Your AI tool just filed for a $965B IPO.

Tried to explain to my accountant this week that the AI tool I pay £20 a month for is now worth nearly a trillion dollars. He did not find this as funny as I did. But he did ask if the price was going up.

On June 1, Anthropic confidentially filed an S-1 with the SEC — the first formal step toward a public stock listing. The timing wasn't subtle. Three days earlier, the company closed a $65 billion Series H at a $965 billion valuation, overtaking OpenAI to become the most valuable private AI company in the world. Revenue run rate hit $47 billion in May, up from $10 billion last year. That's an 80-fold increase in annualised revenue over 18 months.

For a UK SMB paying £20 a month for Claude Pro, this is legitimately interesting news. Not because you should care about Anthropic's IPO — you probably shouldn't — but because a public listing changes the dynamics. Right now, Anthropic can price Claude to grow. Once shareholders are involved, that calculus shifts. According to Fortune, the company is also on pace for its first profitable quarter, which suggests the subsided-pricing era may not last forever.

Worth noting: Anthropic is paying SpaceX $1.25 billion per month for compute through 2029. Someone is paying for those GPUs. Today it's venture capital. After an IPO, it might be you — in the form of higher subscription fees. No panic required, but it's worth locking in annual plans if you haven't already.

The British Chambers of Commerce says 54% of UK SMEs now use AI. That sounds like a lot. It isn't. The same data shows only 11% use it extensively enough to automate actual operations. The other 43% are asking Claude to fix a typo and calling it AI adoption.

This isn't a dig. It's just what early adoption looks like. Most UK businesses are still in the 'poke it with a stick' phase — which means the gap between you and your competitors isn't about who has access to Claude. Everyone has access to Claude. It's about who has built something that runs without them.

The BCC will keep publishing 'adoption is up' headlines. The more honest story is that six in seven UK businesses still aren't getting measurable returns. The tool isn't the differentiator anymore. The workflow is.

Opus 4.8 launched on May 28 with effort control — a slider that lets you tell Claude how hard to think. Low effort for quick tasks, high for complex ones. It's available now on Pro and Team plans alongside the model selector. Here's where it actually saves you time and money.

• First draft mode (Low effort) — Writing a quick client update or social post? Set effort to Low. Claude writes faster, uses fewer tokens, and for this kind of output it's more than good enough. Save the horsepower for things that need it.

• Proposal review (High effort) — Before sending a proposal, paste the whole thing in with effort set to High and ask Claude to find logical gaps, missing information, and anything a sceptical buyer would push back on. This is where the extra thinking actually earns its place.

• Process audit (Medium effort) — Describe a workflow you do every week — reporting, client check-ins, onboarding — and ask Claude to identify the three steps that are most worth automating. Medium effort is the sweet spot: enough reasoning to give you something useful, not overkill for a 10-minute conversation.

The S-1 filing dropped Monday and the Opus 4.8 launch hit the week before. Between them, the timeline had plenty to say.

You've got a new business pitch to write. You know what you want to say, roughly. But every time you open a blank doc you write one sentence, hate it, and go make a coffee instead. The pitch is due Thursday. It's Tuesday afternoon.

This is exactly where effort control changes the game — not because Claude writes better copy than you, but because it removes the blank page problem entirely. Set effort to Low (faster, cheaper, no overthinking), paste in your rough notes, and ask for a messy first draft. Not a polished one. Messy. Something to argue with.

You'll hate parts of it. Good. That's the point. Now you're editing, not writing, which is ten times easier. When you've got a structure you like, switch to High effort and ask Claude to find the weakest section and make the case for cutting it. It will. You probably won't. But at least you'll know which part of the pitch you need to strengthen before Thursday.

Two hours saved. One coffee fewer. The pitch goes out on time.

Re-reading the original iPhone launch coverage. Everyone got the 'what' right and the 'so what' completely wrong. Feels relevant this week.