OpenAI & SpaceX S1 Drops | Anthropic Laps OpenAI at $44B ARR | Nvidia Prints $81.6B | Layoffs at Cloudflare & ClickUp

Key points

• **Anthropic laps OpenAI: $44B ARR, and the margin trajectory is the real story.** Rory: *'gross margins expanded from 38% to 70% and they're going to have a $559 million operating profit in Q2 projected.'* The path: negative-60% gross margin (FY24 prior) → +34% → 70%. *'High growth and improving margins meant profits were inevitable.'* **Direct continuation of [Gavin Baker's Anthropic-EBIT-positive framing in Issue 06](/issues/2026-05-24)** — now with the specific margin ramp. Asterisk: some of the profit may be a temporary SpaceX/Colossus discount, so the headline may flatter the trend.
• **The Anthropic bear case — the premium may not be defensible.** *'Claude is twice the price of its competitor.'* If even half the market decides the ROI isn't there for a premium product, Anthropic either cuts price to match or holds premium like Apple but cedes share. The degradation mechanism: *'folks will just use a combination of GPT-4, it's older, and some Sonnet and some DeepSeek because that's how I'm going to spend my budget'* — token-budget optimisation erodes pricing power. **The cleanest articulated short-thesis against the Anthropic-dominance narrative this week.**
• **The ROI reckoning is the week's master theme: experimental spend is about to meet the green eyeshades.** *'When you go from experimental to we spent $300 million on tokens... you get some lift in efficiency, but that means $200 million of OPEX you didn't expect, and you have to terminate a bunch of people — at that point HR comes in.'* The Uber COO data point: a full year's Anthropic credits spent in four months, gains *'probably there but not measurable.'* **Triangulates exactly with [the Fortune-20 $200M-tokens-minimal-results anecdote in this week's All-In](/issues/2026-05-31)** — two independent sources, same enterprise-unwind signal.
• **Nvidia is 'the most profitable company on the planet' and the market is rationally calm.** *'$50 billion plus, 56 billion of profits... if that continued all year that's $200 billion of profit a year'* vs Google's $100-120B. $81.6B revenue (+80%), $91B Q2 guide, $80B buyback. Why the stock barely moved: *'the stock moves on the Delta news, not the total news. They beat slightly, it's gone up slightly.'* Framing: *'Nvidia is 7% of all Americans' life savings — we're all AI investors whether we realise it or not.'* **Continues [the Nvidia-Q1 print from Issue 06 All-In](/issues/2026-05-24)** with the profitability lens foregrounded.
• **Jensen's $3-4T-by-2030 capex call sets up the $64-trillion question.** Quick math from the panel: $3T capex × 50% GPU × 70% Nvidia share ≈ **$1T of Nvidia revenue** (from ~$320B run-rate now). *'Extrapolating the growth rate of the last three or four years gets you there. But the question is, does that growth continue or do you hit economic — not technical — constraints? There has to be an ROI on that next $2 trillion.'* **The single highest-stakes macro question for the entire AI-infra cohort.**
• **Valuation skepticism is loud and named.** On the SpaceX/OpenAI S1s: *'100 times trailing sales, my God... it's SolarCity on steroids... it could be the GeoCities deal of the AI era... I love the S1 but it's all madness, I wouldn't buy a share.'* The constructive counter: *'the picks and shovels of the agentic revolution is just a good place to be investing.'* **A useful sentiment-gauge: even AI bulls are flinching at the IPO-cohort multiples** — pairs with [Gerstner's '$86B of supply vs $100T depth' framing in this week's CNBC hit](/issues/2026-05-31).
• **The layoffs debate: AI efficiency or COVID over-hiring? Panel says efficiency.** Cloudflare + ClickUp layoffs. Andreessen's 'this is just COVID over-hiring unwinding' take is dismissed: *'the dumbest take I've seen since our last show... this is not over-hiring, natural attrition is 15-25% a year.'* The structural read: once AI spend eats meaningful chunks of the wage bill, the layoff/ROI conversation *'becomes more quantitative — you have to prove it'* before terminating people. **Connects the AI-job-loss narrative to the same ROI-measurability gate running through All-In.**
• **Bifurcation thesis: the willingness to spend tracks ideological commitment to AI.** *'I can't remember except maybe a year in the mid-90s internet where corporate America was so convinced of the ROI of something. The sluice gates are open, the money is being authorized.'* But the next shoe: *'okay, that was fun, we spent $10 billion, what did we get?'* The world bifurcates between AI-native digital businesses (spend justified by token-vs-engineering-savings, e.g. DoorDash) and static-product incumbents (Uber) where the gains aren't measurable.

Notable quotes

Themes

• AI IPO supply wave & valuations
• Anthropic dominance vs premium-price bear case
• ROI-on-AI-spend reckoning
• Nvidia profitability & capex extrapolation
• AI layoffs & efficiency measurability