IBM’s sub-1nm 3D chips & New black-hole “direct wave” - News (Jun 26, 2026)
Sub-1nm IBM chip leap, new black-hole signal in LIGO data, global self-driving rules, China’s supercomputer claim, and US asylum ruling—June 26, 2026.
Our Sponsors
Today's Top News Topics
-
IBM’s sub-1nm 3D chips
— IBM unveiled a NanoStack prototype aiming for a sub-1nm-era chip with vertical transistor stacking, pointing to huge gains for data centers, cloud computing, and generative AI—if heat and leakage can be solved. -
New black-hole “direct wave”
— A Nature study reanalyzed LIGO data from GW250114 and reports signs of a predicted post-merger “direct wave,” offering a new way to probe black hole horizons, frame dragging, and strong-field general relativity. -
Human embryo NANOG gene editing
— Cambridge researchers used base editing to disable NANOG in donated human embryos and found early development diverges from mouse results, sharpening both fertility insights and ethical debates about heritable gene editing. -
UN HIV declaration toward 2030
— UN Member States adopted a 2026 Political Declaration on HIV and AIDS, reaffirming the 2030 goal and emphasizing 95–95–95 targets, human rights, sustainable financing, and better access to medicines. -
Global rules for self-driving cars
— A UN vehicle standards body approved the first global regulations for fully autonomous driving systems, creating shared safety validation, life-cycle oversight, and data recording requirements to reduce fragmented national rules. -
South Korea’s drone-first military training
— South Korea will train drone operation as a basic military skill, scaling up procurement and counter-drone defenses after lessons from Ukraine and Middle East conflicts and growing concern over North Korea’s drone capabilities. -
China’s reported exaflop supercomputer lead
— China is reported to have a new supercomputer, LineShine, reaching roughly 2.2 exaflops with domestic components—signaling continued progress despite US export controls and intensifying the AI compute race. -
AI’s pivot to “world models”
— Researchers and startups say chatbots are hitting diminishing returns and are pivoting to “world models” that predict how environments change over time—an approach seen as crucial for robotics and planning-capable AI. -
US Supreme Court on border asylum
— The US Supreme Court ruled the government can deny asylum seekers at the border before they physically enter the country, expanding executive power over border processing and reshaping access to protection.
Sources & Top News References
- → IBM Unveils Sub-1nm ‘NanoStack’ Chip Architecture Aimed at 100 Billion Transistors
- → LIGO’s GW250114 Shows Evidence of a Near-Horizon ‘Direct Wave’ After Black Hole Merger
- → Base-edited human embryos reveal a human-specific role for NANOG and renew ethics debate
- → UN adopts new Political Declaration to speed progress toward ending AIDS by 2030
- → UN Approves First Global Safety Rules for Fully Driverless Vehicles
- → South Korea Plans Mass Drone Training for 500,000 Troops as North Korea Threat Grows
- → China’s LineShine Reportedly Overtakes U.S. El Capitan as Fastest Supercomputer
- → AI Researchers and Startups Pivot From Chatbots to ‘World Models’ for Physical and Simulated Environments
- → Supreme Court Allows U.S. to Block Asylum Seekers From Entering at the Border
Full Episode Transcript: IBM’s sub-1nm 3D chips & New black-hole “direct wave”
A faint leftover ripple in LIGO’s data may be revealing something we’ve never clearly “heard” before—signal tied to the edge of a black hole, right after two of them collide. Welcome to The Automated Daily, top news edition. The podcast created by generative AI. I’m TrendTeller, and today is June 26th, 2026. We’ve got breakthroughs and big policy shifts today: a bold new direction for computer chips, a possible new window into black-hole physics, and fresh global rules that could speed the rollout of driverless vehicles—plus major decisions on public health, gene editing, military drones, supercomputing, and US asylum.
IBM’s sub-1nm 3D chips
Let’s start with computing—because IBM just teased a future that sounds almost impossible on paper. The company revealed a prototype architecture it says could squeeze close to 100 billion transistors onto a silicon die about the size of a fingernail, using technology equivalent to roughly a 0.7-nanometre process. If that holds up, it’s one of the first publicly discussed steps into what people call “sub-one-nanometre” territory. What’s especially interesting is the strategy: instead of only making everything smaller on a flat surface, IBM’s “NanoStack” concept builds upward, stacking transistor layers like a skyscraper. IBM claims early tests show about a 50% performance jump compared with its own 2-nanometre chip, alongside major energy savings. The catch is that prototypes aren’t products. Heat, electrical leakage, and manufacturing complexity are huge obstacles. But if this becomes manufacturable, it’s the kind of advance that could materially change the economics of AI and cloud data centers—where power and efficiency are often the real bottlenecks, not just raw speed.
New black-hole “direct wave”
Staying with the theme of national-scale compute power, there’s also a report that China may now be leading in supercomputing again. A system reportedly called “LineShine” is said to hit around 2.2 exaflops on the standard Linpack benchmark—above the US Department of Energy’s El Capitan, reported at roughly 1.74 exaflops. The striking part is the claim that it’s built entirely from domestic Chinese components, despite export restrictions meant to slow China’s access to advanced chips. Details appear limited and not fully verified, which is common with strategic systems. And rankings like TOPS500 don’t always reflect what private companies are running behind closed doors. Still, if the numbers are real, it’s a loud signal: China is pushing hard toward full-stack independence in high-end computing, and that directly shapes the global AI and national-security race.
Human embryo NANOG gene editing
And speaking of AI—there’s a noticeable shift underway in what some leading researchers and startups are chasing. A growing group says that simply scaling up chatbots is delivering smaller and smaller breakthroughs, and that the next meaningful step is building “world models”—systems that learn how environments behave over time and respond to actions, not just prompts. The reason this matters is practical: robots and interactive systems have to handle space, motion, and physical cause-and-effect—things language models often stumble on. Well-known voices, including Fei-Fei Li and Yann LeCun, are among those pushing the idea that planning and prediction—under real-world constraints—are closer to what we’d call general intelligence. Investors are already backing this direction, even though the near-term applications are less obvious than a chatbot subscription. The long-term prize is big: AI that can reliably plan and act, not just talk.
UN HIV declaration toward 2030
Now to space—and a finding that could expand what gravitational-wave astronomy can tell us. A paper in Nature reports evidence, in the event GW250114, for a “direct wave” component emitted just after two black holes merge. In plain terms: after researchers removed the loud, expected “ringdown” signal, they found a leftover pattern that lines up with a long-predicted signature tied to the newly formed black hole’s horizon region. Why is that a big deal? Because it hints at a more direct way to test near-horizon physics—things like frame dragging in an extreme setting—rather than relying only on the standard after-merger vibrations. The study says the frequency and the way the signal fades match what general relativity predicts for a spinning, Kerr black hole remnant. This is still careful, statistical science—future events will matter a lot. But it’s a potential new handle on one of the most inaccessible places in the universe: the boundary of a black hole itself.
Global rules for self-driving cars
Shifting to health and ethics, researchers at the University of Cambridge report new results from gene editing in donated human embryos—work that’s scientifically informative and politically sensitive. Using a precise method called base editing, the team disabled the NANOG gene and cultured embryos for about a week. The edited embryos didn’t form a normal epiblast—the group of cells that eventually becomes most body tissues and organs—while other supporting cell types still appeared. The big takeaway is that humans don’t always follow the same developmental script as mice, even for genes scientists thought were well understood. That matters for fertility research and reproductive medicine, because it helps clarify what’s truly human-specific. At the same time, the work intensifies the bigger debate: as embryo-editing techniques improve, societies face more pressure to decide what should be permitted, and under what safeguards—especially because safety problems like mosaicism, where not all cells are edited consistently, remain unresolved.
South Korea’s drone-first military training
On global public health, UN Member States have adopted a new Political Declaration on HIV and AIDS at the 2026 UN General Assembly High-Level Meeting in New York. This is the last such high-level meeting before the 2030 deadline to end AIDS as a public health threat, and the declaration is meant to keep momentum as funding and attention get pulled in many directions. Countries reaffirmed the push toward the 95–95–95 goals, emphasized integrating HIV services into broader health systems, and highlighted human rights—targeting stigma, discrimination, and harmful laws as core barriers, not side issues. The declaration also calls for more reliable financing, stronger domestic investment, and better access to affordable medicines, including support for local production and research. Whether this translates into on-the-ground progress will depend on budgets and enforcement, but politically, it’s a marker: the world is trying to hold the line on HIV progress rather than let it drift.
China’s reported exaflop supercomputer lead
Now to transport and regulation: a UN vehicle standards body has approved what it calls the first global regulations for fully autonomous driving systems. This matters because one of the biggest obstacles to driverless deployment isn’t just technology—it’s inconsistent rules. The new framework, adopted under UNECE’s World Forum for Harmonization of Vehicle Regulations, sets shared safety expectations and common ways to validate these systems, aiming to prevent a patchwork of national standards. It includes oversight across the full life cycle, from development to real-world operation, and it requires ongoing monitoring after vehicles are on the road. It also calls for data storage so safety-relevant events can be reviewed—key for accountability. Major markets are backing the regulation, and it’s expected to take effect in about a month. It doesn’t mean driverless cars will suddenly be everywhere—but it does lay a clearer global track for how they’re judged and governed.
AI’s pivot to “world models”
On defense and changing warfare, South Korea is making drone operation a basic skill across its military. The defense ministry says it wants every soldier to be able to use drones as naturally as a standard piece of equipment, planning to train roughly 500,000 authorized personnel as what it calls “drone warriors.” The decision is heavily influenced by what conflicts in Ukraine and the Middle East have shown: low-cost drones, used at scale, can reshape tactics fast. Seoul also points to a direct regional driver—North Korea’s growing drone capabilities. A 2022 incident, where North Korean drones entered South Korean airspace and crossed even sensitive zones, exposed gaps that South Korea now wants to close. Alongside training, South Korea plans to expand counter-drone defenses, including advanced interception tools, and accelerate domestic development of loitering munitions. It’s modernization under pressure—because in this area, the learning curve is happening in real time.
US Supreme Court on border asylum
Finally, a major legal shift in the United States: the Supreme Court ruled, 6 to 3, that the federal government can turn away would-be asylum seekers at the border before they physically enter the country. The majority said that under immigration law, asylum protections apply to people who “arrive in” the United States—and that being stopped at the border means you haven’t arrived yet. The dissent argued that presenting yourself to authorities at a legal entry point is effectively the beginning of arrival, and that blocking people at that stage undermines access to asylum. The practical impact is significant: the decision gives the executive branch broader authority to limit asylum processing at the border, especially during surges. And it’s likely to reshape how policy is enforced—who gets a chance to make a claim, and where that threshold is drawn.
That’s the Top News Edition for June 26th, 2026. If you’re watching the big themes, today’s throughline is limits—and what happens when we push past them: chips stacking upward when shrinking hits a wall, AI moving beyond chat to systems that can predict and plan, and policymakers trying to set guardrails for everything from autonomous vehicles to asylum. Thanks for listening. I’m TrendTeller, and this was The Automated Daily. Check back tomorrow for the next briefing.
More from Top News
- June 24, 2026 Interstellar comet older than Sun & Early galaxy and cosmic reionisation
- June 23, 2026 Ancient interstellar comet clues & NHS rollout delays type 1
- June 22, 2026 AI CEOs at the G7 & Canada moves to regulate chatbots
- June 21, 2026 Global push for slavery reparations & AI as new geopolitical leverage
- June 20, 2026 Global push for slavery reparations & AI export controls reshape power