In partnership with
News of the day
1. Meta enters AI API market with Muse Spark 1.1, undercutting OpenAI and Anthropic significantly and intensifying the AI price war. → Read more
2. SpaceXAI and Cursor launch Grok 4.5, a powerful, fast, and efficient AI model that rivals frontier pricing and performance. → Read more
3. 1X unveils Neo robot's advanced five-finger hands, boasting human-like dexterity and rapid movement for domestic tasks. → Read more
4. Ollama, a popular open-source AI tool for developers, has secured $65M in Series B funding, bringing its total to $88M and reaching nearly 9M users. → Read more
Our take
Hi Dotikers!
Yesterday we watched the demand side lose its loyalty: Coinbase and Vercel refusing to marry a single lab, routing every task to the cheapest model that does the job. Today the supply side answers, and it answers with a hammer. Meta just walked into the API business, and it walked in through the price.
The model is Muse Spark 1.1, a multimodal reasoning system built for agents, coding and computer use, with a one-million-token context window. On benchmarks it trades blows with the frontier, leading four of twelve tests, a hair behind Anthropic's Opus 4.8. Respectable, but not the headline.
The headline is the price tag. Meta charges 4.25 dollars per million output tokens. Anthropic and OpenAI ask between 25 and 50 for the same thing, roughly six to twelve times more. Meta even undercuts xAI's Grok 4.5, which launched the day before and held the crown of cheapest near-frontier model for a grand total of a few hours. The price war is now measured in hours, not quarters.
Here is why it stings. Meta makes more than 60 billion dollars in profit a year, from ads, not from tokens. Its API does not need to earn a cent any time soon, it is a gateway into an ecosystem. Google sits in the same seat. OpenAI and Anthropic do not: they sell intelligence and little else, and they burn billions doing it. They need fat token margins to survive. Their rivals can price those margins to zero and barely notice.
So the squeeze closes from three sides at once. Customers who no longer pick a favorite. Conglomerates pressing down from above with balance sheets that dwarf a startup's. Chinese open models pressing up from below at rock-bottom prices. And notice the quiet irony: Meta, once the open-source hero with Llama, ships Muse Spark closed. The champion of openness now fights the margin war with a proprietary, cut-price API.
For those of us who build, yesterday's lesson only hardens: do not marry a model, since the cheapest near-frontier option now changes between morning and afternoon. But cheap tokens are not cheap results. A model that burns more tokens per task, or stumbles in production, can erase the entire discount. You do not remove the hard problem, you relocate it, this time from the invoice to the engineering. The price of raw intelligence is sliding toward the cost of the electricity behind it, and the ones with the most to lose are those who sell nothing else.
M.
How owning AI deployment expands your career
Across product, ops, and CX teams, a new kind of role is taking shape: the person responsible for making AI actually work, day to day.
On July 16, three people living this shift join a live roundtable: Simone Santiago Broad (Yoco), Yelva Espinoza (Zumba Fitness), and Fin's Dave Lynch. You'll hear what the job really looks like across industries, how they carved out these roles, the skills they'd hire for, and the challenges they're tackling now. Bring your questions, since the best moments happen live.
Register for the roundtable to save your spot.
Meme of the day
