AI & LLM Trends — 2026-06-18
Daily briefing on artificial intelligence and large language model developments
Big Picture
This week, AI development continues to accelerate across multiple fronts: robotics training via AI coding agents is maturing rapidly, Anthropic revises its Claude Agent SDK pricing structure, and international debate intensifies over American AI infrastructure and governance. Meanwhile, data center grid flexibility emerges as a critical bottleneck for AI scaling.
Top Developments
1. AI coding agents taught robots how to install GPUs and cut zip ties — How to train your robot AI coding agents taught robots how to install GPUs and cut zip ties Nvidia's self-improvement program for robots enlists teams of AI coding agents. Jeremy Hsu – Jun 17, 2026 3:25 pm | 27 Teams of AI coding agents can train robots to do tasks such as inserting GPUs into motherboards. Credit: NVIDIA Teams of AI coding agents can train robots to do tasks such as inserting GPUs into motherboards. Credit: NVIDIA Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Minimize to nav What happens when you give AI coding agents a lab full of robotic arms, some compute resources, and a “generous token budget” for teaching the robots various tasks? The agents can apparently figure out a training regimen that teaches the robots to successfully cut zip ties and even insert GPUs into thin sockets on motherboards. That glimpse into how AI can act in a fully autonomous way to automate robot training was made possible by a new agent harness framework—software that wraps around AI models to enable their use of various tools while also providing capabilities such as memory, context, constraint, and feedback loops. That agentic harness, called ENPIRE , was developed by robotics researchers at the Nvidia GEAR (Generalist Embodied Agent Research) lab alongside collaborators from Carnegie Mellon University in Pittsburgh and the University of California, Berkeley. “A part of our NVIDIA GEAR lab now s
2. "Dangerous" AI models are coming no matter what — "it's the general trend of technology" "Dangerous" AI models are coming no matter what AI models with advanced hacking capabilities will soon be the norm. Lily Hay Newman, WIRED.com – Jun 17, 2026 1:50 pm | 73 Dario Amodei on stage at Code with Claude 2026 in San Francisco. Credit: Samuel Axon Dario Amodei on stage at Code with Claude 2026 in San Francisco. Credit: Samuel Axon Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Minimize to nav Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring “any foreign national” from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings. Since Mythos debuted in April , Anthropic has claimed—and warned—that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double-edged sword in its launch of Mythos 5 and Claude Fable 5. “A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors,” the company wrote in a blog po
3. Anthropic "pauses" token-based billing for its Claude Agent SDK — A last-minute reprieve Anthropic "pauses" token-based billing for its Claude Agent SDK Move originally planned for Monday would have heavily increased power users' costs. Kyle Orland – Jun 16, 2026 5:00 pm | 85 Pricing for Anthropic's Claude Agent SDK isn't changing for the time being. Credit: Anthropic Pricing for Anthropic's Claude Agent SDK isn't changing for the time being. Credit: Anthropic Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Minimize to nav Last month, Anthropic announced a billing change that would have substantially increased costs for heavy users of its automation-focused Claude Agent SDK , including many third-party apps. On Monday, though, Anthropic abruptly announced it had paused those pricing changes just as they were set to take effect, allowing Agent SDK users to continue drawing from the more generous usage limits in their existing Claude subscriptions. The plan, as announced on May 13 , would have treated usage of the Claude Agent SDK (including via third-party apps and the programmatic “claude -p” command ) separately from “standard” Claude usage via the chat interface or the official Claude CLI. At the time, Anthropic said that, as of June 15, that kind of outside SDK usage would be billed at Anthropic’s prevailing API rates , with subscribers receiving a simple monthly usage credit equal to their subscription price. That would have been a major chang
4. Pentagon boasts of using AI to write reports mandated by Congress — AI did my homework Pentagon boasts of using AI to write reports mandated by Congress Pentagon also claims 1.5 million personnel are using generative AI tools. Jeremy Hsu – Jun 16, 2026 2:11 pm | 101 US Secretary of Defense Pete Hegseth and US Chairman of the Joint Chiefs of Staff Gen. Dan Caine testify during a House Committee on Appropriations, Subcommittee on Defense hearing to examine the 2027 budget for the Department of Defense in Washington, DC, on May 12, 2026. Credit: SAUL LOEB / AFP via Getty Images US Secretary of Defense Pete Hegseth and US Chairman of the Joint Chiefs of Staff Gen. Dan Caine testify during a House Committee on Appropriations, Subcommittee on Defense hearing to examine the 2027 budget for the Department of Defense in Washington, DC, on May 12, 2026. Credit: SAUL LOEB / AFP via Getty Images Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Minimize to nav The US Department of Defense has a lot of congressionally mandated homework to do every year involving hundreds of required reports on various national security topics. But Pentagon officials have been proudly describing a new shortcut—using generative AI tools to write such reports for Congress. Pentagon Chief Technology Officer Emil Michael highlighted AI-generated reports to Congress as a key example of how the Department of Defense—stylized as the Department of War under the Trump administration—has adopted generative AI
5. Roelof Botha joins SpaceX's board of directors — Roelof Botha joins SpaceX's board of directors | TechCrunch TechCrunch Desktop Logo TechCrunch Mobile Logo Search Submit Site Search Toggle Mega Menu Toggle Topics Latest AI Amazon Apps Biotech & Health Climate Cloud Computing Commerce Crypto Enterprise EVs Fintech Fundraising Gadgets Gaming Google Government & Policy Hardware Instagram Layoffs Media & Entertainment Meta Microsoft Privacy Robotics Security Social Space Startups TikTok Transportation Venture More from TechCrunch Staff Events Startup Battlefield StrictlyVC Newsletters Podcasts Videos Partner Content TechCrunch Brand Studio Crunchboard Contact Us Image Credits: Kimberly White/Getty Images for TechCrunch Space Roelof Botha joins SpaceX's board of directors Sean O'Kane 1:32 PM PDT · June 17, 2026 Former Sequoia Capital managing partner Roelof Botha is joining SpaceX's board of directors, less than a week after the company went public in the largest IPO ever. SpaceX announced the appointment in a filing with the Securities and Exchange Commission on Wednesday. It said Botha was appointed "to fill the existing vacancy on the Board" and that he will serve until SpaceX's next annual shareholder meeting. He will also join the SpaceX board's audit committee. Botha didn't immediately respond to a request for comment. Botha "brings extensive public company experience along with a deep audit committee background, having served on the boards and audit committees of n
6. After unveiling ridiculously expensive AR glasses, Snap's stock takes a dive — After unveiling ridiculously expensive AR glasses, Snap's stock takes a dive | TechCrunch TechCrunch Desktop Logo TechCrunch Mobile Logo Search Submit Site Search Toggle Mega Menu Toggle Topics Latest AI Amazon Apps Biotech & Health Climate Cloud Computing Commerce Crypto Enterprise EVs Fintech Fundraising Gadgets Gaming Google Government & Policy Hardware Instagram Layoffs Media & Entertainment Meta Microsoft Privacy Robotics Security Social Space Startups TikTok Transportation Venture More from TechCrunch Staff Events Startup Battlefield StrictlyVC Newsletters Podcasts Videos Partner Content TechCrunch Brand Studio Crunchboard Contact Us In Brief Posted: 1:24 PM PDT · June 17, 2026 Image Credits: Joe Scarnici / Getty Images for Snap Lucas Ropek After unveiling ridiculously expensive AR glasses, Snap's stock takes a dive Snap's long-awaited AR glasses , Specs, didn't have the best debut. The company's stock hasn't been on the healthiest trajectory lately. It's dropped 30% over the past year. Following Specs' launch, it sank more than 5% — falling from $5.86 a share on Tuesday to a low of $4.83 on Wednesday morning. As of this writing, the stock still hasn't recovered the position it held prior to the announcement. The big concern surrounding Snap's new smart glasses — which the company has been working on for over a decade — is the cost: The company maintains they will retail at nearly $2,200 apiece. It's worth
7. World leaders want American AI. They just don't want America to be able to turn it off. — World leaders want American AI. They just don't want America to be able to turn it off. | TechCrunch TechCrunch Desktop Logo TechCrunch Mobile Logo Search Submit Site Search Toggle Mega Menu Toggle Topics Latest AI Amazon Apps Biotech & Health Climate Cloud Computing Commerce Crypto Enterprise EVs Fintech Fundraising Gadgets Gaming Google Government & Policy Hardware Instagram Layoffs Media & Entertainment Meta Microsoft Privacy Robotics Security Social Space Startups TikTok Transportation Venture More from TechCrunch Staff Events Startup Battlefield StrictlyVC Newsletters Podcasts Videos Partner Content TechCrunch Brand Studio Crunchboard Contact Us Image Credits: Ludovic MARIN/AFP / Getty Images Government & Policy World leaders want American AI. They just don't want America to be able to turn it off. Rebecca Bellan 12:01 PM PDT · June 17, 2026 At the G7 Summit on Wednesday, world leaders like French President Emmanuel Macron and Indian Prime Minister Narendra Modi voiced concerns that the U.S. could cut off their countries’ access to top American AI models at any time. Macron warned G7 leaders and top AI executives — including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump — over lunch that if the U.S. “from one day to the next can turn off the switch,” it could not only harm the economies of European customers but also damage the AI firms themselves. The comments come a few days after the Trump administration blocked Anthropic fro
8. Want to get a data center online quickly? Give it some flex. — Artificial intelligence Want to get a data center online quickly? Give it some flex. As the data-center boom puts pressure on the grid, some companies say the answer isn’t just more power plants but software that dials down centers’ energy-guzzling ways when demand spikes. By Amos Zeeberg June 16, 2026 PETRA PÉTERFFY At the end of a tense and scoreless first half of a soccer match between the English men’s team and rival Germany, millions of Brits let out a collective sigh and did what they so often do in moments of stress: They made tea. That wave of electric kettles clicking on, however, caused a different kind of stress: a huge and sudden increase in demand for electricity. But National Grid, which operates the local transmission network, was ready. Just as those kettles started heating up, an AI program sent instructions to a data center in London to slow down some of the facility’s power-hungry chips. This reduction helped make sure there was enough supply to match demand, staving off potential blackouts or damage to electrical hardware. For data centers, which normally guzzle power without consideration for anyone or anything else’s needs, it was a radical departure. Advertisement It was also a simulation. In December 2025, engineers sought to test a new breed of data center built to be flexible about its electricity needs, so they re-created the energy demand facing the UK’s grid during a match from the 2020 Euro tournament. They wanted to see how their software, calle
Technical Trends
| Trend | Detail |
|-------|--------|
| AI Coding Agents → Robotics | Physical robot training accelerated by LLM-coded manipulation tasks |
| Agent SDK Pricing | Anthropic pauses token-based billing; SDK economics under review |
| Robot Training Data | Labor-intensive data collection beingoutsourced to gig workers at scale |
| Formal Verification for AI | New startups applying mathematical proof methods to LLM outputs |
| Data Center Grid Flex | Power grid flexibility identified as key constraint for rapid data center deployment |
| Pentagon AI Adoption | US DoD deploying LLMs to draft mandated congressional reports |
| AR/AI Hardware | Snap unveils premium AR glasses; AI integration as differentiator |
| International AI Governance | Allied nations seek American AI capabilities without dependency risk |
Sources
Report generated 2026-06-18 from Ars Technica, TechCrunch, and MIT Technology Review.