ai usage costs get real

Ok it's starting to happen: ai usage costs getting real.

Github copilot announced changes to their pricing models to be more usage based, starting in 1 month.
"Instead of counting premium requests, every Copilot plan will include a monthly allotment of GitHub AI Credits, with the option for paid plans to purchase additional usage. Usage will be calculated based on token consumption, including input, output, and cached tokens, using the listed API rates for each model."

Uber reportedly spent their whole 2026 budget for Cursor and Claude code in the 1st four months of 2026. Maybe that's good. Maybe not. Its safe to say CFOs across orgs are paying attention. Burning tokens for no valuable outcome (tokenmaxxing) is not the right strategy (despite when ai vendors might tell you). The outcome has to be worth the investment.

OpenAI and Anthropic are losing billions. OpenAI internal docs say they will lose $14 billion in 2026 and will lose $44bn until they turn profitable in 2029 (wow!). AI usage is being massively subsidized. 

OpenAI and Anthropic are following a well worn strategy: focus on growth then pivot later to profit. Uber, Netflix, Amazon have all succeeded with it. Have you seen Airbnb service charges lately? 
Both OpenAI and Anthropic plan to IPO in 2026. I anticipate the free lunch will continue until then. 

But soon financial gravity will reassert itself. Shareholders will demand returns and costs will have to go up bigly.

By this point we may all be using free/cheap "good enough" ai and not from OpenAI, Anthropic or MSFT.






Comments

Post a Comment

Popular posts from this blog

angular js protractor e2e cheatsheet

Many from the protractor test suite category detailed task syntax example load page load a page browser.get('yoururl') browser btn actions browser back/forward browser.navigate().back() OR browser.navigate().forward() access elements on page See Protractor API list and examples from Protractors test suite access by angular ngmodel name (must match ngmodel in the html) this is the preferred approach element(by.model('person.name')); access by the css id on the page e.g, #user_name (but this can be brittle) element(by.id('user_name')) access angular binding to get generated value i.e. for {{}} element(by.binding('person.concatName')); access angular repeat list data (must match repeat stmt in the html) See Protractor test suit l...
Read more

angularjs ui-router query string parameter support

angularjs ui-router is pretty cool, we use it in our app now but have not yet used query string parameters (e.g. ?p1=age&p2=height). But now we need to. we're now adding a feature which is basically a report generator and its a good use case for query string parameters once on the page then you're basically in that ui-router "state" users can select filters etc. from dropdowns which will reload the data but should not destroy/recreate any controllers or ui widgets (we stay on the page and do not state transition) yet we do want to update the url for deep links and back button support we don't know the complete list of possible filter choices, but know it should be data driven and will change by report  urls to support this report feature are will be something like thus /reports  /reports/sales  /reports/sales?region=west&time=weeks so we need to use query string parameters Below are some of the things I discovered with ui-router query strin...
Read more

Notes on interview with Kent Beck about AI and why it's like a "genie"

TDD, AI Agents and coding with Kent Beck https://www.youtube.com/watch?v=aSXaxOdVtAQ This podcast was really fun to learn from and I recommend. Kent is famous in the programming world, so his take on AI is worth learning from. Kent says he's spending 6-10 hours a day programming and in 50 years of programming it's the most fun he's had, "its a blast". Kents using agentic tools where it generates a bunch of code from a single prompt. Kent says nobody knows what works so try it all. So we should be experimenting.  Kents tried a number of different languages as he's experimented. Kent thinks of AI as an "unpredictable genie". Its a genie in that "it grants wishes and you wish for something and you get it, but it's not what you actually wanted: it volunteers features, doesn't care about design, willing to cheat (delete/modify tests) there can be some good things with this, but some bad I love this metaphor. I have experienced these aspects o...
Read more