The Stack • Part 5 — The Harness

The Two-Hundred-Dollar Seat That Costs Eight Thousand

A flat-rate coding-agent subscription can cost its heaviest users forty times its sticker price to serve — and the incumbent's own banker calls the layer that hides this an escape hatch.

July 2026 A Wall-Street teardown of the AI compute stack
Allegorical oil painting: a customer stands at a polished café counter with a brass placard reading $200 a month and a tiny cup of molten gold; beneath the glass floor, gloved hands shovel gold coins into a furnace piped up to that single cup, its pressure gauge pinned in the red; datacenter halls and a mobile generator trailer sit beyond the windows.
The two-hundred-dollar seat, as served: below the floor, the boiler burns gold to keep one cup warm. The gauge is pinned.Illustration — AI-assisted

In June 2026, Microsoft did something none of the first four layers of this stack ever do in public: it looked at its own internal bill for a tool it didn't build, decided the number didn't work, and pulled the plug. It canceled the majority of its internal Claude Code licenses, moved thousands of engineers to its own GitHub Copilot CLI, and left behind a tidy piece of tape-reader's evidence. The tool hadn't gotten worse. The bill had gotten honest: a $200-a-month subscription seat, run the way an engineer actually runs an agentic coding session, can consume compute that costs as much as $8,000 a month to serve at metered rates — a gap SemiAnalysis sizes at roughly forty times the sticker price.12 Parts 1 through 4 asked who holds the shell, the silicon, the rented cloud, the frontier model itself. Here the question turns personal: who eats the token when the seat costs a fraction of what it actually runs through the meter? And for the first time in six layers, the name for this rung of the stack — “the harness” — wasn't coined by a critic. It's the word J.P. Morgan's own house economist uses for it, in the same report where he explains why it's the industry's escape hatch.

Where this sits

Part 5 of The Stack, one rung up from the frontier models. Same two questions, asked of the orchestration layer: who books the equity, and who eats the subsidized loss? The frontier labs sell intelligence by the token; the harness — Claude Code, OpenCode, Cursor, Devin, the RL-as-a-service shops — is the software that decides, request by request, whether that intelligence has to come from the most expensive model available. That decision is worth real money, and 2026 is the year both sides of the transaction started fighting over who pays for the token in between.

/ 01The Two-Hundred-Dollar Seat That Costs Eight Thousand

Start with the product itself, because the arbitrage is built into its pricing tiers on purpose. Claude Code — Anthropic's own coding agent, launched publicly in May 2025 — sells three consumer-facing plans: Pro at $20 a month, Max 5x at $100, and Max 20x at $200, each buying a multiple of Pro's usage ceiling rather than a fixed token allowance.3 Enterprise self-serve is $20 a seat — and once an account grows past the Team plan's 150-seat cap, that $20 buys access, not tokens; usage past that point is fully metered at API rates, with no included allowance at all.1 The flat fee is a door. What's behind it is priced by the token.

SemiAnalysis ran the experiment the plans are daring someone to run: buy the subscription, use it flat-out to its stated ceiling, and price every token consumed at Anthropic's own posted API rates. Applied to a Claude Max 20x seat, the metered-equivalent bill comes to as much as $8,000 a month — about forty times the $200 charged for it.12 SemiAnalysis ran the same methodology against OpenAI's $200-a-month ChatGPT Pro plan and found a metered-equivalent bill of up to $14,000 for a user who fully exhausts its ceiling — a roughly seventy-fold gap, though the two figures aren't perfectly comparable across different models and task mixes.4

$200 → $8,000Claude Max 20x, sticker vs. metered-equivalent at full usage1
$200 → $14,000ChatGPT Pro, same SemiAnalysis usage-ceiling method4
$39 → $800+A real Copilot bill after the June 2026 usage-based switch5

Who ate that gap before anyone had to talk about it? The vendor, blended across a much larger base of light users. Anthropic's overall gross margin runs around 44% — about $0.71 of compute cost against every dollar of revenue in Q1 2026, projected to improve to $0.56 in Q2.16 That's a perfectly ordinary SaaS pattern: heavy users cost more than they pay, light users pay more than they cost, and the blend works as long as the heavy tail stays a tail. Two things happened in the same quarter that suggested it might not be staying a tail.

First, Microsoft. It canceled the majority of internal Claude Code licenses across its Experiences + Devices division, effective by June 30, 2026, after the tool's token-metered cost “ran past the annual AI budget months ahead of schedule.”5 One contemporaneous account frames the logic in a single sentence worth keeping: “Cost pulled the trigger. What cost exposed is the real lesson: renting intelligence by the token is a price Microsoft does not control.”7 The same reporting cluster claimed Uber burned through its entire 2026 AI-tools budget on Claude Code and Cursor in just four months — a vivid number, but one that traces to a single secondary aggregator with no primary corroboration located; treat it as a flare, not a fact, until Uber or a first-tier outlet confirms it.5

Second, GitHub. On June 1, 2026 — the same week — GitHub Copilot switched every plan from opaque “premium request” credits to literal, token-metered billing. Headline seat prices didn't move ($10 Pro / $39 Pro+ / $19 Business), but the meter underneath them was exposed to daylight for the first time, and users reported real bills jumping from $39 to over $800 a month, with some reporting increases as high as 100x.8910 Nothing about the underlying compute cost changed on June 1. What changed is that Microsoft stopped absorbing the gap between flat pricing and actual agentic-workflow consumption — on its own product, in the same month it stopped absorbing that gap on someone else's.

Allegorical oil painting: a diner counter with a finished breakfast and coins in a worn hand, while the brass taxi meter beside the register prints a receipt so long it loops across the checkered floor and out the door.
The morning the meter appeared: breakfast unchanged, the receipt now runs out the door — the flat rate was someone else absorbing the gap.Illustration — AI-assisted
Flag before you print these

The two SemiAnalysis usage-ceiling figures — $200-vs-$8,000 for Claude and $200-vs-$14,000 for ChatGPT Pro — are the single most load-bearing numbers in this section, and neither was read from a directly fetched SemiAnalysis publication in this research pass. The Claude figure comes via J.P. Morgan's reproduction of SemiAnalysis's work; the ChatGPT Pro figure comes via secondary aggregation of the same underlying SemiAnalysis research. Both are consistent with the pricing tiers and the Microsoft/GitHub repricing that followed — the direction and rough scale check out — but a direct pull from SemiAnalysis's own newsletter archive should replace this sourcing before either number is treated as filed fact rather than well-corroborated estimate.

/ 02The Businesses

Five kinds of company sit at this layer, and only one of them is profitable in any sense a public filing would recognize.

Claude Code is Anthropic's own harness — the reference implementation of “orchestration running a closed frontier model” — and by its own account it's working: over $2.5 billion in run-rate revenue as of February 2026, more than doubling since the start of the year, business subscriptions quadrupling since January, and enterprise customers now more than half of Claude Code's revenue.11 OpenCode is its open-source mirror image: the leading model-agnostic harness, plugging into 75-plus model providers — Claude, GPT, Gemini, open weights, local models — with no single vendor to toll it. As of late June 2026 it led Claude Code on GitHub stars (roughly 180,500 to 135,000) despite launching about two months later, with 900-plus contributors and a claimed 7.5 million monthly developers.12 It has no disclosed revenue, because it isn't really a company in the conventional sense — its business model is closer to “commoditize the harness so no one can toll it” than to venture-scale SaaS, which makes it this layer's cleanest structural counter-example to the idea that every good harness eventually gets captured. There's simply no equity to acquire.

Cursor, built by Anysphere, is the layer's fastest-moving equity story, full stop. Watch the valuation staircase:

1

$9.9B — Series C, early 2025

A well-funded developer tool, not yet a phenomenon.

2

$29.3B — Series D, November 2025

Post-money, at roughly $1B ARR. Three times the Series C valuation in under a year.

3

$50B pre-money — in talks, April 17, 2026

ARR had reached $2B — doubled in five months — and the valuation rose roughly 70% to match.13

4

$60B all-stock — SpaceX acquisition, announced June 16, 2026

Reported ARR by then: roughly $4 billion, up from about $100 million in early 2025.14

Cursor is also, notably, gross-margin-positive at the enterprise-account level as of April 2026 — individual-developer accounts are still loss-making — driven substantially by its own proprietary Composer model handling roughly half of autocomplete requests and routing the rest to cheaper third-party models like Kimi instead of defaulting to frontier tokens.15 That is worth sitting with: the harness business closest to profitable got there specifically by reducing its dependence on the most expensive frontier inference — the strongest evidence in this entire layer that the escape-hatch thesis isn't just a banker's turn of phrase.

Cognition (Devin) sells autonomous, ticket-to-pull-request coding at $20/month plus $2.25 per unit of agent-compute (roughly fifteen minutes of autonomous work), or $500/month with 250 units included.16 Its path here runs through one of the stranger M&A chains in the whole series: OpenAI agreed to buy the IDE maker Windsurf for $3 billion in May 2025; the deal's exclusivity lapsed that July over an IP conflict with Microsoft; Google then paid $2.4 billion for a non-exclusive license to Windsurf's technology and hired away its CEO and top researchers — a “reverse acquihire,” the same structure used on Character.AI and Inflection; and Cognition closed a deal for Windsurf's remaining IP, brand, and roughly 210 employees within 72 hours of Google's move.17 Cognition then raised $400 million at a $10.2 billion valuation two months later, and over $1 billion at $25 billion pre-money in late May 2026, reaching $492 million in annualized run-rate revenue — with Devin reportedly writing 89% of Cognition's own code.1819 No margin figure has been disclosed alongside any of it.

LangChain sits one level up the stack from the coding-specific harnesses — infrastructure for building any agentic application, not a coding assistant itself. It raised $125 million at a $1.25 billion valuation in October 2025, and reports 90 million monthly downloads with adoption across 35% of the Fortune 500.20 Its revenue is the thin spot in an otherwise well-documented layer: a single third-party aggregator put 2025 ARR at roughly $16 million — unverified, company-undisclosed, and if directionally right, a revenue-to-valuation ratio near 78x, an outlier even by 2026 AI-infrastructure standards.20

And then there's the sharpest form of the thesis: RL-as-a-service. Ramp used Prime Intellect's open-source prime-rl stack to post-train a 35-billion-parameter open model that beat Claude Opus 4.6 by several points on Ramp's own retrieval benchmark, running at Haiku-level cost and latency.21 Harvey had Trajectory Labs post-train Nvidia's open-weight Nemotron 3 Ultra on Harvey's own Legal Agent Bench in under 24 hours, reaching “the same band as leading closed models” on legal work at a fraction of frontier token cost — a claim that comes from Trajectory Labs and Harvey themselves and hasn't been independently benchmarked by a third party.22 J.P. Morgan frames the same category of result more starkly still: Opus 4.6 costs $3,700 to run a standard benchmark task set for a score of 56; DeepSeek V4 Pro scores 44 for $186 — roughly twenty times cheaper.1 This isn't routing to a cheaper model mid-request. It's a customer training its own model so it never has to rent the frontier lab's intelligence for that task again.

BusinessLatest disclosed figureProfitability signal
Claude Code (Anthropic)$2.5B+ run-rate, Feb 202611Parent's ~44% blended gross margin; no operating profit disclosed1
OpenCode7.5M monthly devs (claimed)12No revenue disclosed — not a conventional company
Cursor (Anysphere)~$4B ARR, Jun 202614Gross-margin positive, enterprise accounts only15
Cognition (Devin)$492M ARR run-rate18No margin figure disclosed
LangChain~$16M ARR (single estimate)20Too thinly sourced to assess
Compiled from company disclosures, funding announcements, and third-party aggregators as cited; figures marked in red rest on a single secondary source or are simply absent from any public disclosure.
Steelman — the productivity is real

None of this is a shell game. Developers using these tools ship measurably faster; Cursor's enterprise gross margin and Ramp's benchmark win aren't marketing claims, they're the kind of number a finance team signs off on before renewing a contract. The harness genuinely does what Cembalest says it does — it lets a well-built orchestration layer extract most of a frontier model's value from a much cheaper one underneath it. The bear case in this section isn't that the product doesn't work. It's that almost none of the businesses selling it have found a durable price for what it costs them to deliver — and the one that has, Cursor, got there by routing away from the frontier tokens the rest of this stack is built to sell.

/ 03The Escape Hatch, in the Banker's Own Words

Here is the sentence that gives this layer its name, and it did not come from a critic of the AI boom. It came from J.P. Morgan's own house economist, describing exactly the mechanism this section has been building toward:

“Agent harnesses (external systems such as memory, tools, safety boundaries and orchestration layers in Claude Code or OpenCode) that can run open models often improves their output quality and reduces the need to rely on the most expensive frontier models.” — Michael Cembalest, “Semiquincententacles,” J.P. Morgan Eye on the Market, June 2026, p.15

Read that sentence again in the context of everything Parts 2 through 4 documented: Nvidia's eroding accelerator share, the frontier labs' circular profit, the $2 trillion of cloud backlog resting on counterparties whose revenue doesn't exist yet. The word “harness” is Cembalest's own coinage for the layer that lets a customer walk away from all of it — and he's not speculating. The migration is dated, documented, and happening in production, not in a slide deck.

Lindy AI, a 25-person startup, moved its entire inference stack off Claude and onto DeepSeek V4 Flash, cutting inference cost on the migrated routes by roughly 90%, in a public postmortem dated June 24, 2026.23 Its CEO's own account is unglamorous in exactly the way that makes it credible: AI costs had exceeded personnel costs at the company, and moving off Claude was, in his words, “a matter of survival for the business.”24 It's worth noting Lindy didn't move everything indiscriminately — it tested Kimi K2.5, GLM 5.1, and Kimi K2.6 against production traffic and rejected all three on real-world “vibe” despite promising offline benchmark scores, and kept Claude Sonnet for the subset of tasks that needed genuinely higher intelligence. The harness didn't just swap the cheapest model in; it did the discernment a flat “switch to open-weight” headline would erase.

Coinbase's CEO made the same bet in public, in numbers. On June 8, 2026, Brian Armstrong posted that Coinbase is “working hard on routing prompts to cheaper models,” predicting that within twelve to eighteen months, 80% of workloads will run on models that are 99% cheaper than today's frontier tier — with only the top 20%, the scientific-breakthrough and high-level-orchestrator work, staying on frontier tokens.25 And this isn't one company's anecdote against the grain: J.P. Morgan's own OpenRouter data shows Chinese open-weight models — Qwen, DeepSeek V4, Kimi — surging through the API-call leaderboard by April 2026, landing within “a few dozen Elo points” of closed frontier models at ten to fifty times lower token cost.1

Allegorical oil painting: the bank's own gilded frontier gate with turnstile and velvet ropes, a radiant gold gear on a cushion inside, while the banker's arm points a fountain pen at a small service door propped open with a brick, through which the same gear glows on a plain workbench.
The bank’s own gate, and the bank’s own banker pointing past it: the same gear glows through the service door at a fraction of the toll.Illustration — AI-assisted

“The escape hatch from frontier lock-in is documented, dated, and footnoted — by the same bank that underwrites the lock-in.”

Steelman — this is the layer working as designed

It would be easy to read the harness cynically — as marketing dressed up as infrastructure. That reading doesn't survive contact with Lindy's postmortem or Coinbase's public routing bet, both of which are dated, both of which put a real company's own cost line on the record. A well-built agent harness genuinely does decouple “what the developer experiences as quality” from “which lab trained the model underneath it” — that's the mechanical fact this whole section rests on, not a sales pitch. The honest complication isn't that the escape hatch is fake. It's the one the next section takes up: who owns the door.

/ 04Escape, or Relocated Toll Booth?

The single biggest equity event at this layer in the entire research window runs directly against the escape-hatch thesis, and it happened nine days after Cembalest's harness sentence went to print.

SpaceX had just absorbed xAI in a $1.25 trillion merger in February 2026 and completed the largest IPO in history on June 12. On June 16, 2026, it announced it would acquire Cursor — the layer's largest venture-funded, model-agnostic harness — for $60 billion in stock, explicitly to put Grok in front of Cursor's developer base.26 The deal had reportedly been pre-committed as far back as April 2026, ahead of SpaceX's own IPO, structured as either the $60 billion stock acquisition or a $10 billion break-up fee if it fell through — which tells you how much certainty both sides wanted before SpaceX went public.27 Cursor's four cofounders, average age about twenty-five, each became roughly $2.7 billion richer on paper the day the deal was announced.28

Photorealistic editorial illustration: an auction room where the lot on the velvet platform is a plain emergency-exit door under a gallery spotlight, green exit pictogram above it, gavel raised, one absurdly large gilded paddle rising from the front row, and a lot card reading sixty billion dollars.
The exit itself goes under the gavel: the layer built to route around the house, bought by the house for $60 billion.Illustration — AI-assisted

Read straight, that's a genuinely different equity story from every other layer in this series. The value didn't flow up to whoever owns the scarce shell, the scarce silicon, the rented cloud, or the frontier model — it flowed to the team that built the software sitting between the user and whichever model happened to be cheapest that week. Cognition's Windsurf-to-$26B trajectory tells a version of the same story with different names. For once in this series, the escape hatch's builders got paid like the escape hatch was the valuable thing.

But sit with what the buyer is. The instant a harness gets large and valuable enough, a frontier-model-adjacent conglomerate bought it outright — folding the layer built to route around expensive frontier models into the balance sheet of a company that also owns one. As of this writing, the deal had not closed (targeted for the third quarter of 2026), and no public statement commits Cursor to Grok-exclusivity or confirms it will keep routing to Claude, GPT, Gemini, and Kimi the way it does today.27 That silence is the whole ballgame. If Cursor keeps routing wherever the model is best and cheapest after the deal closes, this was capital rewarding the escape hatch. If Cursor's routing quietly narrows toward Grok, this was the house buying the exit.

The open question this whole part rests on

This is the single most consequential unresolved fact in this layer, and it is explicitly the fact Part 7's synthesis needs answered: does the SpaceX–Cursor deal preserve or kill Cursor's model-agnosticism? It is unresolved because the deal itself is unresolved — it had not closed as of the research date behind this article, and Anysphere, as a private company being acquired, has made no post-close routing commitment. Anthropic's own hedge here is instructive: the frontier lab whose token pricing power the harness is supposed to erode still owns the single most commercially successful harness product in the layer, Claude Code, at $2.5B-plus run-rate. If the harness wins regardless of which model sits underneath it, a frontier lab would rather own the harness too than bet everything on being the model.

/ 05Synthesis — Follow the Token

Follow the token: the subsidy at this layer is the cleanest, best-dated one in the whole series — a flat-rate seat priced well below what its heaviest use costs to serve at metered rates, absorbed for as long as the heavy users stayed a minority of the base. Anthropic's blended 44% gross margin is not evidence of a scam; it's evidence the arbitrage was, until recently, an ordinary SaaS subsidy pattern, light users paying for heavy ones. What makes 2026 different is that two of the parties closest to the meter — Microsoft as a buyer, Microsoft/GitHub as a seller — both concluded in the same month that the heavy tail wasn't staying a tail, and both repriced rather than keep absorbing it. The open question the labs haven't disclosed is the one that decides whether this was a one-time correction or the start of a trend: as agentic, multi-step, tool-calling workflows become the default way developers use these products rather than the edge case, does the average subscriber's usage keep drifting toward the ceiling that produces an $8,000-equivalent bill? Microsoft's and GitHub's own internal numbers, whatever they were, said yes loudly enough to act on inside a single quarter.

Follow the equity: this layer's steelman is the most genuine one in the series so far — real developer productivity, a real mechanism (Lindy, Coinbase, the JPM OpenRouter data) by which the harness lets a customer walk away from the most expensive frontier tokens without walking away from quality. The bear case is equally plain: almost nobody at this layer is profitable in a way a filing would confirm, and the one business that is — Cursor, at the enterprise-account level — got there specifically by reducing its own dependence on the frontier tokens the rest of the stack survives on. Whether that business stays the escape hatch or becomes the house's own toll booth is a question the SpaceX deal opened and has not yet answered.

Every layer in this series so far has answered “who holds the risk” with an institution — a builder, a bondholder, a ratepayer, a pension. This is the first layer where the answer is closer to a design choice: a harness that stays model-agnostic genuinely relocates value away from whoever owns the frontier model. A harness that gets bought by one doesn't. The variable that decides which future you're underwriting when you use one of these tools isn't the model. It's who owns the software making the choice for you — and, one rung further up, who owns the data that software is working on. Part 6 follows the token into the ontology.

Sources & Citations

[1] Michael Cembalest, J.P. Morgan Asset & Wealth Management, “Semiquincententacles: the US grip on markets on the 250th anniversary of the Declaration of Independence,” Eye on the Market Special Edition, June 2026, p.15 — the harness quote; the Claude Max 20x $200-vs-$8,000 figure (cited to SemiAnalysis); Anthropic's Q1/Q2 2026 compute-cost-per-revenue-dollar ($0.71/$0.56); the OpenRouter Chinese-open-weight-model data; the Opus 4.6-vs-DeepSeek V4 Pro benchmark-cost comparison ($3,700 for a score of 56 vs. $186 for 44); the Cursor/Anthropic 6x-fast-token example. Extracted text: semiquincententacles-JUN2026.txt, pp.13–16 (source PDF not in repo). Accessed 2026-07-02.
[2] SemiAnalysis, cited via JPM Cembalest (source [1]) for the Claude Max 20x $200/mo-vs-$8,000/mo API-equivalent-usage figure. The underlying SemiAnalysis publication was not independently located and read in full during this research pass; the figure is JPM-reproduced-from-SemiAnalysis, flagged pending direct verification against the SemiAnalysis newsletter archive.
[3] SSD Nodes, “Claude Code Pricing in 2026: Every Plan Explained (Pro, Max, API & Teams),” 2026, corroborated by morphllm.com and Anthropic's own Claude Help Center pricing pages. Pro $20/mo, Max 5x $100/mo, Max 20x $200/mo (20x Pro's usage ceiling); API access metered separately, Sonnet 4.6 from $3/MTok input, $15/MTok output. Accessed 2026-07-02.
[4] SemiAnalysis (via Digg, TechSpot, ZME Science, and Cybernews aggregation, June 2026) — ChatGPT Pro $200/month-vs-$14,000/month-at-full-utilization finding, using the same purchased-subscription/run-to-ceiling/price-at-API-rates methodology. The original SemiAnalysis article was not directly fetched in this pass; all detail here is via secondary aggregation of the same underlying research, flagged for direct-source verification before publication.
[5] SpaceDaily, aggregated report on Microsoft's internal Claude Code license cancellation and the Uber AI-budget claim, June 2026. Confirms the license cancellation (corroborated independently by source [7]); the claim that Uber burned its entire 2026 AI-tools budget on Claude Code and Cursor in four months is sourced only to this secondary aggregator, with no primary corroboration (no Uber statement, no first-tier outlet) located — flagged as needs-verification. Accessed 2026-07-02.
[6] PitchBook, “Anthropic's gross margin is the most important number in tech,” June 2026, corroborated by Morningstar's republication. Source for Anthropic's ~44% overall gross margin and the $0.71/$0.56 Q1/Q2 2026 compute-cost-per-revenue-dollar figures. Accessed 2026-07-02.
[7] Forbes (Jon Markman), “Microsoft Ends Claude Code Licenses As It Shifts Developers To Copilot,” June 1, 2026. Confirms: Claude Code released May 2025; Microsoft engineers using it since Dec 2025 “on company dime”; wind-down deadline June 30, 2026; the direct quote “renting intelligence by the token is a price Microsoft does not control.” Does not disclose exact dollar overrun or engineer headcount. Accessed 2026-07-02.
[8] GitHub (official blog), “GitHub Copilot is moving to usage-based billing,” 2026. Confirms June 1, 2026 effective date, AI Credits mechanism (1 credit = $0.01), headline seat prices unchanged ($10 Pro / $39 Pro+ / $19 Business). Accessed 2026-07-02.
[9] Tom's Hardware, “Github Copilot customers report up to 100-fold price hikes — AI sticker shock bites as Microsoft switches to usage-based pricing,” June 3, 2026. Source for the $39→$800+ real-bill examples and the “up to 100-fold” headline figure; the same article JPM cites in source [1]. Accessed 2026-07-02.
[10] artificialintelligence-news.com, corroborating coverage of the Copilot usage-based billing rollout and user price-hike reports, June 2026. Accessed 2026-07-02.
[11] Anthropic, Series G funding announcement, “Anthropic raises $30 billion in Series G funding at $380 billion post-money valuation,” February 2026. Primary source for Claude Code's $2.5B+ run-rate revenue, quadrupled business subscriptions since January 2026, and enterprise users representing more than half of Claude Code revenue. Accessed 2026-07-02.
[12] OpenCode (opencode.ai) and corroborating comparison coverage (Builder.io, “OpenCode vs Claude Code”; Firecrawl; morphllm.com). GitHub star counts (~180,500 vs. ~135,000), contributor/commit counts, and the 7.5M-monthly-developers figure are drawn from these secondary trackers rather than a direct GitHub API pull in this research pass — directionally consistent across sources but not independently re-verified; flagged. Accessed 2026-07-02.
[13] TechCrunch, “Sources: Cursor in talks to raise $2B+ at $50B valuation as enterprise growth surges,” April 17, 2026, corroborated by TechFundingNews and Value Add VC coverage of the same round and the preceding November 2025 $29.3B Series D. Accessed 2026-07-02.
[14] TechCrunch, “SpaceX to acquire Cursor for $60B in stock, days after blockbuster IPO,” June 16, 2026; CNBC, “SpaceX to acquire the AI coding startup Cursor for $60 billion,” June 16, 2026. ARR trajectory (~$100M early 2025 → ~$4B June 2026) per Forbes (source [28]). Accessed 2026-07-02.
[15] Aggregated coverage (tech-insider.org; TradingKey; Market Clarity, “Is Cursor Profitable Today?”) on Cursor's April 2026 gross-margin-positive status at the enterprise-account level, its proprietary Composer model (launched Nov 2025) handling roughly half of autocomplete requests, and routing to cheaper third-party models like Kimi. As a private company, Anysphere discloses no financials directly — these are analyst/aggregator syntheses, flagged as estimate-grade. Accessed 2026-07-02.
[16] Devin/Cognition official pricing page, corroborated by third-party trackers (Pensero, Brainroad, EasyClaw): $20/mo Core ($2.25/ACU), $500/mo Team (250 ACUs at $2.00/ACU), custom Enterprise; 1 ACU ≈ 15 minutes of autonomous work. Accessed 2026-07-02.
[17] Fortune, TechCrunch, and Elephas timeline coverage of the Windsurf saga (OpenAI's $3B deal collapsing, Google's $2.4B reverse-acquihire, Cognition's 72-hour acquisition of remaining assets), July 2025. Accessed 2026-07-02.
[18] TechCrunch, “AI coding startup Cognition raises $1B at $25B pre-money valuation,” May 27, 2026, corroborated by TechFundingNews' $492M ARR and 89%-self-written-code figures. Accessed 2026-07-02.
[19] CNBC, “Cognition valued at $10.2 billion two months after Windsurf purchase,” September 8, 2025. Accessed 2026-07-02.
[20] LangChain (company blog), “LangChain raises $125M to build the platform for agent engineering,” October 2025, and TechCrunch coverage of the same round. Primary source for the $125M/$1.25B figures, 90M-monthly-downloads, and 35%-Fortune-500-adoption claims. The ~$16M ARR figure is from Getlatka, a third-party revenue-estimate aggregator — not company-disclosed, explicitly flagged as unverified. Accessed 2026-07-02.
[21] Prime Intellect (company blog/case study), “How Ramp Used RL to Beat Frontier Models at Spreadsheet Search,” 2026. Primary source: Qwen3.5-35B-A3B base “FastAsk” model beat Opus 4.6 by 4+ points exact-match on a held-out retrieval eval, running ~27% faster and at Haiku-level cost. Corroborated by Ramp's own Builders blog post (cited via JPM, source [1], but not independently re-fetched in this pass — flagged as corroborated-but-not-directly-read). Accessed 2026-07-02.
[22] Nvidia Developer Blog, “NVIDIA Nemotron 3 Ultra Powers Faster, More Efficient Reasoning for Long-Running Agents,” 2026, and Harvey/Trajectory Labs' own announcement and product materials on the Nemotron 3 Ultra legal post-training result. The “under 24 hours” and “no new engineering” claims are Trajectory Labs' own characterization, cross-checked against Nvidia's Nemotron 3 Ultra technical specs but not independently verified against a third-party benchmark of the Harvey-specific result — flagged. Accessed 2026-07-02.
[23] Lindy AI (company blog), “Migrating from Claude to DeepSeek,” June 24, 2026. Primary source, company's own account: ~90% inference-cost reduction on migrated routes, Sonnet→DeepSeek V4 Flash on Atlas Cloud, Kimi K2.5/GLM 5.1/Kimi K2.6 tested and rejected on production “vibe” despite offline-eval promise, Claude Sonnet retained for the explicit higher-intelligence path. Accessed 2026-07-02.
[24] the-decoder.com, “AI startup Lindy ditched Claude entirely for Deepseek, saving millions as cost pressure mounts on Anthropic,” June 2026. Corroborating secondary coverage with direct CEO (Flo Crivello) quotes: AI costs had exceeded personnel costs at the 25-person company; “It's a matter of survival for the business.” Accessed 2026-07-02.
[25] Brian Armstrong (@brian_armstrong), X/Twitter post, June 8, 2026. Primary source, direct quote: “80% of workloads will be running on 99% cheaper models within 12-18 months — 20% of workloads will still run on latest gen models where IQ maxing is important.” Corroborated by CryptoRank.io and Benzinga coverage. Accessed 2026-07-02.
[26] TechCrunch, “SpaceX to acquire Cursor for $60B in stock, days after blockbuster IPO,” June 16, 2026. Confirms SpaceX's prior $1.25T xAI merger (Feb 2026) and its June 12, 2026 IPO (reported as the largest in history), and the stated rationale of putting Grok in front of Cursor's developer base. Accessed 2026-07-02.
[27] CNBC, “SpaceX to acquire the AI coding startup Cursor for $60 billion,” June 16, 2026. Corroborates deal structure: pre-committed in April 2026, ahead of SpaceX's IPO, as either a $60B stock acquisition or a $10B break-up fee if it fell through; targeted close Q3 2026; no public commitment as of research date to Grok-exclusivity or continued multi-provider routing post-close. Accessed 2026-07-02.
[28] Forbes (Rashi Shrivastava), “SpaceX's $60 Billion Cursor Acquisition Doubles 20-Something Cofounders' Net Worths,” June 16, 2026. Cofounder ages (Truell 25, Sanger 25, Asif 25, Lunnemark 26) and ~$2.7B each net-worth estimate; ARR trajectory ~$100M (early 2025) → ~$4B (June 2026). Accessed 2026-07-02.
A note on method. This layer moves faster than any other rung in the stack — the Microsoft Claude Code cancellation, GitHub Copilot's billing switch, and the SpaceX–Cursor deal all landed inside a single quarter (April–June 2026), which is itself part of the story: the harness's economics are still being renegotiated in real time. Per the founder's sourcing directive for this series, figures attributed directly to Michael Cembalest's J.P. Morgan reports are treated as accurate and attributed without further hedging; non-JPM secondary figures — principally the SemiAnalysis $200-vs-$8,000 and $200-vs-$14,000 arbitrage numbers, and the Uber four-month-budget claim — keep their flags because they were not read from a directly fetched primary source in this research pass. Undiscoverable or company-only figures (LangChain's ARR, Cursor's margin structure, the Harvey/Nemotron benchmark) are stated plainly with their sourcing limitation noted, not hidden. The argument does not rest on any single contested number; it rests on the structural fact that a flat-rate seat and a metered backend can diverge by an order of magnitude, that both sides of that gap started repricing in the same quarter, and that the escape hatch this creates is, as of this writing, in the process of being bought by the very economy it was built to route around.