He stated he’s pursuing this project because semiconductor manufacturers are not producing chips fast enough to meet his companies’ AI and robotics demands. Musk said:

“We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab.”

According to Bloomberg Musk shared his plans during an event in downtown Austin, Texas, with a photo indicating that the “Terafab” facility will be “gigafactory,” located near Tesla’s Austin headquarters.

He also added that the aim is to produce chips capable of supporting 100–200 gigawatts of computing power annually on Earth, as well as one terawatt in space. He did not provide a timeline for the plan.

Image: Presentation of the Terafab project

The founder of the British consumer electronics company that develops smartphones and other accessories made these comments during an interview at the SXSW conference.

In terms of AI in software, I think people should understand that apps are going to disappear,

So, if you’re a founder or a startup and your app is like where the core value lies, that will be disrupted whether you like it or not.

The company is pitching the idea for some time now about a new kind of smartphone using AI and personalization technology accurate enough so its users won’t feel they have to double-check its output.

Some companies already implemented AI features that can execute a command on the users’ behalf, like booking flights or hotels. However Pei believes the AI could begin to learn a user’s intentions long-term. For example, if you wanted to be healthier, the device could give you nudges to help you accomplish your goals.

I think it gets even more powerful when it starts surfacing suggestions for you; you don’t have to manually come up with an idea…when the system knows us so well, it will come up with things that we don’t even [know] we wanted,

Pei explained.

Pei believes AI-first smartphone would do things for its users without needing to be commanded to. This would mean a device with an interface designed for the AI agent to use.

Despite this Pei doesn’t think apps are going away in the near future, but over time the AI will use the “app” in a frictionless way, not mimicking human touch on the smartphones by moving through menus and tapping options.

That’s not the future. The future is not the agent using a human interface. You need to create an interface for the agent to use. I think that’s the more future-proof way of doing it.

Image: TechCrunch. (2019). TechCrunch Disrupt San Francisco 2019 – Day 3 Carl Pei (cropped). Wikimedia Commons

NVIDIA and Global Industrial Software Giants Bring Design, Engineering and Manufacturing Into the AI Era
NVIDIA announced partnerships with Cadence, Dassault Systèmes, PTC, Siemens and Synopsys to bring CUDA-X, Omniverse and GPU-accelerated tools into industrial workflows. The effort targets faster design, engineering and factory optimization across major manufacturers.

Hyundai Motor, Kia and NVIDIA Expand Strategic Partnership for Next-Generation Autonomous Driving Technology
NVIDIA said Hyundai and Kia are expanding their work with the company around autonomous driving built on the DRIVE Hyperion platform. The announcement signals continued momentum for NVIDIA’s automotive stack in next-generation vehicle systems.

NVIDIA Announces Open Physical AI Data Factory Blueprint to Accelerate Robotics, Vision AI Agents and Autonomous Vehicle Development
NVIDIA introduced an open reference architecture designed to automate the way physical AI training data is generated, augmented and evaluated. The blueprint is meant to cut the cost and complexity of training robotics, vision and autonomous-driving systems.

Roche Scales NVIDIA AI Factories Globally to Accelerate Drug Discovery, Diagnostic Solutions and Manufacturing Breakthroughs
Roche is expanding its NVIDIA deployment to more than 3,500 Blackwell GPUs across global operations. NVIDIA framed the project as a major example of AI factories being used across R&D, diagnostics and manufacturing.

NVIDIA, T-Mobile and Partners Integrate Physical AI Applications on AI-RAN-Ready Infrastructure
NVIDIA and T-Mobile said they are working with Nokia and developers to bring physical AI applications onto distributed edge AI networks. The project points to telecom infrastructure becoming a platform for real-time AI services.

Adobe and NVIDIA Announce Strategic Partnership to Deliver the Next Generation of Firefly Models and Creative, Marketing and Agentic Workflows
Adobe and NVIDIA unveiled a broader alliance around AI-powered creation, production and personalization. The partnership includes work on future Firefly models and agentic workflows for creative and marketing use cases.

BYD, Geely, Isuzu and Nissan Adopt NVIDIA DRIVE Hyperion for Level 4 Vehicles
NVIDIA said adoption of DRIVE Hyperion is expanding with BYD, Geely, Isuzu and Nissan, alongside other mobility players. The company presented this as evidence of growing demand for scalable Level 4 autonomous vehicle platforms.

NVIDIA Ignites the Next Industrial Revolution in Knowledge Work With Open Agent Development Platform
NVIDIA launched an open agent development platform centered on the Agent Toolkit and OpenShell runtime. The goal is to help enterprises build and manage autonomous, self-evolving AI agents more safely and efficiently.

NVIDIA Launches Nemotron Coalition of Leading Global AI Labs to Advance Open Frontier Models
NVIDIA introduced the Nemotron Coalition, a new collaboration among open-model builders and AI developers. The initiative is designed to pool research, data, expertise and compute to accelerate frontier open-model development.

NVIDIA Expands Open Model Families to Power the Next Wave of Agentic, Physical and Healthcare AI
NVIDIA expanded its open model families with new releases aimed at developers and scientists building systems that can reason and act across enterprise, robotics and healthcare settings. The company is positioning open models as a foundation for broader AI deployment.

NVIDIA Announces NemoClaw for the OpenClaw Community
NVIDIA unveiled NemoClaw, a stack for the OpenClaw agent platform that lets users deploy Nemotron models and the OpenShell runtime with a single command. NVIDIA said the package adds privacy and security controls for more trustworthy autonomous agents.

NVIDIA Launches Space Computing, Rocketing AI Into Orbit
NVIDIA said its accelerated computing platforms are being extended to orbital data centers, geospatial intelligence and autonomous space operations. The announcement marks a push to bring data-center-class AI performance into space-constrained environments.

NVIDIA Releases Vera Rubin DSX AI Factory Reference Design and Omniverse DSX Digital Twin Blueprint With Broad Industry Support
NVIDIA released a reference design for Vera Rubin DSX AI factories and made its Omniverse DSX digital twin blueprint generally available. The company said the combination should help customers design and simulate large AI infrastructure before deployment.

NVIDIA Launches Vera CPU, Purpose-Built for Agentic AI
NVIDIA introduced the Vera CPU, which it described as the first processor purpose-built for agentic AI and reinforcement learning. The company claims the chip delivers higher efficiency and faster performance than traditional rack-scale CPUs.

NVIDIA Launches BlueField-4 STX Storage Architecture With Broad Industry Adoption
NVIDIA announced BlueField-4 STX, a modular storage architecture designed for long-context reasoning in agentic AI. The system is intended to help enterprises and cloud providers deploy accelerated storage infrastructure more easily.

NVIDIA Vera Rubin Opens Agentic AI Frontier
NVIDIA said the Vera Rubin platform is entering full production with seven new chips to scale large AI factories. The company is presenting Rubin as a key platform for the next stage of agentic AI infrastructure.

NVIDIA DLSS 5 Delivers AI-Powered Breakthrough in Visual Fidelity for Games
NVIDIA unveiled DLSS 5, calling it its biggest graphics breakthrough since real-time ray tracing. The company says the new version pushes image quality and performance further through AI-powered rendering.

From Simulation to Production: How to Build Robots With AI
In a GTC blog post, NVIDIA outlined how its latest open models, simulation tools and embedded compute are being combined to speed cloud-to-robot workflows. The message was that robotics development is moving toward a more integrated software-to-deployment pipeline.

More Than Meets the Eye: NVIDIA RTX-Accelerated Computers Now Connect Directly to Apple Vision Pro
NVIDIA said CloudXR 6.0 now integrates natively with visionOS, allowing RTX-powered simulators and professional 3D applications to connect directly to Apple Vision Pro. The announcement broadens NVIDIA’s position in immersive enterprise and design workflows.

NVIDIA, Telecom Leaders Build AI Grids to Optimize Inference on Distributed Networks
NVIDIA used a blog post to argue that telecom networks are becoming a new layer for distributed inference. The company highlighted operator activity in the U.S. and Asia as AI-native applications expand to more devices, users and agents.

GTC Spotlights NVIDIA RTX PCs and DGX Sparks Running Latest Open Models and AI Agents Locally
NVIDIA said GTC 2026 showcased a new category of “agent computers,” including RTX PCs and DGX systems running open models and AI agents locally. The pitch is that personal computing is shifting toward on-device generative and agentic AI.

Snap Decisions: How Open Libraries for Accelerated Data Processing Boost A/B Testing for Snapchat
NVIDIA highlighted Snap’s use of open NVIDIA data-processing libraries on Google Cloud to speed product development and experimentation. The post positions accelerated analytics as a way to make large-scale A/B testing faster and more efficient.

Smooth Moves: 90 Frames-Per-Second Virtual Reality Arrives on GeForce NOW
NVIDIA said GeForce NOW now supports 90 fps streaming for compatible VR headsets. The update is aimed at making cloud gaming smoother and more immersive for virtual reality users.

Image: NVIDIA

In the conversation with venture capitalist Bill Gurley, Tan said.

I sleep, like, four hours a night right now,

He added jokingly

I have cyber psychosis, but I think a third of the CEOs that I know have it as well,

Tan compared his work with AI to rebuilding a startup that previously required significant time, funding, and even stimulant use.

Once you try it, you’ll realize: It’s like I was able to re-create my startup that took $10 million in VC capital and 10 people, and I worked on that for two years, and I took anti-narcoleptics — I remember, you know, sort of being on modafinil,

He claims now AI has replaced the need for such aids.

I don’t need modafinil with this revolution. Like, I’m up. I slept at 4 a.m. I woke up at 8 a.m., I wanted to sleep more, but I couldn’t because: Let’s see what’s going on with the 10 workers. I’ve got like three different projects going right now.

Shortly before the interview, Tan released his Claude Code setup, called “gstack,” as an open-source project on GitHub. The system includes a collection of reusable “skills” — reusable prompts stored in “skill.md” files that guide AI behavior across roles such as CEO, engineer, and code reviewer.

Currently the gstack GitHub repository lists 13 skills, but Tan continues to tweet about new updates.

I’ve been having such an amazing time with Claude Code, I wanted you to be able to have my exact skill setup,

he wrote on X.

The project quickly gained traction, attracting nearly 20,000 GitHub stars and thousands of “forks”, while also trending on Product Hunt. However, it also sparked criticism after Tan claimed a CTO friend described it as “god mode” for identifying a security flaw.

Some developers dismissed the project as overly hyped. Critics argued it amounted to little more than a set of prompts, noting that many engineers already use similar workflows.

The youtube video “AI is making CEOs delusional” by Vlogger Mo Bitar is one example of the many critics.

Despite the backlash, AI systems themselves responded positively when asked to evaluate gstack. ChatGPT described it as “reasonably sophisticated prompt workflows” and highlighted the value of simulating an engineering team structure. Gemini called it a “Pro” configuration that improves correctness, while Claude praised it as “a mature, opinionated system built by someone who actually uses it heavily.”

In a follow-up post, Tan reiterated his enthusiasm for AI coding, writing,

I took modafinil just to stay awake longer to be able to turn the momentary crystalline structures I had in my brain into lines of code before sleep or human distraction turned it to grains of sand. I love coding but I love coding with AI even more. I speak it listens and we create. I see the structure and it is built. There is no more powerful an experience to me than that.

Image: Garry Tan LinkedIn Profile

NVIDIA has unveiled a new “space computing” initiative aimed at bringing AI processing to satellites, orbital data centers and autonomous spacecraft. The company says the effort will extend accelerated computing beyond Earth-based infrastructure and enable “data-center-class performance” in space-constrained environments for geospatial intelligence, real-time sensing and mission autonomy.

At the center of the announcement is the NVIDIA Space-1 Vera Rubin Module, which the company says is built for space-based AI inference. NVIDIA claims the Rubin GPU on the module can deliver “up to 25x more AI compute for space-based inferencing” than the NVIDIA H100 GPU, positioning it for orbital data centers, geospatial processing and autonomous mission operations. NVIDIA also highlighted its IGX Thor and Jetson Orin platforms for edge AI in orbit, and said its RTX PRO 6000 Blackwell Server Edition GPU can accelerate ground-based geospatial analysis by “up to 100x” versus legacy CPU-based batch systems.

NVIDIA CEO Jensen Huang framed the move in typically expansive terms. “Space computing, the final frontier, has arrived,” Huang said in the release.

As we deploy satellite constellations and explore deeper into space, intelligence must live wherever data is generated.

He added that

AI processing across space and ground systems enables real-time sensing, decision-making and autonomy,

turning “orbital data centers into instruments of discovery and spacecraft into self-navigating systems.”

The announcement comes as the commercial space sector pushes for more on-orbit processing instead of sending every workload back to Earth. In practice, that means running AI models closer to where imagery, sensor and communications data is generated — a shift NVIDIA is betting will matter as satellites collect larger volumes of data and missions demand faster decisions. NVIDIA explicitly said the technology is aimed at orbital data centers, geospatial intelligence and autonomous space operations.

A group of space companies is already lining up behind the effort, at least on paper. NVIDIA said Aetherflux, Axiom Space, Kepler Communications, Planet Labs PBC, Sophia Space and Starcloud are using its accelerated computing platforms for next-generation missions across orbital and ground environments.

Their quotes sketch out the broader pitch. Aetherflux CEO Baiju Bhatt said the Space-1 Vera Rubin Module enables “high-performance, energy-efficient AI at the edge in orbit,” while Kepler Communications CEO Mina Mitry said Jetson Orin will let the company “intelligently manage and route data across our constellation.” Planet cofounder and CEO Will Marshall said NVIDIA’s platform is helping the company move “from raw pixels to actionable insights in near real time.” Starcloud CEO Philip Johnston, meanwhile, said his company aims to bring “hyperscale-class AI computing to orbit.”

NVIDIA is also tying the space push to geospatial intelligence on Earth. The company said growing volumes of orbital data will still need to be combined with “hundreds of petabytes of historical archive on Earth” for large-scale analysis, and argued that GPU-accelerated systems can improve response times for disaster response, climate and weather modeling, and infrastructure monitoring.

The immediate availability picture is mixed. NVIDIA said IGX Thor, Jetson Orin and the RTX PRO 6000 Blackwell Server Edition GPU are available today, while the Space-1 Vera Rubin Module will be available “at a later date.”

Image: NVIDIA 

For NVIDIA, that matters far beyond headline optics. Blackwell is at the center of the company’s current AI server push, while Rubin is being positioned as the next major step in its hardware roadmap. On stage, Huang framed the shift as a reflection of how fast demand has moved in just a few months, as hyperscalers, cloud providers and AI model companies continue racing to lock in more compute.

Now, I don’t know if you guys feel the same way, but $500 billion is an enormous amount of revenue,

Huang said during the keynote.

Well, I’m here to tell you that right now where I stand — a few short months after GTC DC, one year after last GTC — right here where I stand, I see through 2027, at least $1 trillion.

The number lands at a moment when NVIDIA is trying to show that its growth story extends well beyond one blockbuster chip cycle. Rubin, which Nvidia had previously described as its next-generation AI architecture, is expected to outperform Blackwell significantly on both training and inference workloads, with production ramping in the second half of 2026. That makes Huang’s trillion-dollar claim not just a forecast about demand, but a statement about how central NVIDIA expects its future platforms to remain in the economics of AI.

Image: Keynote GTC, Screenshot

The legal action comes after Defense Department officials designated Anthropic a supply chain risk, citing national security concerns. The move followed a statement by the company’s CEO, Dario Amodei, who said Antropic would not permit Claude’s AI model to be used for autonomous weapons, or for surveillance of U.S. citizens.

According to the lawsuit, the administration’s decision effectively places the AI company on a blacklist that blocks Pentagon suppliers from using Claude. That’s an attempt to punish the company over its AI guardrails.

Anthropic at a Crossroads: Pentagon Tensions, $380 Billion Valuation and the Future of AI

The federal government retaliated against a leading frontier AI developer for adhering to its protected viewpoint on a subject of great public significance — AI safety and the limitations of its own AI model — in violation of the Constitution and laws of the United States,

according to Anthropic, also adding that Trump officials “are seeking to destroy the economic value created by one of the world’s fastest-growing private companies.”

The supply-chain risk designation came after a meeting in February between Defense Secretary Peter Hegseth and Anthropic’ CEO Dario Amodei. According to national security experts, such a label is usually reserved for foreign adversary contractors that could pose a threat to U.S. interests, which makes the use of the blacklist against an American company highly unusual.

Following the designation, Donald Trump made a social media post stating that all federal agencies would stop using Anthropic’s AI tools.

While Anthropic was the first AI frontier lab used by U.S. officials on classified networks since the feud began, Pentagon officials have said Elon Musk’s xAI and OpenAI’s ChatGPT have now been cleared for use in classified systems.

Despite Anthropic’s strong resistance against the administration on lethal weaponry and mass surveillance, the company notes in it’s lawsuit that since 2024 it has collaborated with national security contractors, such as Palantir, to support the government in operations. Some of these activities include “rapid processing of complex data, identifying trends, streamlining document review, and helping government officials make more informed decisions in time sensitive situations.”

Image: Flickr/World Economic Forum/ Sandra Blaser; Edited – 10.03.2026

Image: U.S. Department of Defense / Chad J. McNeeley (Public Domain), via Wikimedia Commons. – 10.03.2026

During testimony, The New York Times quoted Musk downplaying the impact of his social media activity:

I tweet what’s on my mind and the market decides if it’s material.

The newspaper also quoted him adding that

If this was a trial about whether I made stupid tweets, I would say I’m guilty.

The dispute centers on whether Musk’s statements were misleading and intended to influence the stock during negotiations. The acquisition ultimately closed at $54.20 per share, valuing the transaction at about $44 billion, after months of legal wrangling.

Image: Flickr/World Economic Forum / Ciaran McCrickard/Edited– 06.03.2026

The outgoing General Manager, Georgi Ganev, continues his career in a new international role within the company, taking on the position of General Manager Data & AI Central Eastern Europe Territories. Georgi Ganev is one of the most recognizable leaders of IBM Bulgaria and in the country’s IT sector in recent years. He has played a key role in the development of the local office, the business partner ecosystem, the expansion of the business portfolio, and the positioning of IBM as a strategic partner for both the business community and the public sector. His transition to an international role reflects the high appreciation the global organization has for his long-standing experience and achievements.

Images: knowbox

According to Reuters, Mr. Altman told attendees that the rapid expansion of advanced A.I. systems will require “staggering amounts of energy,” adding that breakthroughs in clean and abundant power are essential to sustain progress.

Compute is the limiting factor for A.I.,

he said, noting that the industry must prepare for energy demand on a scale few policymakers currently anticipate.

Mr. Altman, who leads OpenAI, has increasingly linked the development of frontier models to physical infrastructure. At Davos, he argued that while model capabilities continue to advance, the ability to deploy them widely will depend on reliable, scalable electricity generation.

Why It Matters Now

Mr. Altman’s remarks come amid a global race to secure semiconductor supply chains and build hyperscale data centers. Utilities in parts of the United States and Europe have reported record requests for new grid connections from technology companies constructing A.I.-optimized facilities. In some regions, regulators have raised concerns about transmission bottlenecks and grid stability.

The challenge is not only total generation but reliability. A.I. training clusters require uninterrupted, high-density power. While renewable sources such as wind and solar are expanding rapidly, their variability creates integration challenges unless paired with storage or firm generation.

At Davos, Mr. Altman suggested that long-term solutions could include advanced nuclear technologies and other forms of high-output clean energy. He has previously invested in energy ventures, arguing that abundant power is foundational to unlocking A.I.’s economic potential.

How Analysts Interpreted the Message

Energy and technology analysts see Mr. Altman’s comments as part caution, part strategic positioning. By highlighting energy as the next bottleneck, A.I. leaders may be signaling to governments that infrastructure policy is now inseparable from digital competitiveness.

Efficiency improvements are underway. New generations of A.I. accelerators deliver greater performance per watt, and researchers are developing smaller, more specialized models that reduce computational load. Techniques such as quantization and sparsity can lower inference costs significantly.

Yet many experts warn of a rebound effect: as computing becomes more efficient and less expensive, overall usage tends to increase. In that scenario, total energy consumption may continue rising even as individual tasks become more efficient.

Pushing Back on Water and Per-Query Claims

Days after Davos – last week, Mr. Altman addressed a related set of concerns while speaking at an event hosted by The Indian Express during his visit to India for a major A.I. summit.

There, he pushed back forcefully against viral claims about A.I.’s environmental footprint. Concerns about the water usage of systems like ChatGPT are “totally fake,” he said, acknowledging that water consumption had once been higher “when we used to do evaporative cooling in data centers.”

Now that we don’t do that, you see these things on the internet where, ‘Don’t use ChatGPT, it’s 17 gallons of water for each query’ or whatever,

Mr. Altman said.

This is completely untrue, totally insane, no connection to reality.

He also rejected comparisons suggesting that a single ChatGPT query consumes the equivalent of 1.5 iPhone battery charges.

There’s no way it’s anything close to that much,

he said when asked about the figure.

At the same time, Mr. Altman acknowledged that broader concerns about total energy use are legitimate. It is “fair” to worry about

the energy consumption — not per query, but in total, because the world is now using so much AI,

he said. In his view, that reality strengthens the case for accelerating investment in nuclear, wind and solar power.

The world needs to move towards nuclear or wind and solar very quickly,

he added.

The Data and the Debate

There is currently no global legal requirement for technology companies to disclose detailed energy and water consumption figures for specific A.I. workloads, leaving researchers to estimate impacts independently. Some academic studies have linked large data center clusters to localized increases in electricity demand and, in certain cases, higher wholesale power prices.

The International Energy Agency estimates that global data center electricity consumption could exceed 1,000 terawatt-hours annually by the end of the decade, roughly equivalent to Japan’s total electricity use today. A.I.-specific workloads are projected to account for a rising share of that total.

Mr. Altman has argued that public discussions often frame the issue unfairly, particularly when they compare the energy required to train a large A.I. model with the energy needed for a single human task.

Many discussions about ChatGPT’s energy usage are unfair,

he said in India, especially when they focus on

how much energy it takes to train an AI model, relative to how much it costs a human to do one inference query.

But it also takes a lot of energy to train a human,

he added.

It takes like 20 years of life and all of the food you eat during that time before you get smart. And not only that, it took the very widespread evolution of the 100 billion people that have ever lived and learned not to get eaten by predators and learned how to figure out science and whatever, to produce you.

In his telling, the more appropriate comparison is between a trained A.I. system answering a question and a human doing the same task.

If you ask ChatGPT a question, how much energy does it take once its model is trained to answer that question versus a human?

he said.

And probably, A.I. has already caught up on an energy efficiency basis, measured that way.

What Can Be Done

Policy experts point to several responses. Accelerating grid modernization, streamlining permitting for new transmission lines and investing in next-generation clean energy projects could help absorb the surge in demand. Co-locating data centers with dedicated renewable or nuclear facilities is another approach already being explored by major technology firms.

There is also a growing push for “energy-aware A.I.” — designing models and systems that optimize not only for accuracy and speed, but also for power efficiency.

At Davos, Mr. Altman framed the challenge as an opportunity rather than a deterrent. If artificial intelligence can dramatically increase productivity and scientific discovery, he suggested, then investing in abundant clean energy may be one of the most consequential economic decisions of the decade.

In tying A.I.’s trajectory to the electrical grid, Mr. Altman underscored a broader reality: the digital revolution now rests squarely on physical infrastructure. And the next frontier in artificial intelligence may depend less on algorithms than on kilowatts.

Material by Iva Abadjievа

Image: The Indian Express Youtube Channel, “Sam Altman Unfiltered: ChatGPT, AI Risks & What’s Coming Next, 40 Questions in 60 Minutes”

Image: Sam Altman at World Economic Forum from Benedikt von Loebell at Flickr