🎧 Listen ADS-FREE: https://podcasts.apple.com/us/channel/djamgamind/id6760446113

Visit our Research Hub at https://djamgamind.com/pdfs

Subscribe to our Newsletter at

AI Unraveled

Demystifying Frequently Asked Questions on Artificial Intelligence: Latest AI News & Trends, ChatGPT, Gemini, DeepSeek, Gen AI, LLMs, Agentic AI, Ethics, Bias

By Etienne Noumen

#DJAMGAMIND

Summary: In today’s deep-dive, we deconstruct “The Architectural Flip from Cloud to Local Compute.” We analyze Nvidia’s comprehensive offensive at Computex 2026, including the ARM-based RTX Spark superchip, the agent-optimized Vera CPU, and Cosmos 3—a 20-trillion-token multimodal world model built to train physical robotics. We detail Microsoft’s Surface Laptop Ultra, designed to execute large models locally to escape corporate token strain. We explore the financial macroscape as Anthropic confidentially files for its IPO at a $900 billion valuation, while Senator Bernie Sanders outlines a legislative framework to reclaim AI equity for the public via an American Sovereign Wealth Fund. Additional coverage includes China’s landmark approval of the invasive NEO brain-computer interface and SoftBank’s $87 billion infrastructure investment in France.

Important Topics:

• Nvidia Launches the Vera CPU: Nvidia drops an enterprise CPU designed specifically for agentic loops, executing tool use and local coding workflows 1.8x faster than traditional x86 chips.

• The RTX Spark and Surface Ultra: Microsoft introduces the Surface Laptop Ultra powered by Nvidia’s 3-nanometer RTX Spark superchip, supporting 128GB of unified memory and local execution of 120B parameter models.

• Anthropic Files Confidentially for IPO: Anthropic moves to go public with an internal private valuation hitting $900 billion, widening its competitive spread over OpenAI’s $730 billion private valuation.

• The American AI Sovereign Wealth Fund: Senator Bernie Sanders publishes a radical proposal to levy a 50% equity tax on major AI players like OpenAI, Anthropic, and xAI, distributing shares directly into a public trust.

• China Approves First Invasive BCI: The Chinese government clears Neuracle Technology’s “NEO” chip for public medical use outside trials, utilizing an innovative non-cortical-piercing dura mater sensor placement.

• The Data Scraper Crackdown: Strava enforces strict paywalls on its public profiles and transitions to an $11.99 monthly flat API fee to block automated collection by frontier AI labs.

• MicroAGI Traps Physical Household Data: German startup MicroAGI launches its “Shift” app in New York, offering free apartment cleaning in exchange for point-of-view camera footage to train autonomous humanoid systems.

🔗 RESOURCES

Stop scrolling through generic “Top 10” lists and head over to the AI Executive Toolkit at https://djamgamind.com/toolkit

• Find REMOTE AI Jobs (Mercor): Apply Here

• AI & ML TUTOR PRO: https://apps.apple.com/ca/app/ai-ml-tutor-pro/id1610947211

⚗️ PRODUCTION NOTE: We Practice What We Preach.

AI Unraveled is produced using a hybrid “Human-in-the-Loop” workflow.

Nvidia jumps into PCs with new Arm-based chip LINK

• Nvidia has unveiled the RTX Spark, a new superchip for Windows PCs built with Microsoft that the company is calling the most efficient PC chip ever made, arriving this fall.

• The RTX Spark fuses a Blackwell GPU with a new Arm-based N1X CPU custom designed by MediaTek, paired with 128 gigabytes of unified memory and built on TSMC’s 3-nanometer technology.

• Nvidia plans more than 30 laptops and 10 desktops from Dell, HP, ASUS, Lenovo, MSI and Microsoft, with the first laptops as thin as 14 millimeters and aimed at creators, AI developers and gamers.

Nvidia bets it has the CPU for the AI agent era

It’s official. Nvidia is no longer just a GPU company. Its a full-stack AI company.

The chipmaker may have backed into its plum role as the chief supplier and kingmaker of the AI revolution, but no one can argue that it isn’t seizing the moment.

On Monday, Nvidia used the annual Computex event in Taiwan to unleash a flurry of announcements that ranged across the AI stack from CPUs to data centers to open models. Here’s our quick analysis of the most important news:

• Vera CPU now designed for agents: Nvidia declared that its new Vera CPU is specifically designed for agentic workloads, including tool use, writing code, and processing data. It claimed that Vera is a new type of CPU that can complete these tasks 1.8x faster than traditional x86 CPUs. It has signed up OpenAI, Anthropic, SpaceXAI, CoreWeave, Lambda, Dell, HPE, Lenovo, and others to use the chips.

• New platform for AI factories: The company launched the DSX software platform for designing, simulating, building and operating AI factories. That includes DSX MaxLPS, which optimizes power and cooling to run 40% more GPUs on the same power budget, and DSX OS, which the company is pitching to become the open-source operating system for AI factories.

• AI supercomputer for Windows desktops: The new DGX Station for Windows can run AI models up to 1 trillion parameters locally for better performance, security, and cost savings. This is aimed at engineers, researchers, developers, and enterprises.

• New models for robots and robotaxis: The company announced Cosmos 3 (covered by The Deep View’s Nat Rubio-Licht), a new world model that can work across text, images, video, sound, and actions to power physical AI and robotics. It also announced Alpamayo 2 Super, a new open reasoning model for robotaxis and H2 Plus, an open reference design for humanoid robots that combines hardware, software, and onboard compute.

• Vera Rubin hits full production: The next-gen AI infrastructure platform that combines GPUs, CPUs, networking, storage, and security and was unveiled at CES is ramping into full production across 350 factories in 30 countries. It delivers 10x higher agent throughout than the previous generation Grace Blackwell platform, Nvidia claims. Expect this to become the system that will power the world’s largest and most powerful AI factories in the years ahead.

Pay special attention to the CPU news. Nvidia CEO Jensen Huang said, in a statement, “AI agents will be the largest users of computing. Vera is the first CPU designed for that future — built to run agentic AI at hyperscale with extraordinary performance, efficiency and programmability.”

Nvidia expands into the world. By Ina Fried

Image: Nvidia

Nvidia unveiled Cosmos 3, an open AI world model designed to help robots, autonomous vehicles and other physical systems better understand and predict real-world environments.

Why it matters: Nvidia is continuing its move beyond chips into AI models and software, positioning itself to become a foundational platform for physical AI development.

Driving the news: Nvidia says it trained Cosmos 3 on 20 trillion tokens of multimodal data, including nearly a billion images, 400 million real and synthetic videos, ambient audio, text and action data from humans and robots.

• That action data is what makes Cosmos different from a regular video generator. It’s meant to model how machines move, not just how scenes look, Ming-Yu Liu, VP of Nvidia’s Cosmos Lab, told Axios. Autonomous actions are key.

• Developers can use Cosmos 3 to simulate actions in physical environments, then build task-specific models for robots and other machines on top of it.

• Cosmos 3 is designed to generate action data — such as robot joint angles, gripper positions and trajectories — that can help train machines to navigate and manipulate the physical world.

Between the lines: Cosmos is an open model, similar to its early Nemotron family, making it easier for hardware makers to customize Cosmos to their needs and ensure that future versions more closely align to the needs of the industry, Liu said.

• Nvidia is also establishing a coalition of companies supporting the effort. Initial partners include Agile Robots, Black Forest Labs and Runway.

• Nvidia says Cosmos can create rare or dangerous scenarios — such as robot collisions or unusual road events — that are difficult, expensive or unsafe to capture repeatedly.

Zoom in: Nvidia is releasing two versions immediately: a “super” model for tasks requiring high physics accuracy, such as training robots and autonomous vehicles, and a “nano” model that can produce results in fractions of a second.

• An “edge” model that can run locally is coming soon, Nvidia said.

Zoom out: World models have become a key growth area for AI as companies increasingly want to take the smarts of chatbots and agents and allow them to perform real-world tasks.

• Among the hot startups in this area are Fei-Fei Li’s World Labs and Yann LeCun’s AMI Labs.

• “Ultimately what a world model wants to achieve is to help physical agents to become more generalizable,” Liu said. “To become more generalizable, you need to understand the world so you understand how it works, so you can make a plan.”

Bottom line: Nvidia’s bet is that the next wave of AI won’t just answer questions or generate images — it will need to predict, simulate and act in the physical world, and Nvidia wants its open models and infrastructure to be the place developers start.

Apple targets eyewear market with 2027 smart glasses LINK

• Apple is preparing to enter the eyewear market in 2027 with smart glasses that will compete not only against Meta and Samsung but also traditional brands like Oakley, Ray-Ban, and Warby Parker in the $200-$500 range.

• According to Bloomberg’s Mark Gurman, Apple is skipping the luxury end of eyewear, a market worth an estimated $180 to $200 billion annually, after its $10,000 gold Apple Watch failed to gain traction with high-end buyers.

• Apple is betting that its brand, industrial design, iPhone integration, ecosystem of over 2 billion active devices, global retail footprint, and artificial intelligence features will convince regular glasses shoppers to pick an Apple pair instead.

China approves first invasive brain-computer chip LINK

• China has cleared NEO, the world’s first invasive brain-computer interface approved for use outside clinical trials, giving patients aged 18 to 60 with limb paralysis from spinal cord injuries access to the implant.

• Built by Shanghai startup Neuracle Technology with Tsinghua University, the coin-size NEO places eight sensors on the dura mater rather than piercing the cortex like Neuralink’s N1, lowering the risk of hemorrhage and scarring.

• The device reads brain signals and sends them to a computer that drives a soft robotic glove during daily 2.5-hour training, and China is already adding NEO to its health insurance system through a unique code.

US closes loophole on Nvidia chips reaching China LINK

• The US Commerce Department has shut a loophole that let Chinese AI companies buy Nvidia’s top chips through overseas subsidiaries, with new guidance tying export-licence rules to where a company is headquartered rather than physically located.

• The rule now covers advanced processors including Nvidia’s Rubin and Blackwell and AMD’s MI350x, and one industry source told Reuters that hundreds of thousands of these chips may have reached Chinese-linked entities abroad while the gap was open.

• The guidance applies only to future sales, so data centres already running the chips can keep using them and servicing of servers continues, closing ambiguity left after the Trump administration declined to enforce the Biden-era AI Diffusion rule in May 2025.

Strava declares war on scrapers LINK

• Strava is cracking down on AI scrapers by locking public profiles and fitness club listings behind logins, and charging developers a flat $11.99 monthly fee for API access that used to be free under a tiered program.

• The company will retire some API endpoints that let outside apps pull data like club details, give developers a 90-day grace period, and add support for Model Context Protocol so AI assistants get structured, controlled access.

• CEO Michael Martin said Strava refused data licensing deals from leading AI labs and called out Perplexity for routing scraping through aggregator services to hide its origin after being turned away by the company.

Anthropic files to go public:

The TWiST 500 is about to lose one of its most prominent, headline-grabbing members: Claude maker Anthropic. The AI company confidentially filed to go IPO on Monday, joining SpaceX and OpenAI as one of the three largest and highest-profile private companies in history to make the switch this year. This continues Anthropic’s closely-watched, industry-defining rivalry with OpenAI, which also has an IPO brewing, and which is just surpassed in valuation while taking on $65 billion in new financing last week. (Anthropic is now valued at $900 billion, rocketing ahead of OpenAI’s $730 billion.) No details yet on the specific timing or exact size of this mega-offering; Anthropic suggests it will “depend on market conditions” among other factors. Look for it on stock exchanges (and maybe your retirement portfolio) starting in the fall.

Mecka joins the “robot surveillance training” trend:

Seems like not a day goes by any longer when we don’t hear about a startup using surveillance videos of humans to train robots. The latest is Mecka, a New York startup that reportedly just raised $60 million in Series A and follow-on investments. They’re using a combination of techniques — including human workers wearing “body sensors” and recording themselves on iPhones — to collect data for robot training. Co-founder and CEO Josh Gao says that, unlike so many other AI companies collecting training data, Mecka hopes to also work on the “front lines” of the robotics industry, actually applying the data to fine-tune and deploy in physical space.

Sen. Bernie Sanders wants the public to own AI:

In a new op-ed for the New York Times, the Vermont independent suggests that it’s a foregone conclusion that AI will change the world, and the question that remains is: “Who will own and control that future?” He suggests the answer is “everyone” as opposed to “a handful of billionaires,” and further posits that AI’s current level of innovation is only possible because the models have been trained on the “learnings of humanity” and our “collective experience [and] knowledge.” Therefore, Sen. Sanders calls for the passage of an American AI Sovereign Wealth Fund bill, which would tax 50% of the stock of OpenAI, Anthropic, xAI, and other major industry players, then distribute the shares to the public via a sovereign wealth fund. Sen. Sanders also notes that a number of AI executives have made similar, though less specific, calls to nationalize AI profits in the past. He cites quotes from Anthopic’s own blog and xAI chief Elon Musk about creating national AI sovereign wealth funds or distributing “universal high income checks” to every American. Will these and other AI leaders be enthusiastic about what’s now transitioned from an abstract concept into a specific legislative proposal? Time will tell.

Startup cleans apartments in exchange for AI data

Image source: Shift

The Rundown: German startup MicroAGI’s Shift app just opened a free home-cleaning service in New York City that records its cleaners through head-mounted cameras, trading chores for first-person data to both sell to AI labs and use in its own AI research.

The details:

• A vetted cleaner shows up wearing a camera that co-founder Bercan Kilic calls a “magic hat,” filming the roughly two-hour job point-of-view style.

• Despite covering the cost of the cleaning, the human footage is worth more to robot makers for training, letting Shift cover the bill and still profit.

• Shift’s site claims to already pay people across the world $20 an hour to film everyday chores, with $5M+ paid out in Q1 across a variety of tasks.

• GM Harry Kilberg said the launch drew “thousands and thousands of bookings,” with New York first and London, Munich, and Zurich next.

Why it matters: As we’ve seen with DoorDash paying couriers to capture task data, the next AI dataset is coming from ordinary human work instead of the internet. Shift pushes that model deeper into the home, where people are both the customers getting free service and the workforce teaching robots how to replace pieces of the job.

Ex-DeepMind group tackles self-improving science AI

Image source: Inherent Labs

The Rundown: Several Ex-Google DeepMind employees came out of stealth with $50M for Inherent Labs, a London startup building an AI science platform that puts scientists alongside self-improving AI to work out which problems are worth pursuing.

The details:

• Co-founders Tantum Collins, Edward Hughes, and Louis Kirsch came from DeepMind, while Kaloyan Aleksiev previously worked at Reka AI and Microsoft.

• The Faraday platform will pair researchers with self-improving agents built to spot higher-value scientific questions instead of only answering prompts.

• The lab says it will test recursive self-improvement across the research org, including everything from agent training to resource allocation and decisions.

• Inherent is also exploring what “AI taste” looks like in science as the research process shifts, and how humans and machines can best work together.

Why it matters: Self-improving AI is a quest many of the top AI labs and newly funded startups are trying to tackle, and Inherent adds another group with strong pedigrees to the list. Inherent is applying the recursive logic to science itself, making the entire lab and organization the loop instead of just the model training.

YouTube Finally Breaks Through to Hollywood

This feels extremely long overdue, but 2026 has really become the year that Hollywood and YouTube finally found a way to work together in perfect harmony. Hollywood has faced so many challenges over the past few decades. Piracy, better TV shows, streaming, COVID, strikes, competition from other production markets, and many more. Throughout all of that, there was a silver lining of the creator economy growing to fill the gaps. Sure, the number of traditional Hollywood shoot days (and jobs) were in decline, but the number of people making a living in front of or behind a camera was growing overall thanks to the broader content boom. This was always unsatisfying to cinephiles though, TikToks just don’t have the staying power of a cultural moment like Titanic.

The Oscars will officially be streamed on YouTube in 2029, and by then, the trend that really broke out in 2026 will be even more undeniable. Just look at the stats:

• Kane Parsons’ Backrooms opened to roughly $81.5 million in North America and $118 million worldwide on a reported $10 million budget.

• Curry Barker’s Obsession kept climbing in its third weekend and hit $104.7 million domestic, becoming Focus Features’ highest-grossing domestic release, from a movie widely reported to have cost around $1 million.

• Markiplier’s Iron Lung had a reported $3 million production budget and opened to about $18.2 million domestically, before ultimately grossing ~$41.1 million domestic / $51.2 million worldwide.

It’s easy to point to these as “YouTuber with a big audience converts to theatrical sales” but it’s not that clean. Ryan’s World the Movie: Titan Universe Adventure, grossed only $624k on a $10M budget, despite Ryan’s World being an enormous children’s YouTube brand. Smosh: The Movie and Lazer Team from Rooster Teeth were both earlier attempts to cross the YouTube to Hollywood gap that both fell short of expectations.

What’s different about Backrooms, Obsession, and Iron Lung is that the filmmakers had shown the type of creativity and risk-taking that underlies so many Hollywood successes. These creators aren’t just big influencers with millions of casual fans. They do have big audiences, but they also stand above their peers in terms of artistic vision. Curry Barker had a YouTube sketch channel called That’s a Bad Idea where he learned how to quickly and effectively write, act, and edit for a tight audience feedback loop. Backrooms has a similar story, Kane Parsons, known as Kane Pixels, produced his original series, The Backrooms, in Blender and After Effects. The internet myth laid a bit of the groundwork, but he was able to take it in his own direction with lore he built up over several videos. Markiplier did have an objectively huge audience, but still went “full-stack” when producing Iron Lung. He even talked about building a server rack in his bathroom to render VFX shots on a faster turnaround.

Being able to create something engaging for social media virality is probably somewhat important to creating a film that works in theaters, but the bigger value is being a “full stack” filmmaker. Gone are the days of showing up to Hollywood with a manuscript and expecting a studio to do the rest for you. The traditionally segmented teams on productions are simply too expensive to be deployed on anything but existing IP. New projects will come from filmmakers who have experience and a view for every part of the filmmaking process.

Up next, we have two similar projects potentially in the works. One is from Wesley Wang, who went viral for a non-horror YouTube short film called “nothing, except everything.” TriStar picked it up with Darren Aronofsky’s Protozoa producing and Wang set to adapt it as writer-director. Then there’s also the much sillier, but extremely viral Skibidi Toilet. Created by Alexey Gerasimov in 2023, the project reportedly has Michael Bay involved, but there are some IP issues to work through since the series was produced in Source Filmmaker and uses assets from Half-Life 2 and Counter-Strike: Source pretty liberally. There will undoubtedly be other unpredictable breakouts in the coming years, and I wouldn’t be surprised if we see Hollywood executives combing through obscure YouTube playlists for new gems. — John

Jensen gives Computex 2026 keynote

Nvidia CEO Jensen Huang made several announcements during his keynote speech at Computex, Taiwan’s annual technology expo. He introduced the Nvidia RTX Spark Superchip, an ARM-based chip for PCs designed to process AI workloads locally; he announced that the enterprise Vera Rubin CPU is now in full production; and he announced several foundation models — a world model, an open-weight flagship AI model, and a model specifically designed for humanoid robots.

A lot of the attention from the keynote is going to the RTX Spark, which is being viewed as Microsoft’s attempt to create its own ‘Apple silicon moment’. And Microsoft coordinated its announcement of its new Surface Laptop Ultra, which will be optimized for RTX Spark, and will be released this fall.

Watch a cutdown of Jensen’s keynote here. —Brandon

Microsoft debuts Nvidia-powered Surface UltraBy Ina Fried

The Nvidia-powered Microsoft Surface Ultra. Image: Microsoft

Microsoft has officially unveiled the Microsoft Surface Ultra, the first full-fledged Windows PC to run on an Nvidia main processor, saying the device is “made for a kind of work that does not fit in a standard laptop.”

Why it matters: The move, which we scooped Friday night, puts the hottest name in chips behind Windows as Microsoft tries once again to redefine the PC for the AI era.

Driving the news: The new Microsoft Surface Laptop Ultra will be powered by the RTX Spark, a PC chip that is similar to the one that powers the Nvidia Spark line of AI desktop machines, per The Verge.

• Microsoft said the device, which has a 15-inch mini-LED touchscreen, can be outfitted with up to 128GB of unified memory and can run AI models of up to 120 billion parameters.

• Nvidia-powered Windows laptops and desktops are also coming this fall from Asus, Dell, HP and Lenovo, among others.

What they’re saying: “The work from creators, developers and AI builders has a common shape: massive scenes, long compile cycles, local models and datasets that no longer sit politely in the background,” Microsoft corporate VP Brett Ostrom said in a blog post. “We built Surface Laptop Ultra to meet that work without flinching.”

Yes, but: Microsoft and Nvidia were short on the big details. Microsoft said only that its device would come later this year, with no indication of price, though it’s expected to be a very high-end device.

• Nvidia said that Windows laptops with its chips would start shipping this fall.

The bottom line: While most AI work has been done in the cloud, Microsoft’s push to have things run locally could find newly receptive ears.

• Businesses are starting to struggle with massive computing costs that have accompanied the shift from unlimited-use chatbots to agents, which can rack up giant bills as they do their autonomous work.

What Else Happened in AI on June 01st 2026?

Microsoft is reportedly set to merge GitHub Copilot, chat, Cowork, and Autopilot into one super app, matching similar all-in-one plays OAI and Musk’s X are already making.

SoftBank committed up to $87B to build what will be the largest AI data center project in France, extending an infrastructure spree that includes a $60B+ OAI stake.

Nvidia unveiled RTX Spark, a superchip designed for personal AI agents, bringing up to 1 petaflop of AI performance to Windows laptops with all-day battery life.

Sam Altman posted that OpenAI is hiring for its Robotics division, saying that the company imagines “everyone having a personal robot doing anything they need.”

OpenAI launched Rosalind Biodefense, giving the U.S. government and vetted partners access to its biology AI for pandemic-preparedness and outbreak-response.