-generation differences, regional energy costs, and settlement standards, but BlackRock calls them “important but ultimately resolvable.”
    • BlackRock expects standardized products, including exchange-traded compute futures, and tokenized compute claims that can be transferred, pledged as collateral, and settled on programmable rails.
    • Agents could shop real-time compute marketplaces on price, latency, location, and hardware, then pay per use, per model token, or per job via x402.
    • Stripe’s August 2026 agreement to acquire OpenRouter, which routes workloads across more than 400 models from over 80 providers, signals that compute procurement, billing, and programmable settlement are converging.
    • BlackRock’s conclusion: as agents grow more capable, digital assets could become integral to AI’s economic infrastructure across stablecoins, tokenized RWAs, and native cryptoassets.

    Detailed Summary

    Two Technology Waves Become One

    BlackRock opens by naming AI as the defining technology of the era and digital assets as a parallel wave with deep implications for financial infrastructure. For years the two ran on separate tracks. The paper argues that they are now merging because AI is gaining the ability to act on economic networks, not just talk about them. Agentic AI plans and executes multistep tasks, calls external tools, and increasingly makes purchases and initiates financial transactions. Once software can spend money, the question of which money it spends and which rails it uses becomes central, and BlackRock’s answer is that blockchains provide the programmable infrastructure that connects intelligence to economic activity.

    Tokens All the Way Down

    The first pillar is architectural. An LLM tokenizes text into words or sub-words, maps them to numeric IDs, and converts them to embeddings the model can compute on in parallel. A blockchain tokenizes value: cash, a money market fund interest, a security, or another claim becomes a standardized token recorded on a distributed ledger. Transactions are machine-readable data governed by rules. The network verifies authorization, smart contracts apply asset-specific conditions, and once finalized the transfer becomes part of the canonical ledger. BlackRock’s Figure 1 sets these two pipelines side by side, “AI is changing the world” becoming vectors on one side and a $100 money market fund interest becoming a finalized on-chain record on the other. Because both speak structured, machine-readable formats, agents can plug into on-chain assets more directly than into siloed legacy systems, and broader tokenization across asset classes reduces the bespoke integration work that slows automation today.

    Bitcoin as the AI-Native Store of Value

    The paper cites the Bitcoin Policy Institute study “Which money do AI agents prefer?”, in which model outputs across controlled simulations generally chose stablecoins for everyday payments and bitcoin for long-term value preservation. BlackRock is careful to say these are simulated responses rather than observed agent behavior, but it takes the result seriously enough to describe a potential AI-native monetary architecture: stablecoins as transaction money, bitcoin as the store of value. For bitcoin holders, this is the key passage. It places bitcoin at the base of the machine economy’s balance sheet rather than at the edge of it.

    Why Agents Need New Payment Rails

    BlackRock argues that capable agents “increasingly demand payment and asset infrastructure designed natively for machine-speed commerce.” Crypto rails fit high-frequency, sub-cent, 24/7 machine-to-machine payments for API calls, data, and compute. Legacy systems are poorly matched: account setup and credentialing assume a human, merchant fees make very small payments uneconomic, ACH settles in a business day or so, card disputes keep finality open for longer, and volume scaling is uncertain. Modified traditional rails will still serve agents dealing with human businesses and consumers, but the high-velocity machine layer points on-chain.

    The Agentic Protocol Stack: MCP, A2A, x402, ACP, and More

    The paper maps an emerging stack. At the base are Anthropic’s Model Context Protocol, which standardizes how AI applications reach external tools and data, and Google’s Agent2Agent, which lets agents coordinate across platforms. On top sit payment protocols. Coinbase’s x402 uses the HTTP 402 status code to let agents pay inside a web request with near-real-time verifiable settlement, reducing provider counterparty risk. It is chain-agnostic and led by USDC for now, and on permissionless networks its growth could drive demand for blockspace and validator services. Alongside it are Stripe and Tempo’s Machine Payments Protocol, Stripe and OpenAI’s Agentic Commerce Protocol for programmatic checkout on merchants’ existing rails, Google’s AP2 with cryptographic authorization mandates and audit trails, and Visa’s Trusted Agent Protocol for verifying trusted agents. BlackRock’s Figure 2 walks through a trip booking in which a primary agent, a travel sub-agent, x402 data purchases, and ACP checkout combine to return an itinerary and receipts to the user.

    Stablecoins Already Rival the Card Networks

    Stablecoins are expected to lead agent transactions because a stable unit of account makes pricing predictable. They are the largest tokenized RWA category at more than $300 billion in circulation. Adjusted stablecoin volume exceeded $11 trillion in 2025, in the same broad range as Visa and Mastercard (BlackRock’s Figure 3 notes the measures are not directly comparable), and grew at an 80% CAGR from 2020 to 2025, while ACH, still far larger at $93 trillion, grew about 8.5%. The GENIUS Act, MiCA, and Asian licensing regimes add tailwinds. Crucially, the benefits flow to the settlement networks. Stablecoins are issued across multiple chains, from general-purpose networks like Ethereum, where ETH pays for consensus and fees, to purpose-built chains like Circle’s Arc, where USDC itself is the gas asset. More payment activity means more demand for blockspace and potential value capture, subject to each network’s fee and staking design.

    Compute Becomes a Tradable Commodity

    AI needs vast amounts of compute and energy. Investors have focused on capex (estimates above $5 trillion from 2025 to 2030), but BlackRock stresses the operating spend that flows through the cloud compute market. Hyperscaler consensus implies about $1.1 trillion in cloud revenue by 2030. McKinsey projections in the paper’s Figure 4 show inference growing into the largest AI workload, with its share of data center power demand rising sharply. Large resource markets historically develop trading, financing, and hedging infrastructure, and GPU-backed financings are early evidence that compute is on the same path. BlackRock acknowledges real design problems (chip-generation productivity, regional energy costs, cash versus physical settlement) but points to commodity precedents like basis markets and contracts for difference. It expects exchange-traded compute futures and tokenized compute claims that can be transferred, pledged, and settled programmatically, which could broaden institutional participation and open a new market for digital assets.

    Agents That Buy Their Own Compute

    The paper’s Figure 5 imagines an agent running an extended analysis that continuously estimates its own compute needs, queries real-time marketplaces across GPU, specialized, and edge providers for price, latency, and reliability, and provisions capacity just in time, paying per use, per token, or per job over x402. BlackRock cites Stripe’s August 2026 agreement to acquire OpenRouter, which routes workloads across more than 400 models from over 80 providers, as an early strategic signal. Given Stripe’s work in payments, stablecoins, billing, and agentic commerce, the deal points toward agents that autonomously source and pay for compute over blockchains and programmable rails.

    BlackRock’s Conclusion

    BlackRock closes by saying AI and blockchain-based digital assets are converging as machines take a larger role in economic activity. Tokenization gives agents a direct interface to programmable assets, stablecoins and x402 handle high-frequency always-on payments, and liquid compute markets could let agents source, finance, and pay for the resources they run on. The ecosystem is early, with agentic payment activity and compute-market liquidity still limited. But as agents become more capable, the firm expects digital assets to become increasingly integral to AI’s economic infrastructure, expanding utility across stablecoins, tokenized RWAs, and native cryptoassets.

    Notable Quotes

    “At the core of this convergence, AI and digital assets both arise from a common foundation: AI represents machine-native intelligence, while digital assets represent machine-native money.”

    BlackRock, Executive Summary, the thesis of the whole paper in one sentence

    “This paper examines this growing relationship and explains why broad AI adoption may represent an underappreciated source of demand, utility, and application growth across the digital asset economy.”

    BlackRock, Executive Summary, on why the market is mispricing the AI and crypto link

    “These findings reflect simulated model responses rather than observed agent behavior, but point to a potential AI-native monetary architecture in which stablecoins serve as transaction money and bitcoin as a store of value.”

    BlackRock, on Bitcoin Policy Institute research into which money AI models prefer

    “Crypto-native blockchain rails are particularly well suited to high-frequency, sub-cent, machine-to-machine (M2M) transactions that take place around-the-clock, including API calls, on-demand data, and consumption-based compute.”

    BlackRock, on why agentic commerce points on-chain

    “Where settlement occurs on permissionless networks, greater usage could increase demand for blockspace and validator services, creating a potential transmission channel to native cryptoassets.”

    BlackRock, on how x402 payment volume could flow through to crypto assets

    “Adjusted stablecoin volume remained well below the $93 trillion transferred over ACH in 2025; from 2020 to 2025, however, it grew at an 80% CAGR, compared with approximately 8.5% for ACH.”

    BlackRock, on the growth gap between stablecoins and legacy payment rails

    “As this market expands, compute is becoming a distinct, large, and increasingly investable economic resource that could support a new class of digital assets.”

    BlackRock, on the emerging market for tokenized compute

    “In our view, this could support a future in which agents autonomously source and pay for compute over blockchains and other programmable payment rails.”

    BlackRock, on Stripe’s agreement to acquire OpenRouter

    Read the full BlackRock paper, The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute, here.

    Related Reading

  • OpenCode CEO Jay V on 20x Growth in 6 Months: 13 Million Users, 7 Trillion Tokens a Day, the Anthropic Block That Backfired, and the 16-Year Road to Overnight Success

    In this episode of Y Combinator’s Lightcone podcast, Jay V, founder and CEO of OpenCode, the open-source coding agent that works with any model, walks through one of the wildest growth stories in developer tools: 650,000 monthly active users in January to roughly 13 million by June, 7 trillion tokens processed per day, and a business that went from zero to a $40 million revenue run rate in about eight months. He also tells the part almost nobody knows: the company behind this “overnight success” is a 16-year-old legal entity that applied to Y Combinator nine times before getting in.

    TLDW

    Jay V explains how OpenCode grew 20x in six months to around 13 million monthly active users and 4.6 million weekly actives, processing 7 trillion tokens a day (more than OpenRouter’s entire volume), with an inference business annualizing near $40 million plus 160,000 subscribers worth another $18 million. The inflection point came when Anthropic started blocking Claude Code subscriptions inside OpenCode by rejecting requests whose system prompt contained the words “open code,” which backfired by equating the two products and sending curious users flooding in, shortly after which OpenAI’s Codex officially supported OpenCode. The conversation covers OpenCode’s public usage data (DeepSeek Flash dominating token volume despite GLM hype), a global user base led by China at 17% with heavy usage in Indonesia, Brazil, and Vietnam, Fortune 500 companies discovering thousands of employees already using the tool, the shift from ad-based CAC to token-based CAC, the flat 24-hour GPU utilization curve that comes from serving the whole planet, the “betting the field” marketplace thesis on model commoditization, and the founder’s 16-year, nine-application journey from a Waterloo dorm through SST, OpenNext, and selling coffee over SSH to finally catching lightning.

    Thoughts

    The Anthropic block is the most instructive growth story in the episode, because it is a perfect modern Streisand effect. Anthropic had a defensible reason to stop subsidized Claude Code subscriptions from flowing through a third-party harness, but the implementation (rejecting any request whose system prompt literally contained “open code”) turned a quiet policy decision into a public endorsement. As Jay puts it, the block placed OpenCode on the same pedestal as Claude Code in the minds of developers who had never heard of it. The hosts’ Instacart comparison is apt: when Amazon bought Whole Foods, the “death of Instacart” meme drove every grocer in America into Instacart’s arms. Incumbents keep learning this lesson the hard way. You cannot block a product without simultaneously advertising that it matters.

    The deeper story is geographic. Silicon Valley talks about coding agents as if the $200-per-month power user is the market, and Jay’s data says the opposite. China alone is 17% of OpenCode’s usage, with Indonesia, Brazil, and Vietnam each carrying meaningful share, places where a frontier subscription costs more than rent. OpenCode’s $10 Go plan, running DeepSeek and GLM instead of Sonnet and Opus, is how billions of developers will actually have their first coding-agent moment. There is also a hard operational edge hiding in that distribution: because the East works while the West sleeps, OpenCode’s GPU utilization runs a nearly flat 24-hour cycle, which quietly improves unit economics in a way no single-market competitor can match. Serving the whole planet is not just a mission statement. It is a margin strategy.

    OpenCode’s neutrality is turning into one of the most valuable datasets in AI. Because the product is a harness over every model rather than a storefront for one lab, opencode.ai/data shows what developers actually run when they are spending their own money, and it routinely contradicts the Twitter narrative. GLM was supposedly eating DeepSeek’s lunch; the token-volume charts show DeepSeek Flash dipping and then bouncing right back. Users are not loyal, they are rational: they ride frontier limits until they hit caps, then switch to models cheap and fast enough to finish the day’s work. That behavioral reality, boring cost optimization rather than fandom, is what the model market actually looks like once the marketing fog clears, and only a neutral aggregator gets to see it.

    The business model inversion deserves more attention than it usually gets. In the last era, customer acquisition cost meant ads. In this one, it means tokens: the free tier is the marketing budget, spent on giving people the aha moment, and the payoff comes when a fraction of those users become whales paying per token, where OpenCode’s volume discounts become margin. This is the same funnel Anthropic and OpenAI run, except the frontier labs subsidize with investor billions while OpenCode rides the falling cost curve of open-weight models. The enterprise motion follows the same bottoms-up physics: no procurement dance, just inbound emails saying thousands of our employees are already using you, please sign the security questionnaire. That is the purest product-market-fit signal that exists.

    And then there is the 16-year overnight success. Same legal entity since 2010, same two founders from a Waterloo dorm room, nine YC applications and four interviews before acceptance in 2021, years of living with parents and running out of money, a serverless framework, a coffee shop that ran over SSH. Every “dead end” turns out to have been training: the consumer company taught metrics discipline, SST taught open source and building in public, the terminal storefront taught terminal-UI craft that made OpenCode instantly credible with the Neovim crowd. The hosts land the right conclusion: lightning did strike, but the founders spent a decade positioning the bottle. In an industry currently obsessed with six-month-old unicorns, this episode is a useful reminder that most of them are carrying more history than the headline suggests.

    Key Takeaways

    • OpenCode ended June 2026 at roughly 13 million monthly active users and 4.6 million weekly actives, close to Codex’s numbers, a 20x increase from about 650,000 monthly actives at the start of the year.
    • The platform now processes around 7 trillion tokens per day, more than OpenRouter’s total of roughly 6 trillion, up from about 300 billion per day at the beginning of the year.
    • The pay-per-token inference business, launched around late September 2025, annualizes to $31-33 million on June data and $38-40 million on the most recent week, roughly eight months from zero.
    • The subscription product launched in late February has grown to about 160,000 monthly subscribers, roughly $18 million in annualized revenue on top of inference.
    • A Codex lead engineer publicly noted that about 5% of all Codex subscribers use OpenCode as their main harness, and OpenAI officially supports Codex subscriptions inside OpenCode.
    • In the first week of January, Anthropic began blocking Claude Code subscriptions in OpenCode by rejecting any request whose system prompt contained the words “open code.”
    • Jay concedes the block made business sense (Anthropic subsidizes that usage) but says it inadvertently equated OpenCode with Claude Code and drove waves of new users to investigate the product.
    • The hosts compare it to Amazon buying Whole Foods: the “death of Instacart” meme drove every grocer in America to sign with Instacart, fueling its growth instead of killing it.
    • The founding premise is that most people in the world still have not experienced the magic of a coding agent, and frontier per-token prices put that moment out of reach for much of the globe.
    • When OpenCode launched in June 2025 the pitch was using your Claude Code subscription in a better terminal UI; by August and September the first credible open-source models (GLM, Kimi, MiniMax) arrived, roughly six months behind the frontier.
    • February 2026 marked the first four-week span in OpenCode’s data where users ran Gemini more than the Anthropic models (Sonnet plus Opus combined), which convinced the team the non-Anthropic models were ready for real work and triggered the subscription launch.
    • OpenCode publishes its usage data at opencode.ai/data, covering the Go plan where $10 a month buys access to open-source models.
    • DeepSeek Flash leads token volume per day, with the two DeepSeek models plus GLM as the top three, despite social media chatter suggesting GLM had overtaken DeepSeek.
    • By unique users the top models run DeepSeek Flash at about 38,000, DeepSeek Pro at 31,000, and GLM 5.2 near 30,000.
    • A key usage pattern: as users approach daily or weekly limits on premium models, they switch to very cheap models like DeepSeek Flash to finish their work, extending how much coding-agent time their budget buys.
    • Speed matters too: some open models are hosted with far higher tokens-per-second than alternatives, making the agent feel near real time, and users perceive quality niches, like GLM 5.2 being better at front-end design.
    • China is OpenCode’s largest market at 17% of usage, which the hosts note may make it the only YC company in history with meaningful usage in China, partly because Chinese developers want to run Chinese models and OpenCode gives them that choice.
    • Developing countries are huge: Indonesia at 4% of traffic, Brazil at 5%, plus Vietnam and similar markets where a $200-a-month Claude Code subscription is prohibitively expensive.
    • The US, which the team was not even targeting with the Go plan, is growing strongly anyway, which Jay reads as a broader vibe shift toward token budgeting even among Americans.
    • Large US companies with effectively unlimited token budgets also adopted OpenCode early because they did not want to be locked into a specific model or harness.
    • Dozens of forward-leaning Fortune 500 companies have significant OpenCode footprints, often discovered when the company itself emails saying thousands of employees are already using it.
    • Enterprise inbound has inverted the old SaaS procurement dance: companies beg OpenCode to fill out security questionnaires so they can officially use a product their engineers already adopted.
    • Enterprise pull comes in four flavors: officially blessing developer usage, extending the tool to non-technical employees, embedding the agent loop inside their own products, and managing token spend by routing teams to cheaper models.
    • One enterprise asked for deep visibility into exactly what every employee does with the tool, which the team flagged as a should-we-even-build-this question.
    • Ramp built a Slack bot running OpenCode’s embeddable server (the agent loop that works behind the UI) before OpenCode had built anything similar internally, publishing a blog post about it in December.
    • OpenCode is architected as a two-part product: the terminal UI you interact with, and a separately embeddable server that runs the agent loop and calls the LLM.
    • The new CAC is tokens, not ads: the free tier exists to give people the magic moment, the subscription converts them to real work, and whales paying per token feed directly into margin via OpenCode’s volume discounts on inference.
    • The episode references Dylan Patel’s podcast claim that Anthropic reached roughly $50 billion annualized revenue at around 70% margin in Q2, proof that the subsidize-then-harvest funnel can cross into profitability.
    • Global usage produces a nearly flat 24-hour GPU utilization curve (the East works while the West sleeps), improving unit economics versus competitors serving one region.
    • Jay describes OpenCode as a marketplace that showcases model diversity: competition among labs benefits consumers, while vendor lock-in mostly benefits vendor margins.
    • OpenCode is now the largest customer by token volume for most open-source model labs, making the relationship symbiotic: the strategy is not picking a winning lab but betting the whole field.
    • Every bump in OpenCode’s monthly actives traces back to a corresponding release in the open-source model market, making its growth a proxy for open-model progress.
    • The name OpenCode was deliberate positioning: when a market has one or two dominant players, the rest coalesces around an open alternative, and whoever occupies that position first is very hard to displace.
    • To support 70+ models and providers at launch, the team built models.dev, an open-source database of models and providers that Jay calls probably the best such dataset in the world.
    • The origin moment: when Claude Code appeared in February 2025, the team (Neovim users unimpressed by its terminal UI) decided to build a coding agent that met the standard of modern terminal tools, credibility that resonated instantly with the core developer audience.
    • The team had form here: co-founder Dax had built terminal.shop, a complete storefront for buying coffee over SSH, the kind of eccentric-taste project the hosts argue pulls founders toward outlier outcomes.
    • The company is one 16-year-old legal entity, incorporated in 2010, founded by Jay and his college roommate Frank after a Waterloo co-op term convinced Jay he never wanted a normal job.
    • Jay applied to YC nine times between 2016 and 2021 with four interviews before getting in, with his first interview dating back to the era when Paul Graham ran them and an Airbnb founder was hanging around the waiting room.
    • The 2021 YC idea was a serverless platform, Heroku for AWS, which became SST, the team’s first big open-source project and the on-ramp to building in public.
    • Building in public became core identity after co-founder Dax observed that if all your code is public and you work in public, staying silent about it is a disservice to the product; the community now follows the company like a reality TV show.
    • Jay credits survival to stubbornness, visible forward progress, and cheap burn (living with parents after running out of money), while warning founders: don’t try this at home.
    • The hosts’ framing of the whole arc: it took ten years of grinding to get to zero-to-$30-million in eight months, and catching lightning in a bottle requires positioning the bottle correctly first.

    Detailed Summary

    The Numbers: 20x in Six Months

    OpenCode began the year around 650,000 monthly active users and ended June near 13 million, with 4.6 million weekly actives that put it in the same conversation as OpenAI’s Codex. Token throughput grew from roughly 300 billion per day to 7 trillion, a volume larger than all of OpenRouter. The money followed two tracks: a pay-per-token inference business launched in the fall that annualizes near $40 million on recent weeks, and a subscription product launched in late February that reached 160,000 monthly subscribers and about $18 million annualized. Codex officially supporting OpenCode, with around 5% of Codex subscribers choosing it as their harness, added a second frontier on-ramp right as the Anthropic controversy peaked.

    The Anthropic Block That Backfired

    Using a Claude Code subscription inside OpenCode was one of the most common usage patterns until Anthropic moved to stop it in early January, rejecting requests whose system prompt mentioned “open code.” Jay is gracious about the logic (Anthropic subsidizes subscription usage and wants it inside its own product) but the effect was the opposite of containment. The block put the scrappy open-source harness on the same pedestal as the category leader, told every developer who had not tried it that it was worth investigating, and kicked off the year’s 20x run. The hosts draw the Instacart parallel: a supposed death blow that functioned as the best marketing campaign the company never paid for.

    A Global User Base the Valley Doesn’t See

    The product premise is that the coding-agent aha moment is a once-a-generation experience most of the world cannot afford at frontier prices. The Go plan ($10 a month for open-source models) was built for that global audience, and the geography shows it: China leads at 17%, with Indonesia at 4%, Brazil at 5%, and Vietnam prominent, markets where $200 a month is simply not a consumer price point. Two surprises followed. Chinese developers use OpenCode partly to run their own country’s models, which no US-locked product lets them do. And the US, never the target for Go, is growing fast anyway, which Jay reads as the token-budgeting vibe shift reaching even the throw-money-at-it crowd, helped by moments like GLM 5.2’s popularity making the plan the easiest way to try it.

    What the Usage Data Really Shows

    OpenCode publishes per-model usage at opencode.ai/data, and because every data point is an actual end user rather than aggregated API traffic, it is arguably the cleanest picture of what working engineers really run. DeepSeek Flash dominates token volume, the two DeepSeeks plus GLM hold the top three, and the market-share graph shows DeepSeek dipping when GLM launched and then bouncing back, contradicting the Twitter narrative of a GLM takeover. By unique users, Flash leads at 38,000 with DeepSeek Pro at 31,000 and GLM 5.2 near 30,000. The behavioral driver is pragmatic: cheap, fast models let users keep working after they hit premium limits, hosted speeds make some models feel real time, and perceived niches (GLM for front-end design) steer specific workloads.

    Enterprises Arriving Through the Back Door

    Before the open-model wave, companies adopted OpenCode to avoid lock-in to any single model or harness. Now dozens of forward-thinking Fortune 500 companies have significant footprints, and the procurement process has inverted: instead of sales outreach, OpenCode receives DMs saying a few thousand employees are already using the product, please sign the security questionnaire, and often, please don’t tell anyone. Once inside, enterprises pull in predictable directions: extend access to non-technical staff, embed the agent loop in their own products, and manage token spend by restricting expensive frontier models to teams that need them. Ramp exemplified the embedding path, running a Slack bot on OpenCode’s server component before OpenCode itself had tried it. One request, total visibility into employee activity, raised the harder question of what the company is willing to build.

    Token Economics: CAC Is Now Paid in Tokens

    The episode’s sharpest business insight is that customer acquisition cost has migrated from ads to tokens. Becoming skilled enough with coding agents to justify heavy spend is itself expensive, a chasm most individuals and companies cannot cross unaided. Anthropic and OpenAI solve this by subsidizing subscriptions until a percentage of users become whales, and per Dylan Patel’s numbers cited in the episode, that funnel has carried Anthropic to roughly $50 billion annualized at 70% margins. OpenCode runs the same funnel without frontier-scale subsidies: the free tier delivers the magic moment, the $10 plan makes real work affordable on open models, and whales paying per token convert OpenCode’s volume discounts into margin. The flat 24-hour GPU utilization curve from serving every timezone compounds the advantage.

    Betting the Field: The Marketplace Thesis

    Jay frames OpenCode as a marketplace where users pick models by attribute and cost, which keeps labs honest and passes competitive gains to consumers instead of vendor margins. Every bump in OpenCode’s growth traces to a release in the open-model market, so the company is explicitly not picking a winning lab; it is betting the field. That bet has made OpenCode the largest customer by token volume for most open-source model labs, a symbiosis where each side needs the other. On commoditization, Jay’s view is nuanced: the intelligence market is so large that labs will carve defensible niches along the quality-cost-performance axes, the way DeepSeek deliberately owns the cost corner. The positioning strategy has deep roots: as with the team’s earlier OpenNext project, when a market has two dominant players, the rest coalesces around an open alternative, and OpenCode raced to become that default, building models.dev along the way just to support 70+ providers at launch.

    Sixteen Years to Overnight Success

    The backstory reframes everything. Jay started the company after a discouraging Waterloo co-op term in 2006-2007, incorporated with college roommate Frank in 2010, and spent the next decade shipping products that did “reasonably well” while applying to YC nine times across 2016-2021, with four interviews, all as the same legal entity, the same founders, and a rotating cast of ideas. His first YC interview was with Paul Graham, in a waiting room shared with an Airbnb founder. Acceptance finally came in 2021 with the serverless platform that became SST, the team’s gateway into open source and building in public, a practice pushed by YC’s Dalton and crystallized by co-founder Dax’s observation that public code deserves public storytelling. When Claude Code landed in February 2025, the team’s terminal-UI taste (honed on projects as eccentric as coffee-over-SSH) told them exactly what to build. The hosts close on the honest version of the lightning-in-a-bottle myth: ten years of grinding taught the team consumer metrics, open source, marketing, and positioning, so when the strike came, the bottle was already in place.

    Notable Quotes

    “Most people in the world still haven’t experienced the magic of a coding agent.”

    Jay V, on the founding premise of OpenCode

    “You really know you have product market fit when like enterprises are bugging you to sign the security agreement so they can use your product.”

    Lightcone host, on OpenCode’s inverted enterprise sales motion

    “It’s not that we’re picking a winner in terms of a model lab. We’re just betting the field. We just think the rest of the field is going to do well.”

    Jay V, on OpenCode’s strategy toward the model market

    “With these open-source models, we’re the largest customer for most of them.”

    Jay V, on OpenCode’s token volume relative to open-model labs

    “When you’ve got a dominant or in this case two dominant players in the market, the rest of the market coalesces around an open alternative. And picking that position ends up being really valuable because if you pick it, it’s very hard for somebody else to displace you.”

    Jay V, on the deliberate positioning behind the OpenCode name

    “This is just an unprecedented market, like the market for intelligence has not existed before, everybody should be thinking in a positive-sum grow-the-pie mentality.”

    Lightcone host, on why labs should welcome OpenCode’s growth

    “Look, you know, all your code is public. You work basically in public. If you don’t talk about it publicly, you’re probably doing yourself a disservice and your product a disservice.”

    Jay V, recounting co-founder Dax’s case for building in public

    “It was really more a journey that took 10 years to get to 0 to 30 million in 8 months.”

    Lightcone host, reframing the overnight-success narrative

    “To catch the lightning in the bottle, you actually like have to sort of position the bottle correctly and be ready for it and know what to do with it.”

    Lightcone host, closing the episode on preparation meeting luck

    Watch the full conversation here.

    Related Reading

    • OpenCode the open-source coding agent discussed throughout the episode, including its public usage data.
    • models.dev the open-source database of AI models and providers the team built to support 70+ providers at launch.
    • SST the serverless framework that got the company into YC and established its open-source, build-in-public roots.
    • Terminal the coffee-over-SSH storefront that proved the team’s terminal-UI chops before OpenCode existed.
    • Y Combinator the accelerator behind the Lightcone podcast, which Jay applied to nine times before getting in.