Anthropic's quiet engine: getting API buyers to adopt its apps
Issue 01The Growth IssueVAIRALITYWhat's taking off in AI, and whyOct 26Cover storyAnthropic's quiet engine: getting API buyers to adopt its appsSector watch: pricingLLM prices are collapsing. Spending is still climbing.Breakout: world modelsGenie 4 pushes world models out of the demo reelSector watch: cloudMicrosoft finally puts a dollar figure on Azure: $33.2 billion
Issue 01ContentsContents
04Anthropic's quiet engine: getting API buyers to adopt its appsAnthropic's S-1 shows revenue exploding. The more telling number is how many big customers now buy both the d…
06LLM prices are collapsing. Spending is still climbing.OpenAI and Anthropic have cut prices by as much as 80% in months. Cheaper tokens are real, but the evidence p…
08Genie 4 pushes world models out of the demo reelGoogle DeepMind's new model holds a scene together for two minutes, takes text edits mid-simulation and gener…
10Microsoft finally puts a dollar figure on Azure: $33.2 billionFor the first time, Microsoft reported Azure's quarterly revenue in dollars. The number lets investors compar…
12Wabi's $1.25B bet: your AI agent builds its own appsWabi raised $55 million to turn a chat agent into an app factory. Alongside Thrive's ADT deal and Dream's see…
14Apple Intelligence finally gets a China release dateApple will bring a first set of AI features to Chinese iPhones with iOS 27.3. The slimmed-down launch, built …
16Five small models, one cheap shot at a frontier coderAn open-source workflow lets five modest Qwen models work as a team on hard coding problems. Each run costs u…
02Vairality · Issue 01
The Vairality IndexWeek of Oct 26This week7 stories on what's taking off in AI, and why. The Vairality Index joins the issue once a full week of tracking is in.03Vairality · Issue 01
900%Breakout: AnthropicAnthropic's quiet engine: getting API buyers to adopt its apps04Vairality · Issue 01
BreakoutsAnthropic's quiet engine: getting API buyer…Anthropic's S-1 shows revenue exploding. The more telling number is how many big customers now buy both the developer platform and its apps, a pairing that makes accounts harder to pry loose.The number everyone will quote from Anthropic's S-1 is $16.3 billion: revenue for the first half of 2026, against $1.2 billion in the same period a year earlier. That is more than 13 times higher in twelve months.A smaller number deeper in the filing says more about where the business is heading. More than half of customers spending at least $100,000 a year now use both the Claude Platform and a first-party app such as Claude Code or Cowork. At the end of last year, the share was about 30%.The enterprise base is large and getting top-heavy. As of September 30, Anthropic counted more than 11,900 customers spending $100,000 or more annually, more than 1,900 spending $1 million or more, and about 250 spending $10 million or more.Continue reading →05Vairality · Issue 01
80%Sector watch: pricingLLM prices are collapsing. Spending is still climbing.06Vairality · Issue 01
SectorsLLM prices are collapsing. Spending is stil…OpenAI and Anthropic have cut prices by as much as 80% in months. Cheaper tokens are real, but the evidence points to better models, not lower bills, as what drives usage.The tidy story is that AI adoption has been held back by cost, and that OpenAI and Anthropic are now removing the barrier. The first half of the story is well documented. The second half is only partly supported.Since June, the two labs have repeatedly lowered what they charge for their most popular models. The pattern looks less like a one-off promotion than a price war.OpenAI's small model, GPT-5.6 Luna, launched at $1 per million input tokens and $6 per million output tokens. OpenAI later cut that by 80%, to $0.20 and $1.20.Continue reading →07Vairality · Issue 01
two minutesBreakout: world modelsGenie 4 pushes world models out of the demo reel08Vairality · Issue 01
BreakoutsGenie 4 pushes world models out of the demo…Google DeepMind's new model holds a scene together for two minutes, takes text edits mid-simulation and generates its own sound. Rivals are consolidating around the same idea.World models have spent years as impressive clips with no obvious use. Genie 4, from Google DeepMind, is an attempt to change that. It is a simulation you steer while it runs, and it is now showing up in the places where people go to try new software.It is still a research preview, and Google says plainly that it is early. But the feature list reads less like a lab paper and more like a product spec.Genie 3 arrived as a research preview a little over a year ago. Genie 4 makes four visible jumps.First, it holds a consistent scene for up to two minutes, twice its predecessor's limit. Second, it generates audio natively: ambient sound, music and spoken dialogue come out of the same model that draws the world. DeepMind calls it the first real-time world model to do this.Continue reading →09Vairality · Issue 01
$33.2 billionSector watch: cloudMicrosoft finally puts a dollar figure on Azure: $33.2 billion10Vairality · Issue 01
SectorsMicrosoft finally puts a dollar figure on A…For the first time, Microsoft reported Azure's quarterly revenue in dollars. The number lets investors compare the big clouds directly, and shows how much of the growth rests on a few AI labs.Microsoft has never told the market what Azure earns in dollars. Until this week it offered growth percentages and the occasional vague milestone. On Wednesday it dropped the habit: Azure brought in $33.2 billion in the July-September quarter, up 48% from a year earlier.The timing matters as much as the number. Amazon and Alphabet already report their cloud businesses in dollars every quarter. Microsoft's move puts all three on the same scoreboard, and it comes just as AI demand is deciding who wins the next round of enterprise spending.Azure now makes up roughly 36% of Microsoft's $92.3 billion in quarterly revenue, and Microsoft calls it its largest and fastest-growing product line. It beat expectations on both counts. Analysts had pencilled in $32.3 billion, so Azure landed about $850 million higher. The 48% growth rate also cleared the 44% to 45% that CFO Amy Hood had forecast in July.Continue reading →11Vairality · Issue 01
$1.25 billionBreakout: agentsWabi's $1.25B bet: your AI agent builds its own apps12Vairality · Issue 01
BreakoutsWabi's $1.25B bet: your AI agent builds its…Wabi raised $55 million to turn a chat agent into an app factory. Alongside Thrive's ADT deal and Dream's seed round, it shows where AI money is going.Wabi has raised $55 million at a $1.25 billion valuation. The company is a few months into its second act, and that act is what investors are paying for.It started as a tool that let anyone build an app by typing a prompt. Last month it recast itself as a "personal agent" delivered through an AI messenger. The agent doesn't just answer questions. It builds the small interface you need for the task at hand.The examples the company gives are deliberately ordinary: a meal planner, a tracker for local restaurants, a date-night planner, a fitness plan, a custom trivia game. None of these is worth a standalone app store listing. Together they describe a different product: software that appears when you need it and has no reason to exist beyond that.Continue reading →13Vairality · Issue 01
Breakout: ChinaApple Intelligence finally gets a China release date14Vairality · Issue 01
BreakoutsApple Intelligence finally gets a China rel…Apple will bring a first set of AI features to Chinese iPhones with iOS 27.3. The slimmed-down launch, built around local partners, shows how AI products get adapted for a market.Apple Intelligence is coming to China. The company says a beta of iOS 27.3 arrives on November 4th, with the full release scheduled for December. Apple announced the plan on its Chinese newsroom site, and MacRumors picked it up.The launch is narrower than the one users elsewhere got. The new Siri AI is not part of it. Chinese users will instead get the earlier batch of features: notification summaries, real-time translation, rewriting and proofreading tools, visual intelligence, and the ability to remove objects from photos.The more telling detail is who sits underneath. Apple says it worked with Alibaba to support the Qwen chatbot in some experiences. That looks like a close cousin of the ChatGPT integration iOS added in 2024: a third-party model plugged into the phone's interface, chosen to fit the market.Continue reading →15Vairality · Issue 01
$0.39Small-model teamsFive small models, one cheap shot at a frontier coder16Vairality · Issue 01
What's NextFive small models, one cheap shot at a fron…An open-source workflow lets five modest Qwen models work as a team on hard coding problems. Each run costs under a dollar, and the post claims the team beats a frontier model.The claim, in the title of a Hacker News post: a team of five Qwen3.8-27B models beats a model called Fable 5 on LiveCodeBench Hard. The excerpt we have doesn't include the score table, so treat the head-to-head as the author's claim. The setup behind it is documented, and it is more interesting than the leaderboard line.The roles are a planner, two ideators, a solver and a manager, with a repair worker on call. No agent inherits another's chat history. Each starts clean, reads shared files and writes its output back to disk. That file trail is the only channel between them.Success is not a model's opinion. A script runs the candidate program against sample inputs and outputs and compares the results. If it passes, the team is done.Continue reading →17Vairality · Issue 01
VAIRALITYThe weekly breakout, in your inbox.One email every Sunday: the Index, the biggest breakout, and what to watch next.Subscribe free →vairality.com · Next issue Sunday
Cover
Drag a corner, use the arrow keys, or tap the buttons.
