Anthropic closed a $65 billion funding round that values it at 965 billion, making it the most valuable private startup in history and overtaking OpenAI (852 billion) for the first time. The same day it launched its new model, Claude Opus 4.8. The figure is hard to grasp, so a regional reference point helps: Coatlicue, Latin America’s most ambitious sovereign compute project (Mexico’s plan to invest 6 billion pesos over two years), is equivalent to barely 0.5% of this single round (Bloomberg).
There is a detail worth a closer look. Among Anthropic’s new strategic partners are three Asian semiconductor giants (Samsung, SK Hynix and Micron), alongside global institutional funds such as Brookfield and Blackstone. In other words, the company is diversifying its chip and capital base beyond Google and Amazon, and relying on a deeply international network. That raises a legitimate question for the region: when infrastructure, money and models are concentrated at this scale and with these actors, how much bargaining room does Latin America have left, and with whom? With annualized revenue that has already passed $47 billion and a possible IPO projected for between September and November, the answer is no small matter.
Also today
- Claude Opus 4.8 arrives with subagent orchestration and effort control — Anthropic’s new model, announced the same day as the round, promises to coordinate many tasks in parallel and to significantly reduce “overconfidence” when reporting erroneous results.
- Demis Hassabis at Stanford: AI is a “species-level transition” — The CEO of Google DeepMind and Nobel laureate in Chemistry moved his forecast for artificial general intelligence forward to 2029 and called for international coordination over the next 5 to 10 years. He is the third frontier lab leader in two weeks to use language of urgency.
- Global ChatGPT outage: API, DALL-E, Codex, Sora and login affected — A massive outage knocked out much of OpenAI’s tools, with “too many concurrent connections” errors and no officially confirmed cause. A reminder of the fragility of an infrastructure that companies and governments already depend on.
In the region
The big regional story of the day is democracy under digital pressure. Colombia holds presidential elections on May 31 without any regulatory framework for AI in elections, just the week after Brazil approved its artificial intelligence law. Colombia’s Electoral Observation Mission warns that the central challenge is no longer the physical tampering of votes, but the systematic deception of voters through voice cloning and hyperrealistic videos circulating on WhatsApp and social media. In parallel, Bolivia has accumulated 29 pieces of fake AI-generated content about President Rodrigo Paz since May 12 (among them a cloned audio clip announcing a fake resignation), also without specific legislation against deepfakes. It is the first time two countries in the region have simultaneously faced an alert over AI-driven disinformation in the middle of a political moment. For those who want to have a say from outside the big blocs, there is a concrete window: the public consultation on the European Union’s AI Act regarding the classification of high-risk systems is open until June 23.
Launches
- Claude Opus 4.8 — A model geared toward autonomous tasks with two new features: “Dynamic Workflows,” which makes it possible to coordinate many subagents in parallel within a single session, and “Effort Control,” which regulates how much reasoning the model devotes to each task. It keeps the price of the previous version and has been available since day one on claude.ai, the API and GitHub Copilot, which broadens its reach for development teams in the region.
Threads we’re following
The wave of deepfakes in Colombia and Bolivia adds to a pattern we have been observing in Latin America: the technology to fabricate convincing deceptions is already within anyone’s reach, while regulation tends to arrive after the damage is done. Brazil took a step forward with its new law and its electoral court’s anti-deepfake measures; the rest of the region is still watching closely to see how it goes. On Sunday, May 31, with Colombia at the polls, there will be a real test of how resilient an election is when there is no regulatory net to support it.
If building AI at the limit of what is possible now requires figures close to a trillion dollars and alliances that span continents, what role does Latin America want to play in that conversation: spectator, market or partner?
Correction (September 30, 2026). The original version said Mexico planned to invest $6 billion over five years in the Coatlicue project; the correct figure is about 6 billion pesos (about $326 million) over two years of construction, according to El Financiero. At that figure, the project is equivalent to about 0.5% of Anthropic’s round, not 9% as the original version said.
About this entry. It is generated automatically from public sources, without human review before publication. It may contain errors of interpretation or summary; please check each story against its original source (the links lead there) before citing it or making decisions based on it.
Doble Click is written with Anthropic models.