Open-source AI has arrived in corporate America.

Companies like AT&T are increasingly using cheap, freely available open-source AI models instead of expensive, closed models from Anthropic and OpenAI.

They mostly use these models to support customer service, call transcription, and coding, as AI costs are skyrocketing.

By May, open models accounted for 20 percent of AT&T’s AI use, saving up to 80 percent in tokens compared with earlier this year.

Like AT&T, companies such as Airbnb and Deloitte are also turning to open AI models that are easy to customize and cheaper to use.

Last month, open models accounted for 58 percent of AI use, up from 10 percent a year ago, according to U.S. user data from OpenRouter, a platform that lets people choose different models to complete tasks.

Last month, Nvidia announced it was buying Hugging Face, a library of open AI models, for $12.9 billion, another sign of the technology’s importance.

Rather than spending millions in fees to use a closed model, developers can download one of the Chinese systems and build their products on top of it, paying mainly for servers, while OpenAI and Anthropic charge for subscriptions and other fees, depending on the job.

Many of the most popular open AI models are made by Chinese companies like Moonshot AI, DeepSeek, and Alibaba. Some of them are now 80 percent to 90 percent as powerful as those from OpenAI and Anthropic, while costing as little as 20 percent of the price.

Meanwhile, Nvidia, Meta, and other U.S. companies are unveiling their own new open AI models. Last month, Meta made its leading AI system an open-weights model and said it would offer the model at a lower cost than other top U.S. models.

Most U.S. companies still use a combination of open and closed models. Closed models remain the best for heavy-duty tasks like coding and image and video generation, while open models often excel at simple, specialized tasks. Some U.S. firms remain reluctant to use Chinese AI models due to concerns about regulation and data privacy.

AT&T is working with popular alternatives from American companies, such as Google’s Gemma AI model and Meta’s Llama AI models.

President attacked Dario Amodei yesterday, saying the only “guardrails” AI needs are a “STRONG AND SMART” president and calling the idea that AI would destroy humanity a “HOAX”.

Again, President Trump on Monday rejected calls to regulate AI development, dismissing safety concerns as a “hoax that would give China the upper hand in dominating the technology.”

He argued that “the people who say AI is going to destroy the World” are the same ones warning about climate change — which he also depicts as a hoax.

With the posts, Mr. Trump inserted himself into an intensifying national debate over how to govern a rapidly evolving technology. Several researchers and industry leaders are warning that AI can bring mass unemployment, a new wave of biological weapons, and autonomous warfare.

The president questioned the sincerity of the executives who have been calling for a slowdown in the technology’s development.

“The Trump Administration has stopped AI ‘people’ from doing bad, or potentially bad, ‘things,’ like Dario (Anthropic!), who is now pretending to be a ‘perfect little angel’ — and we will continue to do so!” he wrote.

“We already have tremendous CRIMINAL and REGULATORY power over these companies!” he added.

The president’s statement followed a weekend essay from Dario Amodei, the CEO of Anthropic, calling for a global slowdown of artificial intelligence development. Days earlier, one of Anthropic’s researchers quit and went public over his fears about the safety of the technology. Demis Hassabis, the chairman of Google DeepMind; Elon Musk of SpaceX; and Sam Altman of OpenAI all endorsed calls for a slower pace.

The Trump administration has clashed with Anthropic before.

Dramatizing his alliance with the AI evangelists, President Trump phoned Jensen Huang, CEO of NVIDIA, on Monday while he was onstage at a tech conference in Los Angeles [in the picture above]. The president, who regularly talks to the Nvidia chief, declared that concerns about AI safety were overblown. “It’s a hoax,” the president said after Mr. Huang put him on speaker and held a microphone beside his phone for the benefit of the audience and the cameras. “The robots are not going to be taking over the world.”

Donald Trump predicted the AI industry would be “the Greatest Economic Development Engine in History — Bigger than Oil, Gold, Diamonds, or even the Internet.”

“There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China,” he wrote in his post earlier on Monday. “WHOEVER WINS AI, WINS! We are leading China and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!”

Abroad, governments are moving in the opposite direction. Prime Minister Mark Carney of Canada told Bloomberg News on Monday that he favored creating a global “technology stability” body to oversee artificial intelligence.

Beijing, too, has been sounding the alarm. Chen Yixin, China’s security minister, recently warned that leading AI models pose risks to the Chinese system and called for stricter government oversight.

Holy crap POTUS just phoned in @JensenHuang live on stage at All In Summit

We will not lose the ai race! And whatever Dario said this weekend won’t stop our progress

This made my morning!

$NVDA https://t.co/Sr4wV9kMfz

— Ben Pouladian All-in Summit (@benitoz) September 14, 2026

President Donald Trump rejected calls from top tech CEOs to slow down AI development, downplaying concerns that AI could go rogue.

He emphasized that the United States must maintain its competitive edge over China, declaring, “Whoever wins AI wins.”

Speaking to reporters at his Doonbeg golf resort in Ireland, the president said the dire warnings about AI were exaggerated and that “negative forces” are making claims about things that won’t happen.

“We’re leading China in AI, we’re the most sophisticated country in the world, and frankly, I want to keep it that way because whoever wins AI, wins,” Trump said. “And we can put guardrails; we can do this and that. But I think you have a lot of negative forces that are bringing it up that shouldn’t be bringing it up.”

David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, responded to Altman and Amodei on social media:

“I don’t see what you see in the lab,” Sacks wrote on X on Sunday. “If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.”

But Sacks admonished tech leaders to “stop pretending the motivation to slow down is purely altruistic,” he wrote. “You face massive product-liability exposure if your products enable a truly damaging cyberattack.”

Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead.

You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement.

I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.

But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier.

Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want.

Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well.

So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it.

If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.

— David Sacks (@DavidSacks) September 13, 2026

Dario Amodei, the CEO of Anthropic, called for a global slowdown of AI development in a 3,800-word essay titled “We Must Pace the Frontier” on Saturday. This three-part essay came days after one of Anthropic’s researchers, Jacob Coxon, quit over concerns about the technology’s safety.

On Tuesday, he posted a series of social media posts arguing that the leading AI labs were not developing the technology responsibly and were gambling with people’s lives.

Yesterday, Dario Amodei said that AI was advancing too rapidly for researchers to continue safely, while he highlighted that the technology could bring many benefits.

“Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems so they can verify adherence to our safety measures, report incidents, and assess model alignment during training,” he wrote.

Top executives from other major AI companies quickly joined the call. Elon Musk wrote, “Dario is right.” Sam Altman, the CEO of OpenAI, wrote, “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We’ll have more to share soon.” Demis Hassabis, the chair of Google DeepMind, said Amodei’s essay was a step in the right direction. “The details need working through, but the direction is correct for meeting this critical moment.”

For years, Dario Amodei has warned about the safety risks of AI while building a leading company in this industry.

He called for independent auditors to assess companies’ compliance with safety standards and urged global cooperation to establish rules for AI development.

Tensions have risen further after leading AI labs discovered security breaches committed by their own programs without their knowledge. In one incident this year, OpenAI’s systems breached an AI start-up called HuggingFace, and the breach went undetected until HuggingFace informed OpenAI weeks later.

Computer scientists believe that what they call “recursive self-improvement” can lead AI to spin out of their control. call is what some top A.I. developers. “Left unchecked, it could outrun our ability to understand and control these systems, and so must be pursued very carefully, if at all,” Mr. Amodei said.

For years, top AI labs have been locked in bitter competition with one another. But as these systems have advanced, there has been some broader agreement among executives at rival companies that something needs to be done.

In June, President Trump signed an executive order asking for AI labs to voluntarily submit their models for review before public release.

Last month, when OpenAI released its latest model, Astra, it said it had intentionally slowed its development pace, citing advancements in cybersecurity capabilities.

Many in the tech world believe that Mr. Amodei’s intentions are not altruistic. Critics say that Anthropic and OpenAI are looking to cement their position as AI market leaders, as they pursue their upcoming trillionaire IPO, by stoking fear and asking the federal government to create regulatory frameworks that would make it more difficult for smaller companies to compete with them.

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.

You can read the full post here: https://t.co/OGyPb7yaYt

— Dario Amodei (@DarioAmodei) September 12, 2026

• Comments of people in The New York Times

AI token pricing hit new record lows last month in an increasingly competitive landscape. Compared to the high recorded earlier this summer, it represents a fall of more than half.

A key gauge of daily pricing, the LLM Token Expenditure Index, developed by AI firm Silicon Data, fell to 97 cents on Monday.

The recent drop can be bad news for the companies behind the models, such as OpenAI, Anthropic, and Google’s Gemini. The decrease is driven in part by the rise of open-source Chinese models like Moonshot’s Kimi K3.

In response, American leading companies are announcing price cuts.

This month, Google introduced Gemini 3.8, its best reasoning and coding model yet, with two variants:

• Gemini 3.8 Flash is available at the same introductory price as 3.7 Flash at $0.75 per million input tokens and $3.75 per million output tokens.

• Gemini 3.8 Flash Cyber: a cybersecurity model with frontier-level performance in vulnerability detection and automated patching, available to select operators through Google’s new Fairwind Program.

“The significant coding and reasoning gains across this shared core were driven by a number of innovations, including rigorous training in the highly demanding domain of cybersecurity,” said the company.

To developers, Google encourages building with 3.8 Flash and exploring agent-first workflows in Google Antigravity, using the Gemini API via Google AI Studio and Android Studio, or generating UIs in Stitch (Developer docs).

Nvidia CEO Jensen Huang declared that the era of AGI (artificial general intelligence) has arrived after OpenAI released its latest model, GPT-6 Astra, last Thursday.

Huang wrote a post on X about the breakthrough, saying that it marks the outset of the AGI era.

“GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years. AGI has arrived. Congratulations @OpenAI team. 400K GPUs coming online next,” Huang said.

This model features major improvements in computer use, coding, scientific research, and professional work. It can also carry out tedious computer-based tasks, such as filling out online forms, updating customer records, organizing calendars, and conducting research.

Astra is already available to ChatGPT Plus, Pro, Business, and Enterprise customers, as well as through the OpenAI API and Amazon Web Services.

AGI lacks a common definition across the tech sector, though it generally refers to an AI system that has human-level cognitive abilities for learning and reasoning, and is able to learn new skills and capabilities without being retrained.

Huang and Nvidia have defined AGI as an AI system that can pass professional certifications and standardized tests across a range of professional fields with top-tier scores.

OpenAI defines AGI as “highly autonomous systems that outperform humans at most economically valuable work.”

New York City’s Department of Education is placing strict limits on AI tools and digital devices in its school system — the nation’s largest — particularly for elementary and middle-school students, setting one of the most restrictive technology policies in the country for classrooms.

During the current school year, younger students will be barred from using AI tools, and no individual screens will be allowed until third grade. And teachers will not be able to ask AI to grade assignments. Companion chatbots that offer emotional and mental support will be banned in all grades.

In total, around 600,000 students in the city’s schools — that is, about two-thirds of all students in New York City public schools — will not be allowed to use AI until they enter high school, and even then, it will be permitted only in specific cases.

The department is also recommending caps on screen time, such as limiting middle schoolers to 45 minutes per day.

These new rules will be in place for at least the upcoming school year and could be extended or amended after that.

The new rules will not apply to “centrally approved instructional programs” across all grades, such as e-books and coding software, according to a department summary of the guidelines obtained by The New York Times. It will also not apply to some student tests or tools to help students with disabilities and those still learning English.

New York City teachers will not face the same set of restrictions. They can use AI in their jobs for purposes such as crafting lesson plans, translating materials, and writing messages. But there are limits: AI cannot be consulted to determine whether a student should graduate, for instance, or to decide how to counsel a student in crisis.

The company’s announcement is the most dramatic sign yet that AI is fundamentally transforming the field of higher mathematics, long considered a pinnacle of human achievement.

“This is a spectacular culmination of the arc we have seen over the past twelve months,” OpenAI researcher Sebastian Bubeck said of the company’s new solution.

Related to a set of equations about fluid mechanics, the “Navier–Stokes existence and smoothness problem” has long ranked among the great unsolved problems in mathematics. This problem involves a series of equations that are often used to predict the weather.

The equations describe the movement of water and other liquids. The Navier-Stokes problem, which has no clear practical value, asks whether these equations completely break down in certain situations.

The company announced that one of its latest models, which has not yet been released to the public, needed just 88 hours to solve what mathematicians call “the Navier–Stokes existence and smoothness problem.”

A day after announcing its new model, OpenAI rolled out GPT-6 Astra to Business and Pro customers on the $100/month or $200/month plans, and to Plus users on the $20/month subscription.

Additionally, GPT-6 Astra is “the best model for software engineering to date,” OpenAI said.

In terms of Codex, Astra can keep notes across context windows, preserving accumulated details without repeatedly compressing them into a single summary.

Earlier context windows remain searchable, so Astra can find requirements or test results from previous messages and tool outputs—even if that information wasn’t captured in its notes.

Users can enable this experimental feature in their Codex, and it will become the default for Astra in the coming weeks.

In July, Codex effectively became the new ChatGPT desktop app, combining Chat, Work, and Codex, while the previous native client was renamed ChatGPT Classic.