Anthropic passes OpenAI on enterprise revenue.

For the first time since the category existed, the lab perceived as the underdog is winning the enterprise stack — at lower cost, with a smaller burn, and a tighter Claude Code wedge. The capex/revenue gap, meanwhile, just hit 10:1.

Frontier LabsCapexBubble RiskCoding Wedge

ANTHROPIC ARR

$30B

from $1B in Jan '25

ENTERPRISE SHARE

40%

vs OpenAI 27%

CAPEX : REVENUE

10 : 1

$725B vs $51B

CLAUDE CODE ARR

$2.5B

single-product wedge

FEATURED // ANTHROPIC vs OPENAI — MONTHLY REVENUE RUN-RATE ($B)

JAN '25 → APR '26

OPEN IN /LABS

CROSSOVER: Mar '26 — Anthropic $19B vs OpenAI $18B run-rate.

DELTA: +$6B in 30 days, the largest monthly Δ in the dataset.

SOURCE: Sacra · The Information · primary disclosures.

EXECUTIVE SUMMARY

The story of the last 30 days is not that Anthropic has won — it's that the assumption of an OpenAI monopoly on enterprise frontier has cleanly broken. Run-rate parity in March became a +$6B lead in April, against a roughly 4× lower training spend and a tighter, code-anchored product wedge (25% premium).

Underneath, the macro picture is unchanged and increasingly uncomfortable. Hyperscalers will spend $725B on infrastructure this year against perhaps $51B of direct AI revenue — a 10 : 1 gap that Goldman now calls "elongation." Power has replaced GPUs as the binding constraint. 95% of GenAI pilots still fail to reach production.

For operators, the operative question this week is no longer "which model" but "which wedge." Claude Code is the proof. The next $1B of enterprise AI revenue does not accrue to whoever ships the highest Elo — it accrues to whoever owns a workflow narrow enough to escape the pilot graveyard.

WHAT TO WATCH // 7 DAYS

T+1d
Anthropic $50B round close
≈$900B post-money, per FT
T+2d
GPT-5.5 Pro pricing reset
OpenAI counter-move likely
T+3d
OpenAI DevDay '26
Workspace Agents GA
T+4d
Q1 hyperscaler capex prints
MSFT / GOOGL / AMZN / META
T+5d
EU AI Act Article 51 deadline
GPAI compliance

DRILL DEEPER // OPEN MODULES

FRONTIER LABS

See Anthropic vs OpenAI head-to-head: Elo, ARR, valuation, enterprise share.

→ FOCUS: HEAD-TO-HEAD

INFRASTRUCTURE

Trace the $725B capex flow and the 10:1 gap by hyperscaler.

→ FOCUS: CAPEX-GAP

ENTERPRISE SAAS

Drill the spend stack: $4B coding, $2.5B Claude Code, NRR by archetype.

→ FOCUS: SPEND-STACK

AGENT LAYER

Where Anthropic's wedge converts: MCP at 78% adoption, Claude in the loop.

→ FOCUS: MCP-GAUGE

FORECASTS

Project Anthropic's trajectory to 2030 with the timeline scrubber.

→ FOCUS: HORIZON

RISKS

Capex bubble (78), model commoditisation (55) — the counter-narrative.

→ FOCUS: RADAR

GEO LANDSCAPE

Where US frontier ARR is booked vs where Chinese open-weights win share.

→ FOCUS: COUNTRY-REGISTRY

WHITESPACE

What Anthropic is NOT building: vertical AI, blue-collar, agent identity.

→ FOCUS: MATRIX

STANDING CONTEXT // 8 THINGS THAT MATTER MOST

REPORTING BASELINE — MAY 2026

1. Anthropic overtook OpenAI on enterprise revenue
$30B+ run-rate by April 2026 (from $1B Jan '25). 40% enterprise spend share vs OpenAI's 27%. Claude Code alone at $2.5B ARR.
2. The model layer is bifurcating
Frontier reasoning (GPT-5.5, Opus 4.7, Gemini 3 Pro, Grok 4.1) stays expensive & closed. Open-weights bench (DeepSeek V4, Qwen 3.6, Kimi K2.6, Llama 4) matches them at ~10× lower cost.
3. Coding is the only category with proven unit economics
55% of departmental AI spend. Labs now compete with their own customers (Cursor, Cognition). Legal next: Harvey $190M ARR; Legora fastest-ever to $100M.
4. Sovereign AI is the defining geopolitical force
UAE, Saudi, India, Korea, Japan, FR, UK, EU, Brazil all building national capacity. Only China attempts full-stack independence — most 'sovereign' efforts run on NVIDIA + US clouds.
5. Agent layer is real but earlier than the marketing
Agentforce $1.4B ARR. MCP at 78% enterprise adoption. But only 31% of orgs have an agent in prod. 88% of agent pilots stall.
6. Capex has decoupled from revenue
$725B infra spend vs ~$51B direct AI revenue in 2026. Goldman calls it 'elongation' — buildouts stretch, gap widens, no one blinks.
7. Pricing is genuinely unsettled
Salesforce runs 3 models in parallel: per-seat, per-conversation ($2), Flex Credits. Intercom Fin's $0.99-per-resolution is the new primitive. IDC: 70% of vendors refactor away from per-seat by 2028.
8. The largest underbuilt areas aren't where capital is
Vertical AI for regulated industries, blue-collar/field work, non-English markets, agent identity & governance, AI-native services that sell completed work.

PRIOR BRIEFS // ARCHIVE

W18DeepSeek V4 Pro lands at MIT license — and on benchmark parity.ARCHIVED
W17Salesforce kills per-seat for Agentforce. The pricing primitive shifts.ARCHIVED
W16xAI / SpaceX close. Memphis Colossus headed to 1.5 GW.ARCHIVED
W15MIT: 95% of GenAI pilots still fail. The pilot-to-prod chasm widens.ARCHIVED