XTHE

After nearly three years, Intel is returning to the desktop workstation with the anticipated Granite Rapids WS series, now called Xeon 600. The new series covers all of Intel’s bases for desktop workstations with the previous-gen Sapphire Rapids. WS/Intel split its products between two lineups: Zeon W2500 and Zeon W3500, but all Granite Rapids workstation chips will use the same Xeon 600 branding.  

Granite Rapids has been in the data center for about 1.5 years. During that time, AMD introduced its Zen 5-based Threadripper 9000 chips, leaving an open slot for Intel to release its next-generation workstation CPUs. Intel has eleven SKUs for Xeon 600, five of which will be available as boxed models for retail sale. Intel hasn’t provided a firm release date for the CPUs, but says the new W890 motherboards and systems from brands like Dell, Lenovo, Supermicro, and Puget will be available starting in late March.  

Compared to refreshed Sapphire Rapids WS parts, Intel claims Xeon 600 delivers up to 9% better single-core performance and up to 61% higher multi-threaded performance. The higher multi-threaded performance comes from much higher code numbers on 8.86 cores and the Xeon 696X with 64 cores.  

Like its data center counterpart, Xeon 600 chips use the Redwood Co. micro-architecture that debuted in Intel’s Meteor Lake mobile chips. Here, the core count is scaled up and, following Intel’s new split of Xeon CPUs into heterogeneous architectures, the Xeon 600 CPUs use only P-cores with hyper-threading.  

The full list of Xeon 600SK ranges from $499 for the 12-core Xeon 634 up to $7,699 for the flagship Xeon 698X. This split largely mirrors AMD’s approach with Threadripper 9000 chips, where the main range tops out at 64 cores, while the Threadripper Pro 9000 WX range reaches 96 cores.  

Short of the bottom three SKUs, some platform features are consistent across the Xeon 600-range X-Series SKUs:  

• They are unlocked for overclocking.  

• All octa-channel memory supports speeds up to 6400.  

• You also get 128p CLE 5.0 Lens, CXL 2.0 Support, and an intense AMX accelerator in each CPU core, which now supports FP16 instructions.  

The chips support up to 40 TB of memory, doubling what’s available on AMD’s Threadripper 9000WX range and quadrupling support compared to the base Threadripper 9000 range. Even an A1 TB kit of DDR5 6400 RDIMMs costs about $28,000 right now, bringing support for MRDIMMs(multiplexed rank DIMMs).   

MRDI MMS are supported in the data center with Xeon 6 CPUs, but this is the first time we’re seeing them in workstation chips. MRDIMMs include two ranks of memory chips, each with a multiplexer chip. The multiplexer combines the bandwidth of both ranks, doubling the transfer rate. In the case of Xeon 600, with support for up to 8000 MT/s.  

MRDI-MMs use the same physical connector as RDI-MMs, so they fit into existing connectors if the CPU supports MRDI-MMs. MRDI-MMs are only relevant in HPC settings where memory bandwidth build is critical, so Intel isn’t supporting them across the full stack. They are only supporting the top 5 SKUs, starting from the Xeon 678X.  

Intel broadly claims a 9% improvement in single-core performance with the Xeon 600 series compared to the Xeon W2500 and Xeon W3500, according to Intel. These numbers come from Cinebench 2026 scores, as you can see. The disclaimers and configurations are included at the end of the gallery below.  

Looking at the spec workstation for Intel, it says the flagship Xeon 698X offers:  

• a 17% improvement in AI  

• 22% in the energy sub-category  

• 61% in financial services  

• 19% in life sciences  

• 10% in media and entertainment  

compared to the Xeon W9-3595X. In the productivity category, the Xeon 698X boasted identical performance, while in product design, it showed an undisclosed regression.  

In specific apps, Intel says the Xeon 698X finished the Blender Junction Render 74% faster than the Xeon W-3595X and sped up AI-powered upscaling with Topaz Labs Video Upscaler by 29%. Intel attributes the latter speedup to the AMX accelerators in the Xeon 600-core chips. To that end, Intel is introducing Open Image Denoise 2.4, which it says is accelerated by FP16 instructions, Available in Xeon 600 AMX.  

In development, data analysis, and AI interpretation:  

• 24% better linear algebra performance (as measured in algorithms in Intel’s form of NumPy/SCIPY)  

• 18% faster large data set analysis with spec workstation 4s data science workload  

• 16% faster AI inference with spec workstation 4s ONNX inference test  

In particular, Intel didn’t share any competitive benchmarks for Threadripper 9000 during a press Q&A session. Intel’s Jonathan Patten said, “We’re looking to be very competitive within the market, offering better performance per dollar for more value for the workstation spend. This is a very highly expandable platform. We have up to 4 TB of memory capacity, supporting two DI/MMs per channel. Our competitors do not. We have advanced instruction sets, AMX, as we mentioned a little bit, and our vPro technology, so we continue to offer a very competitive platform.”  

Hopefully, we’ll have those comparisons soon. Intel says Xeon 600 motherboards with the W890 chipset will launch in late March, as will workstations from OEMs like Dell, Lenovo, and Supermacro. We still haven’t learned about a firm release date, let alone a release window, for boxed Xeon 600 chips.

Source: Intel returns to boxed workstation CPUs with Xeon 600 — Granite Rapids WS delivers up to 86 cores, 4TB of memory, and 128 PCIe 5 lanes 

In mid-February, Amazon sellers received a policy update email, but most people just skimmed it or paid little attention.  

But overlooking it could be a big mistake.  

The updated Business Solutions Agreement, which takes effect on March 4, 2026, is one of the biggest changes Amazon has made to its seller contracts in years. Its impact goes far beyond just your automation tools.  

Let’s break down what’s changed, who needs to pay attention, and what steps you should take.  

What Did Amazon Actually Change 

The new BSA makes official what Amazon has been moving toward for some time: definite legal control over any AI or automated system that interacts with its marketplace.  

The main change is a new agent policy in Section 19. This policy creates a category called Agent for any automated software or AI system that uses Amazon services, such as Seller Central, advertising APIs, or third-party tools linked to your account. Now three rules apply to all of these systems.  

Every agent must identify itself as an automated system when it connects to Amazon. Each agent has to follow the agent policy at all times. Also, every agent must stop accessing Amazon right away if Amazon asks.  

The last rule is often called the kill switch provision. Amazon can tell any automated system to stop working on its platform. If your tools don’t shut down when told to, you could run into compliance issues.  

In addition, Section 4.2 now limits the use of Amazon materials or services for developing or improving AI and machine learning models. This includes reverse engineering, data mining, and creating source code or model parts. Amazon is making it clear that its platform is not meant for training outside AI systems.  

What Does This Mean for Your Tech Stack? 

If you use repricers, PPC automation, listing optimization software, or internal scripts that connect to Seller Central or Amazon APIs, all of these are now covered by this policy. The real question for most brands isn’t whether your tools are included, but whether your vendors have updated their systems to comply with the new rules.  

You should check that every third-party tool connected to your Amazon account can identify itself as an automated system when it logs in, and that you or the vendor can turn off access right away if Amazon asks. Some vendors are already prepared, but others are not. It is better to talk to them now than wait until there is a problem.  

So if you keep selling on Amazon after March 4, you are automatically agreeing to these changes. You don’t need to sign anything new. The agreement takes effect by default.  

The Generative Search Angle (Most people are missing this part) 

There’s another part of the Amazon BSA update that isn’t getting much attention, but it’s important for brands focused on generative search visibility.  

Tools like ChatGPT, Perplexity, and Gemini have always had trouble getting product data from Amazon, even before this update. Amazon used robots.txt rate limits and other technical blocks to stop AI crawlers from scraping product pages. The new BSA makes this official and shuts down any backdoors that could have been used to access data.  

Generative search engines don’t have direct up-to-date Amazon data at scale. When someone asks ChatGPT to recommend the best protein powder or compare running shoes, the answers come from sources beyond Amazon’s listings. This is more important than many brands think.  

If your product information is only on Amazon, you’re mostly invisible to the AI systems that are now influencing buying decisions before shoppers even open their browsers.  

How To Build Generative Search Presence Whenever Amazon Is Off-Limits 

We’ve been telling clients this for a while: Don’t rely on Amazon as your only source of product information online.  

Your DTC site or brand hub should have the most complete and detailed information about every product you sell. Include:  

• Full spec tables  

• Use cases  

• Comparison content  

• Buying guides  

• FAQ sections that answer real customer questions  

Amazon’s listing rules won’t let you share all this there, so your own website is the best place for it.  

Schema markups are more important than ever. Structured data like product info, reviews, FAQs, and how-to guides makes it easier for a generative system to read, summarize, and reference your content. If you haven’t checked your DTC site’s schema recently, now is a good time to do so.  

Being on multiple channels helps, too, if your products are listed on Walmart, Target, specialty retailers, and Amazon. Generative systems can get a better sense of your products without relying on Amazon data. The more places your brand appears online, the more accurately AI can represent you.  

The main idea is that Amazon will remain on your primary sales channel, but your product information shouldn’t live only there. Relying solely on Amazon has always been risky, and this BSA update underscores that risk.  

Amazon AI Agent Compliance Checklist 

Most brands don’t need to panic, but a few actions are worth completing before the deadline:  

1. Audit every third-party tool that accesses your Amazon account and confirm it meets the new agent requirements.  

2. Contact your major vendors and ask directly whether they are compliant.  

3. With the March 4 BSA update, verify that the Kill Switch capability exists for each tool, either on the vendor side or your own.  

If you use any internal information, scripts, or custom integrations, review them with someone who understands the new rules in Section 19. The definition of an agent is broad, so even standard fulfillment automation counts.

Source:Amazon’s New AI Rules Go Live March 4. Here’s What They Mean for Your Business 

Takeaways 

• For the 2026 U.S. elections, we are using the same thorough measures we have relied on for years. We support free expression, encourage voter participation, and are transparent about AI-produced content.  

• Like in past election cycles, we will keep blocking new political ads on our platforms during the last week of the campaign.  

• We are still running our election operations center and focusing on protecting people from scams on our platforms, especially those that misuse images of politicians.  

This November, millions of Americans will vote in the U.S. elections. Every election is different, but Meta has built a thorough approach for our platforms. We support free expression, encourage civic participation, and offer strong transparency for political and social issue ads.  

As we prepare for the 2026 elections, we are explaining how our established policies and safeguards will operate on our platforms.  

Industry-Leading Transparency Around Political Ads 

Advertisers who want to run ads about elections, politics, or social issues must go through an authorization process and add a “Paid for by” disclaimer. We have kept these ads in our public ad library for seven years. Right now, our ad library has over 18 million U.S. entries.  

Restriction period column as before. We will block new political, electoral, or social issue ads during the last week of the US election campaign. Ads that have already shown at least once before this period can keep running. We do this because in the final days of an election, there may not be enough time to challenge new claims in ads.  

Political ads created or edited using AI: advertisers must tell us whether they use AI to create or change ads about social issues, elections, or politics. In certain cases, when they do, we add this information to the ad and our ad library. We are always improving our labeling so people can more easily spot ads that may have been edited or made with AI.  

Identifying Artificial Intelligence-Generated Content Beyond Political Ads 

Use industry standards and technologies, such as C2PA, to spot content created or altered with AI outside ads on our platforms. We label this content to keep things transparent. If we detect that the content was made by an AI tool, we show an AI info label and explain if the label comes from industrial signals or self-disclosure. People must use this label when posting organic content that includes photorealistic video or realistic audio that was digitally created or altered. If they do not, we may apply penalties.  

Continuing To Connect People with Information From State And Local Officials 

Voting Information and Election Day Reminders: As in the past years, we show notifications at the top of Facebook and Instagram feeds to connect people with local and state voting information, including during primaries. On Facebook, these notifications are now easier to understand because we show them the input the users choose a language and another language if they interact with content in that language. If someone has moved or we get their location, they can select “Change state” to get the correct government information.  

Search Engine Results: When people search for election-related ITAC terms on Facebook, we show results that link to state government websites for more information.  

Voting Alerts: On Facebook, we keep working with state and local election officials to send timely voting alerts about registering and voting to people in their communities. State and local officials have sent over 1 billion notifications through Voting Alerts on Facebook.  

Instagram stickers: On Instagram, we continue to highlight stickers on stories that guide people to official voting information before registration deadlines and Election Day.  

Empowering the community to add expanded context 

Last year, we launched Community Notes so people can add context to posts on Facebook, Instagram, and Threads that might be misleading or confusing. Unlike terrestrial third-party fact-checking, Community Notes are written and rated by contributors, not by Meta or a small group of fact-checkers. Notes can be added to most public content, including posts by Meta, our executives, public figures, and politicians. Contributors decide which content could use greater context. A note is only published if it reaches consensus, which means contributors who usually disagree on it. This helps reduce bias and improve the quality of published notes. Anyone in the US can sign up to be a contributor if they are over 18, have an account that is more than 6 months old and in good standing, and either have an authenticated phone number or use two-factor authentication.  

Combating Scams 

Focusing on protecting people from scams, including those that misuse politicians’ images, we consistently review and update our approach. Our policies do not allow ads that use images of public figures to scam or defraud people. We continue to invest resources to fight these fraudulent ads and are continually improving our enforcement, including suspending and deleting accounts, pages, and ads that break our rules.  

Fraudsters use elections to trick people into engaging with content that appears to be political campaigns. To help people spot and avoid these threats, we have created resources in our Scam Prevention Hub. We also use facial recognition technology to detect and stop celeb bait ads that use images of public figures to defraud people. We remove imposter accounts that are reported for impersonating others.  

Operational Readiness: We have invested over $30 billion in safety and security over the past decade, including efforts to protect elections on Facebook, Instagram, and Threads. We enforce our rules against voter interference, electoral violence, and false information about when, where, and how to vote. We remove content that breaks our rules as soon as we find it. We also remove content that could cause physical harm or disrupt elections.  

They are also continuing to expose and disrupt foreign influence operations, including those targeting elections, and have identified 200 networks of coordinated inauthentic behavior since 2017. Our Election Operations Center, which brings together subject matter experts across Meta, including our threat intelligence, data science, engineering, research, operations, and legal teams for real-time monitoring to tackle potential abuse flowing across our network, as we have since 2018. This is for elections around the globe, including the Brazilian presidential election.

Source: How Meta Is Preparing for the 2026 US Midterm Elections 

Microsoft is quietly building a new Canvas-style workspace for Co-Pilot, and leaked screenshots suggest it is a full-fledged AI-powered whiteboard.  

As posted on x by Windows enthusiast Walking Cat, the feature/app is internally referred to as Project Firenze; however, the leaked interface shows the name Co-Pilot Canvas.  

Co-Pilot Canvas appears to be a web-based environment where users can create and manage canvases drawn with digital ink tools and interact with content in a free-form layout, similar to the existing Microsoft Whiteboard app.  

The landing screen shows a prompt to create your initial canvas to start drawing and taking notes. Like the Microsoft Whiteboard app, Co-Pilot Canvas can automatically save your work.  

We found references to both development and production of Azure endpoints, which suggest Co-Pilot Whiteboard is being actively tested internally and is not a static mockup. There is also a generic-looking logo, but we are not sure if this is the final version.  

Despite no major updates recently, the Microsoft Whiteboard app remains a fully functional collaborative tool, so it is unclear whether Microsoft is replacing it with Co-Pilot Canvas.  

CoPilot Canvas Integrates AI Image Generation, Streaming, And Advanced Feature Controls 

As expected, Co-Pilot whiteboard will rely on AI as its main differentiator compared to the original Microsoft whiteboard. Seven developer-style options point to a system designed for instant AI interaction.  

One of the most telling switches is labeled “Create with AI streaming”, suggesting the canvas may support live generative responses as you draw or type instead of waiting for a completed prompt. Copilot Whiteboard may generate diagrams, layouts, or visual elements incrementally as you work, similar to brainstorming with an assistant that updates the board in real time.  

Another menu shows an image model selector with options such as GPT-4o Image Gen (default), GPT-4o Image Gen 1.5, and GPT Image 1.5, which are not the latest models. However, the presence of multiple selectable models shows that the Co-Pilot canvas can handle multi-modal generation directly in the workspace.  

Auto-naming for Canvas titles could be useful for collaborative work during or after a meeting. Co-Pilot Whiteboard may examine the content of a board and produce a meaningful name automatically.  

The Co-Pilot Canvas app also reveals a long list of AI-related configuration panels under developer mode, including:  

• Debug gates/AI gates.  

• Meeting summary  

• One Shot Grounding  

• Post grounding  

• Intent detection  

• Solve Math  

• Delegate actions to Augloop and hand off actions  

These are not typical whiteboard features. They support agent-style behaviours in which the AI can reason over content, summarize discussions, interpret intent, and potentially trigger additional actions, which aligns with Microsoft’s strengths.  

Co-Pilot Canvas could bring AI to help with brainstorming on whiteboards. 

Today, most interactions with AI still take place in a chat box. Co-Pilot cameras may be placed closer to a visual workspace where users can collaborate, map, and execute ideas with Co-Pilot’s help.  

Although many modern canvas-style apps like Notion, Visual Pages, Figma, Miro, and Microsoft Whiteboards exist, Microsoft could be in a unique position to bring AI into the whiteboard environment because it has direct access to enterprises.  

Co-Pilot whiteboard could enable workflows where teams can:  

• sketch  

• draft documents  

• generate images  

• summarize discussions  

• trigger actions  

Microsoft may also be considering portable workspaces with Co-Pilot whiteboard, since there are options to export and import .canvas files. This could allow teams to share AI-assisted canvases, as they do with documents today.  

That being said, everything about Project Forensic looks like an early developer toggle feature, and internal endpoints point to something in testing rather than something prepared for release.  

Microsoft has not made any public announcement about a possible replacement for Microsoft Whiteboard or any roadmap. We will update as soon as Microsoft makes Copilot Canvas official.

Source: Microsoft Copilot Canvas leak reveals an AI-powered Whiteboard with image generation, AI streaming, and more 

News summary 

• Top operators and infrastructure providers such as Booz Allen, BT Group, Cisco, Deutsche Telekom, Ericsson, Maitreya, Nokia, ODC, SK Telecom, SoftBank Corp., and T-Mobile plan to use open, trusted software-defined wireless platforms.  

• This commitment adds to NVIDIA’s ongoing work with industry and governments in Europe, Japan, Korea, the UK, and the US to push forward AI-native 6G innovation.  

At Mobile World Congress, NVIDIA announced a joint commitment with Booz Allen, BT Group, Cisco, Deutsche Telekom, Ericsson, Maitreya, Nokia, ODC, SK Telecom, SoftBank Corp., and T-Mobile to build the next generation of wireless networks using AI-native, open, secure, and trustworthy platforms.  

This effort demonstrates a shared commitment to making 6G infrastructure, which will serve as a foundation for future worldwide connectivity, open, intelligent, resilient, and able to drive innovation while protecting global trust.  

6G wireless networks will go beyond basic connectivity and become the backbone for physical AI. They will support billions of self-driving machines, vehicles, sensors, and robots, requiring much higher levels of security and trust. Older wireless systems were not built for these needs, so new challenges will arise as networks become more complex.  

To address these problems, NVIDIA is uniting the industry to develop AI-native software-defined wireless platforms based on open trusted principles by adding AI throughout the radio access network, edge, and core 16 network. We’ll need to provide:  

• secure sensing and communication  

• smart decision making  

• support for interoperability  

• supply chain resilience  

• faster innovation  

AI is changing computing and driving the largest infrastructure expansion ever. Telecommunications is the next step, said Jensen Huang, founder and CEO of NVIDIA, with a global group of industry leaders in media building AI RAN to run telecom networks into AI infrastructure everywhere.  

Uniting On Openness And Trust For The AI-Native Software-Defined Era Of Connectivity 

6G will be AI-native and software-defined, allowing wireless networks to keep up with rapid innovation. Networks built on AI/RAN architecture will evolve through software, providing real-time intelligence and quick progress. This change will welcome a wide range of participants, including global operators, tele-technology providers, start-ups, researchers, and developers, all working together on open and programmable platforms.  

Alison Kirkby, Chief Executive of BT Group, said: “Connectivity is the backbone of economic growth, and with this cooperation, we are helping lay the foundations for a future ecosystem that is intelligent, sustainable, and secure by building on open and trustworthy AI-native platforms. We can simplify future technologies like 6G, making sure they build upon the strengths of today’s 5G networks while still unlocking powerful new capabilities at scale.”  

Tim Hottges, CEO of Deutsche Telekom AG, said: “Best network, best customer experience: that remains our promise. With an open, intelligent, and trusted 6G infrastructure, we are laying the foundation for the era of physical AI and bringing new value to our customers for industry and for society.”  

Arielle Roth, Assistant Secretary of Commerce for Communications and Information and Administrator at the National Telecommunications and Information Administration, said America’s 6G leadership will be critical to our nation’s global economic prosperity, national security, and global competitiveness. Today’s announcement demonstrates that the United States and our allies and partners around the world are leading in this next-generation technology. We look forward to the next steps from this International Industry Coalition as they advance and implement their common 6G vision.  

Jung Jai-hun, President and CEO of SK Telecom, said, “SKT is evolving telco infrastructure to serve as the foundation for the AI era, in which connectivity acts as a platform for intelligence and innovation. Together, we can build an open, trusted infrastructure that drives a global ecosystem of AI innovation.”  

Hideyuki Tsukuda, Executive Vice President and Chief Technology Officer of SoftBank Corp., said 6G will transform wireless networks into secure, software-defined infrastructure that supports the next wave of global innovation. SoftBank Corp. is driving this advancement with NVIDIA by advancing open and trusted platforms that enable interoperability, resilience, and steady evolution at scale.  

Inigo Palan, CEO of T-Mobile, said we are at a key moment. In the US, we’ve laid the foundation for advanced, AI-native 5 G networks, where intelligence lives within the network. As 6G becomes the backbone of the AI era, telecom will serve as the nervous system of the digital economy, enabling autonomous and intelligent industries at scale and unleashing new value for customers and enterprises alike. T-Mobile is proud to help define what’s next through deep ecosystem collaboration and sustained innovation.  

A Common Vision for 6G: Open Software-Defined AI-Native 

NVIDIA participates in global private and public initiatives to advance 6G innovation, contributing open-source software, accessible platforms, and joint research and development projects:  

• In the United States, NVIDIA has joined the Future-G office-led OCUDU initiative, collaborating with government and industry partners to accelerate open, software-defined, and AI-native 6G architectures.  

• NVIDIA is the founding member of the AI/RAN Alliance, which now has over 130 participating companies driving AI/RAN innovation.  

• NVIDIA, along with Booth Allen, CISCO, T-Mobile, MITRE, and ODC, launched the AI-native wireless networks (AI-WIN project) in October, an all-American AI RN stack to accelerate the pack to 6G.  

• In Korea, NVIDIA is collaborating with an industry consortium to help shape intelligent, secure, programmable 6G networks from the ground up.  

• In the UK, NVIDIA is collaborating with the Department of Science, Innovation, and Technology to advance applied research, ecosystem development, and trusted AI-native network design.  

• Across Europe and Japan, NVIDIA is actively engaged with public and industry programs created to strengthen open innovation, interoperability, and trusted infrastructure.  

Together, these joint efforts show a unified commitment, supported by like-minded governments, operators, and technology partners, to create secure, intelligent, and trusted worldwide connectivity for the next generation of wireless technology.

Source: NVIDIA and Global Telecom Leaders Commit to Build 6G on Open and Secure AI-Native Platforms 

Highlights 

• Google has finished rolling out its February Discover Core update.  

• For the first time, Google has publicly called a core update for a discovery core update.  

• The rollout ended on February 27 after about 22 days.  

According to the Search Status Dashboard, Google’s February 2026 Discovery Core update finished rolling out at 2:02 a.m. PT on Feb 27.  

The rollout began on February 5 and lasted about 22 days, about 8 days longer than Google’s original estimate of up to 2 weeks.  

Google announced the update on the Search Central blog and noted that this was the first time it publicly labeled a Core update as a Discovery Core update.  

At launch, Google outlined three main goals for the update:  

1. Show users more locally relevant content from sites in their country.  

2. Reduce exaggerated or clickbait content.  

3. Highlight more in-depth, original, and timely content from expert sources.  

The update first rolled out to English-speaking users in the U.S. Google plans to expand it to all countries and languages in the coming months but has not given a specified timeline.  

Insights From Third-Party Data 

Early third-party tracking gives a first look at what changed during the rollout.  

NewzDash released a scoreboard comparing the pre-update period, January 25-31, and post-update period, February 8-14, for the top 1000 domains and articles in the U.S., California, and New York.  

Earlier this week, we reported that the data revealed three main patterns.  

NewzDash’s data suggests that regional personalization increased. New York’s local domains appeared roughly five times more often in the New York field than in the California field, and the reverse was true for California’s local domains. The fields still share most of their top hundred items, but each state now gets a meaningful local layer on top of that national core.  

Newer domains are now getting top placements in the US. Unique domains in the top 1000 fell from 172 to 158. After the update, California saw a similar drop from 187 to 177. New York was the exception, with unique publishers remaining about the same, while publisher diversity shrank. Unique content categories increased across all three geographic views, while unique domains decreased. That suggests Discover is covering more topics but concentrating on top placements among a narrow set of publishers.  

News Dash also found that posts from institutional accounts on x.com rose from 3 to 13 in the US top 100 Discover placements. Most of these came from well-known media brands posting on X. News Dash has tracked x.com’s growth on Discover since November 2025, and this update appears to have accelerated the trend.  

Wider Context 

This Core update comes as Discover’s importance as traffic sources continue to grow.  

A study of more than 400 news publishers found that Discover’s share of Google-driven traffic almost doubled in two years, rising from 37% in 2023 to about 68%. At the same time, traditional web-search traffic to news publishers fell from 51% to around 27%.  

While this data does not show why Google changed Discover’s score, it does show that a Discover-only core update is important. When a platform sees so much traffic for publishers, any changes to content can have real effects on revenue.  

The roll-out is complete, so US sites can now compare their Discover performance in Search Console for both the pre-update and post-update periods. Google suggests waiting at least a week after a core update finishes before making any conclusions and recommends comparing data from before the update started. Publishers with strong regional relevance and clear topic focus may have benefited. At the same time, those without topic-level authority may have lost ground. Discover covered more topics in the post-update window, but fewer sites were appearing in top placements in the US and California. That combination is worth monitoring as more data comes in.  

The rollout lasted about 22 days, longer than Google’s 2-week estimate. As a result, some NewsDash data was collected while the update was still happening. Examining data from after the rollout finished could reveal different trends.  

What’s next? 

Google hasn’t said whether Discover will continue to get its own Core updates going forward. This was the first time Google labeled a core update as a Discover core update, so it’s too early to know whether this will become a recurring pattern.

Source: Google’s Discover Core Update Finishes Rolling Out 

Yesterday, we signed an agreement with the Pentagon to install advanced AI systems in classified settings. We also asked that these systems be made available to all AI companies.  

We believe our agreement includes stronger safe plots than any previous deal for classified AI deployments, even compared to Anthropic’s. Here’s why:  

Three main red lines guide our work with the Department of War and other leading AI labs; we generally share these.  

• No use of OpenAI technology for mass domestic surveillance.  

• No Use of OpenAI technology to direct autonomous weapons systems.  

• No use of OpenAI technology for high-risk automated decisions (e.g., systems such as social credit).  

Some AI labs have lowered or removed their safety guardrails and now rely mostly on usage policies for national security deployments. Our approach delivers better protection against misuse.  

Our agreement protects these red lines with a broad, multifaceted approach:  

• We keep full control over our safety systems.  

• We use cloud deployment more clearly, OpenAI staff.  

• We have strong contract protections.  

These measures add to the protections already in US law.  

We are committed to democracy because this technology is so important. AI development needs to work closely with the political process. We also know our technology brings new risks, and we want those defending the United States to have the best tools available.  

Our agreement covers the following points:  

1. Deployment Architecture Cologne. This is a cloud-only setup that uses a safety system that follows these and other principles. We are not giving the Department of War any models without safety features, and we are not deploying our models on edge devices, which could be used for autonomous lethal weapons.  

Our deployment setup lets us independently check that these red lines are not crossed, including running and updating classifiers.  

1. Our contract: The key terms are as follows:  

The Department of War may use the AI system for all lawful purposes consistent with applicable law, operational requirements, and well-established safety and oversight protocols. The AI system will not be used to independently direct autonomous weapons in any case where law, regulation, or department policy requires human control, nor will it be used to assume other high-stakes decisions that require approval by a human decision-maker under the same authority. Per DoD Directive 3000.09 (dated 25 Jan/2023, any use of AI in autonomous and semi-autonomous systems must undergo strict verification, validation, and testing to ensure they perform as intended in realistic environments before deployment.  

For intelligence activities, any handling of private information will comply with:  

• The Fourth Amendment  

• The National Security Act of 1947  

• The Foreign Intelligence Surveillance Act of 1978  

• Executive Order 12333  

• applicable DoD directives that require a defined acceptance, as permitted by the Posse Comitatus Act and other applicable law.  

1. AI Expert Involvement: Clear AI OpenAI engineers will work directly with the government, and clear safety and alignment researchers will also be involved.  

FAQs 

Why are you doing this? 

First, we believe the US military needs strong AI models to support its mission, especially as potential adversaries increasingly use AI in their systems. We did not sign a contract for classified deployment right away because we felt our safeguards and systems were not ready. We have worked hard to ensure that, when classified deployment occurs, it includes safeguards to prevent any red lines from being crossed.  

I’ve never been willing to remove important technical safeguards to improve performance on national security work. We do not think that is the right way to support the US military.  

Second, we wanted to ease tensions between the Department of War and U.S. AI labs. Constructing a better future will require real collaboration between the government and AI labs. As part of our agreement, we asked that the same terms be offered to all AI labs and that the government try to resolve issues with Anthropic. The current situation is not a good way to start this next phase of working together.  

Why could you reach a deal when Anthropic could not? Did you sign the deal? Wouldn’t they? 

From what we know, our contract offers better guarantees and stronger safeguards than earlier agreements, including Anthropic’s original contract. Our red lines are more enforceable because deployment is limited to the cloud. Our safety stack works as intended and has remained in OpenAI staff R-in mode throughout.  

We do not know why Anthropic could not make this deal, but we hope they and other lands will consider it in the future.  

Do you think Anthropic should be designated as a supply chain risk? 

No, and we have made our position on this clear to the government.  

Will this deal enable the Department of War to use OpenAI models to power autonomous weapons? 

No. Based on our safety stack, our Cloud-only deployment, the contract language, and existing laws, regulations, and policies, we believe that this cannot happen. We will also have OpenAI personnel in the loop for additional assurance.  

Will this deal enable the Department of War to use OpenAI models to conduct mass surveillance on U.S. persons? 

No. Based on our safety stat, the contract language, and existing laws that heavily restrict DoW from domestic surveillance, we believe that this cannot happen. We will also have OpenAI personnel in the loop for additional assurance.  

Do you have to deploy models without a safety stack? 

No, we keep full control over the safety stack we use and will not deploy without safety guard rails. Our safety and alignment researchers will also be inward and help improve our systems over time. We know some other AI labs have reduced model guardrails and rely mainly on usage policies, but our layered approach delivers better protection against misuse.  

What happens if the government violates the terms of the contract? 

As with any contract, we could end the agreement if the other party breaks the terms. We do not expect this to happen.  

What if the government changes the law or existing D.O.W. policies? 

Our contract clearly refers to the current laws and policies on surveillance and autonomous weapons. Even if these laws or policies change in the future, our systems must still comply with the standards set out in the agreement.  

How do you address the arguments Anthropic made in their blog post (opens in a new window) about their discussion with the DOW? 

In their post, Anthropic lists two red lines. We share those two and add a third: automated, high-stakes decision-making. Anthropic explained why they did not think these red lines would be upheld in the contracts they saw from the Department of War at that time: “We think these red lines would be upheld in our contract.”  

• Mass Domestic Surveillance. In our discussion, it was clear that the Department of War sees mass domestic surveillance as illegal and did not plan to use our technology for this. We made sure our contract clearly states that this is not allowed under lawful use.  

• Fully autonomous weapons. Our contract only allows cloud deployment, which cannot power fully autonomous weapons that require edge deployment, which is not permitted.  

Along with these protections, our contract includes extra safeguards such as our Safety Stack and OpenAI technical experts who are involved throughout.

Source: Our agreement with the Department of War 

Apple has announced it will start revealing new products on X (formerly Twitter) on Monday, March 2. CEO Tim Cook posted a big week ahead. It all starts on Monday morning. He also used the hashtag #AppleLaunch, suggesting several announcements are planned before the event on March 4.  

Apple has not shared official details about what it will launch, but reports say several products are close to release. These may include the iPhone 17e, new MacBook Pro models, and a MacBook Air with the M5 chip. There are also rumors of a cheaper MacBook that might use an iPhone-level processor instead of a standard M-series chip.  

What to expect from Apple’s March for Special Experience event? 

According to a report by 9to5Mac, Apple’s March 4 event will be invite-only, with media gatherings expected in cities including New York, London, and Shanghai. The report indicates there may not be a traditional keynote presentation. Instead, Apple is likely to make product announcements online across the week.  

Here’s a closer look at the devices Apple might introduce:  

iPhone 17E 

The iPhone 17E will likely follow Apple’s approach of giving its budget phones the latest hardware. After the iPhone 16E got the A18 chip, the new model is expected to use the A19 processor, first seen in the iPhone 17 series. This should improve performance, graphics, and neural engine features.  

A report by Mako Takara suggests the device may also feature Apple’s newer C1X modem, which is said to offer faster speeds than the C1 modem used in the iPhone 16E. It could also include Apple’s in-house N1 Networking chip to handle Wi-Fi/Bluetooth and thread connectivity.  

The iPhone 17E is expected to keep its current design, with a notch rather than the Dynamic Island and a single rear camera. The phone may have slimmer bezels but will still use a 6.1-inch display. It is likely to stick with a 60Hz screen without ProMotion or Always-On features. The front camera could be upgraded to an 80MP center-stage camera for better video calls, while the back camera may remain a single 48MP sensor. MagSafe support is also expected.  

MacBook Pro with M5 Pro, M5 Max 

Apple introduced a 14-inch MacBook Pro with the standard M5 chip in October, but did not launch the higher-tier Pro and Max versions at that time. The company is now expected to round out the line-up with MacBook Pro models powered by the M5 Pro and M5 Max chips, as well as a 16-inch display.  

Aside from the expected performance boost from the new chips, no major design changes or additional hardware upgrades are expected.  

Low-cost MacBook. 

Apple could also introduce a new entry-level MacBook that uses an iPhone-grade A-series processor instead of an M-series chip. Reports indicate it may run on the A18 Pro chip, with performance said to approach that of M1-based MacBooks in certain workloads. The device could feature a smaller 12-inch display with lower brightness than the MacBook Air’s. The new low-cost MacBook could get either an aluminum or a plastic chassis. Apple may also expand the color palette beyond the traditional silver and gray finishes.  

M5 MacBook Air 

A new MacBook Air with the M5 chip is expected soon. The design will likely stay the same following recent trends. The 2026 model might include Apple’s N1 networking chip, which could make wireless connections faster and more reliable.  

M5, Mac Studio 

Apple is said to be working on a new Mac Studio. The current model uses M4 Max and M3 Ultra chips, but the next version will likely feature M5 Max and M5 Ultra chips. These updates could offer better performance for professional and creative users.  

Base iPad (11th generation) 

The 11th-generation entry-level iPad is expected to launch next week. Its design will likely stay the same, but it could get a significant speed boost from the A18 chip, which may enable Apple Intelligence features. It might also use Apple’s own Wi-Fi and 5 G modems.  

Other Devices 

Apple Studio Display 2: Apple is said to be developing two new Studio Display models with mini-LED screens in 27-inch and 32-inch sizes. The bigger model may have a 6K resolution. Both displays could use the new A19 Pro chip instead of the old A13 Bionic.  

New Apple TV 4K: The next Apple TV 4K may use either the A18 or A17 Pro chip, which could support Apple Intelligence features. It is also expected to include a built-in FaceTime camera. The device might include Apple’s N1 networking chip.  

HomePod Mini 2: The next HomePod Mini may come with a newer S-series chip, like the S9 or S10, to boost performance and add Apple Intelligence features, like the new 4 K TV. It could also use the N1 networking chip for better stability and faster Wi-Fi 7.

Source: Apple may launch multiple devices next week: iPhone 17e, low-cost MacBook 

HP’s CEO says rising memory costs will soon make the company raise prices and offer products with less powerful configurations. 

During an earnings call on Tuesday, Enrique Lores talked about rising memory costs which have pushed DDR5 RAM prices up by more than 200% in recent weeks. He said HP has a stockpile of memory, so he expects the company to lessen the impact of these costs. In the first half of our fiscal year, which began this month. 

However, starting in May, higher memory costs will start to reduce HP’s PC product margins, so the company will need to respond. Lores said they include: 

• qualifying lower-cost suppliers 

• redesigning the portfolio to reduce memory configurations 

• accelerating our AI-enabled transformation to drive further cost savings 

• raising prices in close partnerships with our channel and direct customers 

At the end of the call, HP’s CEO added: “What we have seen in the past in these situations, from a demand perspective, is usually the more low-end categories, those that are impacted.” 

The company also plans to raise prices only as needed, depending on the country and product category. 

Other smaller PC makers are also warning about price increases. CyberPower PC will raise prices on Dec. 7. MainGear is telling customers to buy PCs and parts now before prices are expected to go up after Black Friday sales. 

The memory shortage is being blamed on high demand from AI data centers, which is reducing the supply of RAM and SSDs. The shortage is especially worrying because it could last for years if demand for AI stays high. 

This month, the CEO of Fashion, a major supplier of SSD controllers, warned that electronics makers might cut storage capacity in their products by up to 50% because of the shortage. Meanwhile, Lenovo has been stockpiling memory, and experts have expected to have enough to last through 2026

Source: HP to Raise Prices, Lower Configurations Due to Surging Memory Costs 

News Highlights 

• Meta is teaming up with AMD to quickly expand AI infrastructure and accelerate the development and deployment of advanced AI models.  

• AMD and Meta have signed a long-term partnership to deploy up to 6 GW of AMD Instinct GPUs over several product generations.  

• The first Gigawatt deployment will use the AMD Helios rack-scale architecture, announced at the 2025 Open Compute Project Global Summit. Shipments are expected to start in the second half of 2026, using a custom AMD Instinct GPU based on the AMD 450 architecture and designed for Metal’s needs.  

• The AMD Meta is strengthening their partnership by aligning their plans for GPUs/CPUs/systems and software.  

And Meta today announced a 6 GW agreement to power its next-generation AI infrastructure across multiple generations of AMD Instinct GPUs.  

This agreement builds on the companies’ current partnership and aligns their hardware and software plans to create AI platforms tailored to Meta’s needs. The first rollout will use a custom AMD Instinct GPU based on the MI450. Its architecture is optimized for Meta’s large-scale workloads. Shipments for the first Gigawatt deployment are set to begin in the second half of 2026 using the custom-built MI450-based GPU and sixth-generation AMD EPYC CPUs called Venice running ROCM software and built on the AMD Helios rack-scale architecture. AMD and Meta develop Helios together through the Open Compute project to support scalable rack-level AI infrastructure.  

We are proud to grow our partnership with Meta as they advance AI on a massive scale, said Dr. Lisa Su, AMD’s chair and CEO. This long-term collaboration across Instinct GPUs, EPYC CPUs, and Rackscale AI systems brings our plans together to deliver high-performance, energy-efficient infrastructure for Meta’s needs. It will help speed up one of the largest AI deployments in the industry and put AMD at the heart of the global AI expansion.  

Excited to start a long-term partnership with AMD to deploy efficient computing for AI and deliver personal super intelligence, said Mark Zuckerberg, founder and CEO of Meta. This is a key step for Meta. While we expand our computing options, AMD will be a valuable partner for many years.  

Along with their work on GPUs, AMD and Meta are also expanding their partnership around AMD EPYC processors. Meta has worked closely with AMD for several generations, using millions of AMD EPYC CPUs and large numbers of AMD Instinct MI300 and MI350 GPUs in their global systems. AI infrastructure becomes larger and more complex. CPUs play a key role in making the system efficient, scalable, and well-coordinated alongside GPUs. With their plans closely aligned, Meta will be a major customer for the sixth-generation AMD E-PYC CPU, such as the Codenine and Verano. A new E-PYC processor designed for specific workloads to deliver top performance at lower cost and energy use. The agreement to further align strategic interests. AMD has issued Meta a performance-based warrant for up to 160 million shares of AMD common stock structured to vest as particular milestones associated with instant GPU shipments are achieved. The first tranche vests with the initial 1 GW of shipments, with additional tranches vesting as Meta’s purchases scale up to 6 GW. Vesting is further tied to AMD achieving certain stock price thresholds, and exercise is tied to Meta achieving key technical and commercial milestones.  

We expect this partnership to bring strong revenue growth over several years and add to our non-GAAP earnings per share. This is another big step toward our long-term financial goals,” said Jean Hu, EVP, CFO, and treasurer of AMD. “The performance-based structure also closely aligns AMD and Meta on execution and creating long-term value.”  

AMD and Meta are working together on hardware, systems, and software to build global AI infrastructure. The partnership will speed up AI innovation and bring AI-powered services and experiences to billions of people.

Source: AMD and Meta Announce Expanded Strategic Partnership to Deploy 6 Gigawatts of AMD GPUs