XTHE

Today, OpenAI is launching a research preview of GPT-5.3 Codex Spark, a smaller version of GPT-5.3 Codex, and its first model built for live coding. Codex Spark is the first result of its partnership with Cerebras, announced in January. It’s designed to feel almost instant on ultra-low latency hardware, delivering over 1,000 tokens per second while remaining highly effective for actual programming tasks.  

We are making Codex Spark on Cerebras available as a research preview to ChatGPT Pro users. This lets developers start experimenting early while we work with Cerebras to increase data center capacity, improve the user experience, and prepare for the launch of our larger models.  

Our latest models are especially good at handling long-running tasks, working on their own for hours, days, or even weeks. Codex Spark is our first model built for instant work with Codex, so you can make targeted edits, adjust logic, or refine interfaces and see results right away. Now Codex supports both big, ongoing projects and quick, in-the-moment work.  

We look forward to learning from developers and using your feedback as we expand access.  

At launch, Codex Spark has a 128K context window and supports only text. During the research preview, it will have its own rate limits, and usage won’t count toward standard limits. If demand is high, you might see limited access or short waits as we keep things reliable for everyone.  

Speed and Intelligence 

Codex Spark is built for interactive work where speed is just as important as intelligence. You can work with the model in real time, interrupt or redirect it as needed, and quickly try out new ideas with fast responses. Since it is tuned for speed, Codex Spark keeps things simple by making only minimal targeted changes and running tests only when you ask.  

Coding 

Codex Spark is a powerful, small model designed for fast results on the HWE Bench Pro and Terminal Bench 2.0, which tests software engineering skills. GPT-5.3 Codex Spark performs well and completes tasks much faster than GPT-5.3 Codex.  

Latency Improvements for All Models 

While training Codex Spark, we realized that speed alone wasn’t enough for instant collaboration. We also needed to reduce latency throughout the entire request and response process.  
 
We made improvements that will help all models, such as:  

• streamlining how responses move between client and server  

• updating our inference stack  

• making sessions start faster so you see the first token sooner  

By adding a persistent WebSocket connection and upgrading the responses to API, we reduced the client/server round-trip overhead by 80% per token and the first-time end-to-end time for the first token by 50%. Codex Spark uses the WebSocket path by default, and soon all models will too.  

Powered by Cerebras 

Codex Spark runs on the Cerebras wafer-scale engine 3, a specialized AI accelerator built for high-speed inference, providing Codex with a low-latency serving option. We worked with Cerebras to add this fast path into our main production system, so Codex works smoothly and is ready to support future models.  

What excites us most about ChatGPT 5.3 Codex Spark is partnering with OpenAI and the developer community to discover what fast inference makes possible: new interaction methods, new use cases, and a fundamentally different model experience. This preview is just the beginning: Sean Lee, CTO and co-founder of Cerebras.  

GPUs are still the backbone of our training and inference systems, providing the most cost-effective solution for broad adoption. Cerebras adds to this by handling tasks that require very low latency, making Codex feel more responsive as you work. You can also combine GPUs and Cerebras for the best performance on single workloads.  

Availability & Details 

Codex Spark is launching today as a research preview for ChatGPT Pro users in the latest versions of the Codex app, CLI, and VS Code extension. Since it runs on special low-latency hardware, it has its own rate limit that may change based on demand. During the preview, we are also making Codex Spark available in the API for a small group of design partners to see how developers want to use it in their products. We’ll expand access in the coming weeks as we continue improving our integration.  

Right now, Codex Spark is text-only, with a 128K context window, and is the first in a new line of ultra-fast models. As we learn from the developer community about where fast models work best for coding, we’ll add more features, such as:  

• larger models  

• longer context windows  

• support for different types of input  

Codex Spark has the same safety training as our main models, including cybersecurity training. We reviewed Codex Spark as part of our usual deployment process, which checks for cyber and other abilities. We found that it does not have a realistic chance of reaching our preparedness framework threshold for higher capability in cybersecurity or biology.  

What’s Next 

Codex Spark is just the beginning. The goal is to create a Codex in two main modes:  

1. One for longer-term reasoning and execution  

2. Another for live collaboration and quick changes  

Over time, these modes will come together. Codex will let you stay closely involved while it handles longer tasks in the background or spreads work across many models at once when you need speed and variety. This way, you won’t have to pick just one mode from the start.  

As models get better, the speed of interaction becomes more important. Faster responses make Codex easier to use and open new possibilities for everyone who wants to turn an idea into working software.

Source: Introducing GPT‑5.3‑Codex‑Spark 

OpenAI is considering adding advertisements to ChatGPT as a practical way to resolve its financial and business challenges. The company faces high operating costs, including a multi-billion-dollar investment plan, and most users do not pay for the services (hundreds of millions use it for free, while only a small percentage are paid subscribers). It could provide a familiar and scalable way to generate revenue. Both industry analysts and company executives note that advertising could help cover losses and generate income beyond subscriptions. Offering an ad-supported option could also make ChatGPT more accessible by helping to cover the costs of free use, especially for people who cannot or do not want to pay subscription fees.  

The numbers highlight the financial challenge. By early 2025, ChatGPT had over 400 million weekly active users, but only about 5% paid for premium features. As of mid-2025, OpenAI’s computing and research costs are rising rapidly, with nearly $2.5 billion spent in the first half of 2025 alone. Supplying a mostly free, high-quality service is becoming harder to sustain. Advertising is appealing because it helps OpenAI diversify its income (the company intends to derive up to 20% of revenue from ads and connect it to a large digital advertising market).  

Advertising in ChatGPT must be implemented carefully, but many stakeholders maintain that it can be done carefully. Even ChatGPT’s leadership has not categorically forbidden ads, though they stress the need to preserve audience trust in markets or user segments. Resistant to subscription fees, an ad-based alternative could broaden reach and delay or eliminate paywalls. Crucially, ChatGPT already integrates commerce and shopping features (e.g., product recommendations, ad-like model suggestions, affiliates, or commissions in development). From multiple perspectives (financial, competitive, and user experience), introducing ads into ChatGPT is defensible as a logical step for OpenAI to ensure long-term viability and maximize utility. This report examines this proposition in depth, combining data, expert analysis, and case studies to evaluate why an ad-supported ChatGPT could make strategic sense.  

Introduction and Background 

OpenAI launched ChatGPT as a free AI assistant in late 2022, and it promptly gained worldwide attention. Within a few months, millions were using it for everything from writing help to coding. By early 2025, Reuters reported that ChatGPT had over 400 million weekly active users, up from 300 million in late 2024. The rapid growth shows how popular the tool is, but it also emphasizes the challenge of keeping it running. OpenAI started as a non-profit research lab in 2015 but now operates under a capped-profit model and has focused on making money, especially through its partnership with Microsoft. By mid-2025, OpenAI was making billions in revenue (about $4.3 billion in the first half of 2025), but its expenses were even higher (about $2.5 billion in that period), resulting in big net losses. Some industry sources think OpenAI could spend tens of billions more in the coming years.  

From the beginning, OpenAI offered both free and paid versions of its products. The basic ChatGPT was free, while ChatGPT Plus, launched in early 2023, gave users extra features like GPT-4 access, faster replies, and higher limits for $20 a month. Later, a Pro Plan costing $200 a month was introduced for heavy or professional users. Businesses and organization’s also pay for OpenAI’s API and custom solutions, with over two million business users on paid models by early 2025. Since most users are on the free tier, OpenAI has sought ways to monetize this large group while continuing to develop new models.  

In parallel to building its AI, OpenAI has pursued partnerships and new features. It signed content licensing deals with publishers (e.g., DotDash, Meredith, Conde Nast) to legally incorporate their material into ChatGPT, boosting the model’s responses while compensating creators. Launched in chat, shopping features provide product recommendations and links (initially ad-free per official statement). OpenAI is even working on an in-chat checkout system to earn purchase commissions. These moves illustrate OpenAI’s ambitions to intertwine e-commerce with AI assistance, creating new revenue avenues.  

Even with these new features, the main problem is still how to pay for the high costs of advanced AI. OpenAI’s five-year plan includes more than $1 trillion in spending, so leaders are looking to diversify revenue and raise more funds in this situation. OpenAI advertising has come back as a possible solution. OpenAI executives have publicly discussed adding ads to ChatGPT in late 2024. CFO Sarah Friar said the company was carefully looking at ad models to help cover high operating costs. In mid-2025, ChatGPT’s head, Nick Turley, also said ads could be acceptable if they are done in a way that keeps the product’s quality and trust.  

This report looks at why adding ads to ChatGPT could make sense. We review OpenAI’s financial needs, compare what other companies have done, and consider technical and user experience issues, as well as wider effects, by using data such as user numbers, revenue, and forecasts. By referencing published statements and studies, we aim to provide a clear, evidence-based view of why advertising could be a sensible next step for ChatGPT.  

OpenAI’s Financial and Business Context 

Surging Costs and Spending Commitments 

Large language models like ChatGPT require significant computing power. Each new version of ChatGPT needs more resources, and ongoing user activity adds to cloud and hardware expenses. Reports show the scale of OpenAI’s spending, according to the Financial Times. The company has committed over $1 trillion to AI infrastructure over five years, including $300 billion for Oracle and more than $100 billion for hardware vendors (projections suggest OpenAI will spend over $100 billion by 2029 to keep developing its AI). In the short term, OpenAI raised its 2025 revenue forecast to about $12.7 billion, but it expects to reach cash flow positivity only by 2029. In the first half of 2025, the company reported about $4.3 billion in revenue and $2.5 billion in costs, showing heavy investment and losses. Sam Altman said the latest GPT-5 stretched resources so much that the company was running out of GPUs and plans to invest trillions in more data centers.  

There is a clear gap between OpenAI’s revenue and its investment needs. Most of its revenue comes from subscriptions and services, and it could reach $20 billion by the end of 2025 and $125 billion by 2029. However, research and infrastructure costs are rising quickly. To close this gap, OpenAI needs to find new ways to make money beyond its current sources.  

User Base and Revenue Conversion 

ChatGPT’s user base makes this problem even larger, with hundreds of millions of users each week. Even small costs per query add up quickly. Only a small percentage of users pay. In July 2025, about 35 million users (about 5%) subscribed to the ChatGPT Plus ($20) or Pro ($200) plans. There are also about two million business customers paying for enterprise or API services. The rest, hundreds of millions, use the free version. For example, Reuters reported 400 million weekly active users in early 2025, and projections suggest there could be 2.6 billion weekly users by 2030, with about 8.5% (220 million) paying. This conversion rate of about 8-9% is normal for large subscription services, but it still means that over 90% of users would use the service for free.  

The numbers show that most users use OpenAI’s resources for free, which does not bring in direct revenue. Sources say that both researchers and executives are well aware of this issue. From a business perspective, it makes sense to explore ways to monetize free users. Advertising is a proven way to monetize large free user bases in the tech industry. CFO Sarah Fr said ad revenue could help offset losses and reassure stakeholders amid OpenAI’s high costs. Without ads or similar income, it may not be possible to keep offering so much for free.  

Competitive and Industry Comparisons 

OpenAI’s situation is similar to the rest of the tech industry. Search engines and social platforms have long relied on advertising as their primary source of revenue. For example, Google gets over 80% of its revenue from ads (more than $200 billion each year) and offers a search for free. ChatGPT is a new type of search or assistant tool. If it gets as many queries as predicted (over a billion searches per week in 2025), not using ads could give competitors an advantage. EMarketer predicts that AI-driven search ad spending in the US could grow from about $1.1 billion in 2025 to $26 billion by 2029. Companies like Google and Microsoft are rapidly integrating AI into their search and ad platforms. OpenAI faces strong competition, so using proven ways to make money may be necessary to stay in business.  

OpenAI is not the only one encountering these problems. For example, Meta recently decided to add ads to WhatsApp to reach its large user base, realizing that subscriptions and business features were not enough. Telegram has also found it difficult to balance user growth and profitability and sees advertising as a way to increase revenue; however, it also brings extra costs for content moderation. These examples show that big platforms usually start using ads when they have many users but not enough to pay customers. With ChatGPT’s size and growth, adding ads fits this common pattern.  

Monetization Approaches and Models 

Current OpenAI Revenue Streams 

OpenAI generates revenue through several methods. Table 1 presents the main sources of revenue and provides examples of each.  

Consumer Subscriptions	Direct user fees	ChatGPT Plus ($20/MO) and Pro ($200/MO) plans Premium Access and Limits
Enterprise/ API services	Business Contracts API usage fees	Over 2 million subscribers by early 2025 (Custom ChatGPT Enterprise and API)
Content licensing & partners	Licensing Publisher Content, Partnerships	Malta here deals with Dot Dash Meredith Condé Nast Financial Times, etc., to use content.
E-commerce commissions	In chat shopping tools with transaction fees	Upcoming ChatGPT checkout with Shopify partnership earnings affiliate commissions
Potential Advertising	Displaying ads or sponsored suggestions to users	Under consideration by OpenAI would target free users with relevant ads.

Each revenue source has its own limits. Subscriptions make up most of the earnings, with Plus and Pro plans, helping to reach $4.3 billion in H1 2025 revenue, but only a small number of users pay for them. Enterprise/API Plus licensing deals generate additional income, but they are often one-time or limited. Shopping and checkout features are new, but Reuters reported that the first product’s recommendations did not include ads or commissions, which kept the focus on users. However, the company is now moving forward, making money from e-commerce. In the short term, these new e-commerce features could bring in some returns, but advertising could quickly and flexibly increase income, especially with so many free users.  

User Plans and Revenue Options 

ChatGPT’s tiered service plan shows how OpenAI is juggling its current revenue options. As of mid-2025, the structure looks as follows (see the table below).  

Share/Offering	Price (USD)	Ads	Notes
ChatGPT-Free (basic)	$0	No (currently).	Access to basic model (some GPT queries, mostly G3.5) Standard features
ChatGPT Plus (GPT-4)	$20/month	No.	Priority Access Faster Response Larger Context Window Popular Among Power Users
ChatGPT Pro (business)	$200/month	No.	High usage limits, advanced feature (soon configurable GPT-5 etc.), geared to professionals
(hypothetical) ad-supported free	$0 (ad-funded)	Yes (Contextual Ads)	A proposed free tier with ads to subsidize use could be targeted at price-sensitive markets/users.

Table 2: ChatGPT Service Tiers and Pricing. ChatGPT Pro pricing confirmed by Reuters. Potential Ads-Supported Tier Based on Discussion with OpenAI Executives.  

Most TA Free users today use the $0 option, which is designed to be ad-free. Adding ads would create a new free-with-ads version or add ads to the current free plan. Although there is no official ad-supported TA yet, OpenAI executives have suggested offering a limited, ad-supported version instead of subscription fees. For instance, the Spanish press reported that CFO Sarah Friar discussed a free, ad-supported version of ChatGPT for markets where users are unwilling to pay, similar to other freemium tech models. Technology journalists also note that some users would rather have a free version with ads than pay high subscription costs.  

If more users switch to a free ads option, OpenAI could cover the costs for people who might otherwise leave because of high prices. In this model, ads serve as a form of payment: users give their attention rather than money. Since only about 5-10% of users subscribe, an ad-supported tier could bring in revenue from the other 90-95% who currently use the service for free. Ads could appear as banners, promoted content, or suggested links, but the main idea is to shift some user access. Advertisers should not rely solely on subscriptions.  

Advertising In Tech And AI Platforms 

Advertising has long been a way for tech companies to make money. Search engines, social networks, and many apps offer free services funded by ads. For example, when Google launched, it offered users free search results with ads alongside them. Today, Google search ads bring in tens of billions of dollars each year. Social media sites like Facebook, Twitter (X), and YouTube are also free to use, with ads supporting their large revenues. Many mobile apps and games use a similar model, supplying a free version with ads and a paid ad-free option.  

We see similar patterns in AI. Google’s BARD uses Google Search, and while BARD itself does not make money directly, Google still earns from ads shown with search results. Microsoft’s Bing Chat, which uses OpenAI technology, adds AI answers to Bing search and keeps its ad spaces. Similar AI assistants are also considering adding affiliate links to ads, according to a recent EMarketer study. AI-driven ad search in the US could grow from just over $1 billion in 2025 to about $26 billion by 2029. These new AI search ads are expected to be more targeted and natural. If ChatGPT does not include ads, it could lose this revenue since AI assistants can be monetized like search engines.  

Advertising in AI can also look different from traditional ads. Instead of banners, AI could offer sponsored recommendations. For example, if someone asks ChatGPT for earphone suggestions, it could highlight a certain brand with a clear notice and earn affiliate income. OpenAI is already testing this idea. Its Shopify checkout feature emphasizes turning user questions into sales. Building on this, ChatGPT could include promotional suggestions naturally as long as it stays transparent.  

User Views And Market Segmentation 

Users have different opinions about ads. Some may say they would rather see ads than pay high subscription fees. As one tech commentary points out, some people prefer a free ChatGPT with ads over paying a lot for access. This is similar to how people use Media: many switch to paid, ad-free video services if they can, but stick with free ad-supported versions when money is tight. Adding ads could help keep loyal free users from leaving because of new paywalls, which might push them to other AI tools. In this way, an ad model helps keep ChatGPT accessible and retain its user base, rather than forcing everyone to pay or leave.  

Some people worry that ads could hurt trust in ChatGPT. Part of ChatGPT’s value is its neutral, user-focused design. Users want answers that are fair and not influenced by user-advertisers. OpenAI leaders recognize this. Nick Turley has said that any ads must be tasteful and not harm the main AI experience. Ads should not reduce accuracy or ad bias.  

Still, many digital services use clearly labeled, unobtrusive ads, and users often accept them when carefully designed. It is possible to show relevant ads (like for product questions) without affecting other answers. For ChatGPT, a good solution might be to limit ads to specific topics, such as Shopping or local businesses, rather than showing them everywhere.  

Adding ads could also raise privacy matters. Traditional ads often use user data and tracking to target people. Right now, ChatGPT does not use much personal data. If ads are added, OpenAI will have to maintain effective targeting while balancing user privacy demands. One option is to use only the current conversation to choose ads (for example, showing cookware ads if someone asks about recipes), which would limit extra data collection. OpenAI’s focus on user security and ethical AI suggests it could prioritize privacy.  

Overall, multiple perspectives point to an ad approach as a logical evolution. It taps a natural monetization channel, leverages ChatGPT’s free base, and aligns with how tech giants monetize similar services. It remains to be seen exactly how OpenAI will implement it, but the need is clear, and the model has industry precedent.  

Case Studies And Market Examples 

WhatsApp’s Move to Advertisements 

A recent example is WhatsApp, which started in 2009 as a paid app and later became free, ad-supported under its founders. In 2025, Meta announced it would add ads to WhatsApp, especially in the updates tab, where status updates and channels are found. These ads do not appear in private chats, so end-to-end encryption on messages remains. With about 1.5 billion daily users, Meta needed to monetize WhatsApp without abandoning its subscription model. Now, WhatsApp’s ads and subscription tools, such as paid business services, are expected to play a big role in Meta’s $160 billion-plus annual revenue.  

WhatsApp’s move shows that a large free user base can become profitable through ads as a major communication tool. The app could not stay completely free since Meta relies on advertising for revenue. WhatsApp ads appear only in status updates, not in conversations.  

OpenAI could take a similar approach by placing ChatGPT ads in certain tabs or separate answer sections, keeping main conversations ad-free. The WhatsApp example shows that even platforms known for being ad-free often turn to ads when they need to make money. Some users protested Meta’s decision, but as Meta’s CFO explained, supplying both a free ad-supported tier and a paid subscription is a common trade-off. For ChatGPT, avoiding ads is only possible if other revenue sources can cover costs, which may not be realistic for OpenAI in the long run.  

Telegrams Profitability Challenge 

Telegram, known for its focus on security and privacy, has also struggled to turn a profit. Reuters’ Breaking News reported in 2024 that Telegram had hundreds of millions of users but made only $342 million in 2023 and still lost money. The report said that moving to an advertising model like Facebook could increase revenue. Still, Telegram would have to change its approach to content moderation, which could raise costs enough to cancel out any gains. Another option is a privacy-focused premium model that Telegram is testing, but this could make it even harder to generate enough revenue.  

The takeaways for ChatGPT are two-fold. First, a transition to ads often carries secondary costs (content oversight, compliance, etc.), which must be weighed, for Telegram targeted advertising implied a dramatic increase in moderation spending, similar to Facebook’s $1 billion-plus per year, which could negate revenue gains. ChatGPT might avoid some of these pitfalls. It doesn’t host user-generated content at scale, where disinformation moderation is a huge effort, and much of its output is synthetic. However, if ChatGPT starts promoting products or links, it may need supervision to prevent abuse (for example, ensure recommendations are safe). Second, Telegram’s strategy shows that preserving a pure ad-free vision is financially challenging. Even with a committed leader, Telegram has flirted with monetization (e.g., paid subscriptions, limited ads) to reach breakeven. OpenAI faces similar pressures: either accept ads, or future profitability could remain elusive. Telegram’s experience suggests that a blended approach (some free, ad-supported, some paid, no ads) is often a compromise platform.  

Search Engines and AI Assistants 

Google and Microsoft illustrate how AI features interplay with ads. Google, the world’s largest search engine, naturally runs ads on search results when it degrades AI (in Bard/Chat and AI-enhanced search), Google largely maintains ads below AI-generated answers or on separate tabs. Microsoft’s Bing integrated OpenAI tech but kept its ad framework. Their strategy has been to experiment with AI answers while still coercing users to click on sponsored links to purchase OpenAI’s ChatGPT; by contrast, it outputs answers without ads or links. It competes by being more conversational.  

Industry analysts suggest that AI-powered search could disrupt existing ad models. Axios report notes that OpenAI’s ad-free AI search challenges Google’s $200 billion advertising market. Google currently earns $100 billion plus from search ads, so if ChatGPT cannibalizes that traffic, Google’s model is at risk. The remedy for OpenAI: become part of the ad ecosystem rather than removing it entirely. If ChatGPT becomes a major research tool, advertisers would reasonably follow (as users wouldn’t want to miss that opportunity). By eventually sharing in search advertising, ChatGPT could capture a share of the massive digital ad spend. In fact, industry data indicates that AI search advertising is set to explode. An eMarketer report sees a 24x jump in US spending by 2029. If ChatGPT remains ad-free, it effectively ignores that opportunity.  

Content and Commerce Partnerships 

OpenAI’s existing deals with content publishers suggest another angle. Agreements with companies like Dot Dash Meredith allow ChatGPT to surface journalistic content with credit, which makes the AI more useful. Linked deals with commerce partners (e.g., Shopify) create sales, and OpenAI can take commissions. These collaborations often yield revenue either directly (through license fees or commissions) or indirectly (by improving the user experience, leading to higher retention). While not traditional advertising, they employ a sponsored, content-like approach. For example, the hire of Shiva Kumar Venkatraman, an ex-Google Search ad executive, suggests that OpenAI is using ad-tech expertise to manage these functions. Tailoring recommendations or purchases in ChatGPT closely parallels inserting an ad when the user’s query has commercial intent.   

These examples show that OpenAI is already close to mixing information with money-making suggestions. Moving to formal ads like working with Google or showing sponsored search results in ChatGPT would not be a big change. Big companies are already testing affiliate-style features in AI tools, as one industry newsletter noted. Even though OpenAI first promised an advertisement-free experience, its focus on shopping and personalization is turning ChatGPT into a commercial platform. The Shopping Assistant feature, launched in 2025 with product links and images, started without ads or commissions, but reports and Reuters data suggest future versions may exclude affiliate revenue. In short, adding ads to ChatGPT would be a natural next step as it becomes more involved in e-commerce.  

Analysis: Why Ads Make Logistical Sense 

Based on the data and examples above, here is an analysis of why adding ads to ChatGPT makes sense:  

1. Covering high operating costs: OpenAI needs substantial funding to support its technical plans, since its revenue is still below its costs. Regular pricing is not enough to close the gap. Ads can bring in extra money, and CFO Sarah Friar has said they could help offset losses and keep the company stable. With only a few ways to make money (new deals and pro plans), ads are one of the last options left in a five-year plan that could cost a trillion dollars—every bit of revenue matters.  

2. Using a huge user base: ChatGPT has one of the biggest audiences in the world. By 2030, it could have billions of users each week. Even a small amount of ad money per user could add up to a lot. For example, YouTube and Facebook generate large ad revenues from tens of millions of users, but ChatGPT’s user base is even larger. Showing ads to just a portion of free users, like on 10% of queries, could bring in as much or more than subscriptions. Studies show that companies with a large reach often find that ads are the best way to monetize it.  

3. Avoiding High Subscription Barriers: Data shows that only a small number of users will pay high monthly fees if there is a free option. Setting ChatGPT’s price at $20 or $200 is already too much for many, notably in emerging markets or for casual users. Reuters recently reported that some advanced ChatGPT features are only available in expensive plans, which has frustrated users. Allowing ads as an alternative helps keep the service accessible. Economists say that making available both free with ads and ad-free premium options lets more people use the service. Those who want no ads can pay, while others can keep using it for free.  

4. Competing with ad-supported alternatives: ChatGPT often faces off against services that use ads, such as Google Search and Bing Chat. If ChatGPT stays ad-freebut limits usage or data, users may not see it as truly free. If it offers an ad-supported version with similar features, it can compete better. Personalized ads are common in the industry, and not using them could make ChatGPT less effective at engaging users. As ChatGPT becomes more of a super-assistant and connects with other web services and apps, adding ads helps OpenAI fit in with how most platforms work.  

5. Industry examples: Many platforms with millions of users, like WhatsApp, Telegram, and Google, have added ads as they have grown. They did not start this way, but financial needs pushed them to do it. OpenAI is now looking at ads, hiring advertising experts, and talking with Google and Meta about ways to make money, which shows its leaders think it is a smart move. Sam Altman has said OpenAI should build widely useful tools rather than exploit users. In reality, even popular tools like Google and Facebook rely on ads. For OpenAI, offering ChatGPT to a wide audience, including free users, could put it within the same group as these big companies.  

6. Protecting value for content creators: There is concern that if ChatGPT remains free and produces all content, it could hurt content creators and publishers, creating legal and ethical problems by adding ads. OpenAI can share revenue with partners. For example, OpenAI’s content deals already include payments to publishers. An ad system in ChatGPT could also allow revenue splits or sponsored answers that help original authors. In this way, ads help support content creation by sending some of ChatGPT’s value back to contributors. This fits with OpenAI’s focus on giving credit for content, and ads could turn that credit into real payments.  

Importantly, introducing ads does not compromise ChatGPT’s core mission. It can be done in ways that ideally preserve user faith and model impartiality. For example, ads might be clearly labeled or optional (e.g., products mentioned in the answer are sponsored), and confined to certain domains (shopping, entertainment, etc.). ChatGPT could limit personalized tracking and target ads only based on conversation context, addressing privacy concerns. Even if the principal justification is financial, the feasible implementation can be fine-tuned (hence CFO’s emphasis on responsible implementation), the bottom line: given OpenAI’s immense scale and need for revenue, the economic logic for ads is compelling, It shares burdens its shares, burdens across the largely free user base, leveraging massive usage to sustain the business.  

Possible Implications and Challenges 

Adding Ads to ChatGPT Comes with Some Risks  

• Trust and bias: If users believe ads promote certain products or information, they may doubt ChatGPT’s neutrality. OpenAI will need to separate editorial content from ads clearly. If earlier versions of ChatGPT warned against giving financial or medical advice for free, showing ads for those products could seem inconsistent. To address this, OpenAI should apply strict rules for ad content, such as banning ads for medical treatments or legal advice, and ensuring ads do not affect factual answers.  

• User experience: Too many ads, as seen on other platforms, may frustrate users. If the free version of ChatGPT becomes overrun with ads, some users may leave for paid options or switch to paid private versions for sensitive tasks, but ChatGPT remains a useful tool. A small number of relevant ads might be acceptable. Surveys will be needed to confirm this, but they are not available yet.  

• Competitive Response: If ChatGPT stays ad-free while competitors use ads, it might attract users who care about privacy, but could miss out on revenue. On the other hand, adding ads too quickly could upset users. You are likely to test ads slowly and make improvements rather than adding many ads all at once.  

• Regulatory/policy: Advertising is carefully monitored by regulators, particularly regarding privacy laws and trust in advertising rules. OpenAI will need to ensure any ads comply with these rules, especially if they use personal data, given OpenAI’s well-known status. It could even help set new standards for ethical advertising as it rolls out ads transparently.  

• Alternative Revenue Paths: Some people suggest that if ads cause problems, OpenAI could use other ways to make money, like:  

• Raising subscription prices  

• Adding micropayments  

• Focusing on enterprise licensing  

• OpenAI has even discussed premium options, such as the Strawberry Model. Still, our review shows these options have limits. Large price increases could drive away regular users. Focusing solely on enterprise or API sales would reach only certain groups, and content licensing has its own limits. In short, ads make sense as an extra source of revenue, but they are not the only answer.  

We expect OpenAI to test ads carefully, possibly starting in certain markets or with specific features, and to monitor user feedback closely. Showing just one or two targeted ads per week to two free users could raise significant revenue without causing much disruption. If this works well, the program could grow over time.  

Conclusions 

To summarize, adding ads to ChatGPT fits with OpenAI’s current financial needs and the market situation. The company has significant expenses, and most users do not pay, so using ads is a proven way to make money at scale. Both business reasons and user options support this step:  

• Many users would rather see a few ads than pay.  

• High subscription fees and ad revenue can help sustain OpenAI.  

• Careful ad design, making ads tasteful, minimal, and clear, can help users trust and maintain the quality of the main AIS service.  

We have looked at several types of evidence:  

• User numbers and revenue forecasts  

• Executive comments about ads  

• Industry spending estimates  

• Case studies like WhatsApp’s ad rollout (and Telegram’s monetization challenges)  

All of this suggests that advertising is a logical, though complicated, addition. ChatGPT ads would give OpenAI more ways to earn money beyond subscriptions, tap into existing ad spending, and help more people use the service.  

As AI assistants become a key part of the Internet, their business models will affect both users and society. OpenAI’s goal is to benefit many people, so it needs to balance making money with providing users with a good experience. If done carefully, ads can provide the funding needed without harming ChatGPT’s value. Based on current data and what we see with other tech platforms, moving to an ad-supported ChatGPT is a sensible way for OpenAI to keep its AI strong, available, and sustainable in the future.

Source: ChatGPT Ads: The Economic Case for OpenAI’s Monetization Strategy 

In February 2026, OpenAI started testing sponsored ads in ChatGPT for free and Go plan users in the United States. These ads show up below AI responses and help keep advanced features free. Paid plans like Plus, Pro, Business, and Enterprise do not have ads.  

Here is how to manage your new sponsored response settings:  

1. Accessing Ad Controls 

You can control your ad experience right in the GPT app or on the website:  

• Go to settings and select Ad controls or Data controls, depending on your version.  

• If you have a paid plan or are outside the US, you might not see these options.  

2. Managing Personalization (Ad Targeting) 

By default, ads are personalized using your conversation topics, chat history, and past ad interactions. You can adjust these settings:  

• Turn off personalized ads to stop ChatGPT from using your ad history and interests. You will still see ads but only based on your current conversation.  

• You can also turn off the use of memory or past chats for ad targeting the same menu.  

3. Cleaning Ad Data 

You can delete any data ChatGPT has collected for ads.  

• Go to settings, then add controls, and choose to delete ad data.  

• This will clear your ad’s history and interests. It may take up to 30 days for all systems to remove this data.  

4. Interacting with ads 

When you see an ad, click the ellipsis menu on it to:  

• Hide the ad if it’s not relevant or you don’t want to see it.  

• Report the ad if it’s inappropriate, misleading, or breaks any policies.  

• You can also share a specific ad with ChatGPT to ask questions about it.  

4. Opting out of ads entirely 

If you’d rather not see ads, you have two main options:  

• Subscribe to ChatGPT Plus or Pro for a commercial-free experience.  

• Go ad-free on the free tier: Select this option in settings > ad controls. You can also choose the ad-free options in settings under add controls, but this will lower your daily message limit and reduce access to some features like image generation. You will not have access to your chat conversations, memories, or personal details.  

• Ads will not appear in conversations about sensitive topics such as health, mental health, or politics.  

• Users under 18 will not see any ads.  

OpenAI is testing ads in ChatGPT, starting with users on free and ChatGPT Go plans. The company shared this update in a blog post on Monday, February 9. Right now, the test is limited to the US and is not a full rollout. OpenAI wants to see how ads affect the user experience.  

Ads are currently shown to a small group of users. Paid plans, including Plus, Pro, Business, and Enterprise, remain ad-free. OpenAI has not said whether or when the test will expand to other regions. This phase is to learn how ads work in ChatGPT.  

OpenAI explained in the article why ads are coming to ChatGPT. Millions use the platform for learning, work, and daily decisions. Keeping the free-and-go tiers fast and reliable requires considerable infrastructure and ongoing investment. Ads help fund that work, supporting broader access to AI through higher-quality, free, and low-cost options. They enable us to improve the intelligence and capabilities we offer continually, the company said.  

OpenAI also addressed whether ads will change ChatGPT’s responses. The company says ads will not affect the answers you get. Ads do not influence the answers ChatGPT gives you. Answers are optimized based on what is most helpful to you. When you see an ad, it is always clearly labeled as sponsored and visually separated from the organic results.  

Many users were concerned about privacy and whether advertisers would see their chats. OpenAI responded by saying, “Advertisers cannot access conversations. Ads are designed to respect your privacy. Advertisers do not have access to your chats, chat history, memories, or personal details. Advertisers only receive aggregate information about how their ads perform, such as the number of views or clicks.” The company said.  

OpenAI said ads will not be shown to all users. During our test, we will not show ads in accounts where the user tells us, or we predict they are under 18. The company explained that ads also won’t appear near sensitive topics like health, mental health, or politics. OpenAI plans to expand these safeguards as testing continues. User safety and privacy will stay a priority as the ad program grows.  

Users will have control over the ads they see. OpenAI says you can dismiss ads, give feedback, and see why an ad is shown. You can delete ad-related data with one tap and manage ad personalization at any time. If you prefer not to see ads, you can upgrade to our Plus or Pro plans or opt out of ads in the free tier in exchange for free daily messages. The company said.

Source: ChatGPT Introduces Ads: Here’s How It Works And What Indian Users Should Know 

On Feb 13, 2026, OpenAI retired several classic models, including GPT-4o, GPT-4.1, GPT-4.1 mini, and GPT-o4 mini, from the ChatGPT web and app interfaces. This change means everyone must now use newer systems, mainly GPT-5.2, as OpenAI focuses on more advanced models that, according to the company, are safer.  

Here’s what the February 13 model update means for your saved chats and how you work:  

Impact on Saved chats 

• You will still have access to your existing conversations and chat history. You can read and refer to your old chats as before.  

• If you keep a conversation that uses GPT-4.0, it will now automatically switch to GPT-5.2.  

• Since the model is changing, you may notice differences in the chats, personality, tone, or creative style. GPT-5.2 is designed to be more precise and less warm than GPT-4.  

Key Changes and What to Know 

• You won’t be able to choose GPT-4o from the model menu anymore.  

• Some GPTs that used the retired models will now automatically use GPT-5.2 as their default.  

• Although these models are no longer in the ChatGPT app, developers can still access them through the API for a limited time.  

• Business, enterprise, and edu customers will still have access to GPT-4o and custom GPTs until April 13, 2026.  

Why Did This Happen 

OpenAI said that only 0.1% of daily users were still using GPT-4o, whilst most had already switched to GPT-5.2. This change helps OpenAI focus on newer, more capable, and customizable models that give users better control over tone and manner, reducing the need to switch between models.  

What Should You Do 

1. You used GPT-4o for certain tasks; try those prompts in GPT-5.2 to see how the mood changes and make any needed adjustments.  

2. If you have important data, use the export data option in your settings to save your chat history.  

3. To make ChatGPT sound more like the older GPT-4o, update your custom instructions to ask for a friendly or less sterile tone.  

OpenAI will retire GPT-4.0, GPT-4.1, GPT-4.1 Mini, and OpenAI O4 Mini in ChatGPT on 13th February 2026. GPT-5 (instant and thinking) will also be retired, as announced earlier. This change is meant to make ChatGPT faster and more reliable by focusing on current models. There are no API changes right now.  

What’s changing and why? 

The following models will leave ChatGPT on 13 February 2026:  

• GPT-4o  

• GPT-4.1  

• GPT-4.1 Mini  

• OpenAI o4 Mini  

The retirement of GPT-5 was announced earlier.  

Retiring these models helps improve reliability, performance, and support. It also lets us invest more in the latest models and features.  

This retirement only affects ChatGPT. There are no changes to the API at this time. If you use legacy model snapshots in the API, please check the deprecations page and Azure notices for updates.  

Practical Guidance (users & admins) 

Individuals (ChatGPT): 

• Let your team know about the change and suggest updated templates or prompts.  

• Check your team’s playbook for any references to specific legacy model names.  

• If your team is regulated, update any screenshots or process documents that show the model switcher.  

Developers 

• If you build API apps, check the deprecations page regularly. If you use Azure OpenAI, be aware that retirement timelines and auto-upgrade plans may differ.  

Example Transitions 

• If you move from GPT-4.0 to GPT-5.2, retest your saved prompts and add tone instructions, such as asking for concise Warm or British English responses.  

• If you switch from GPT-4.1 Mini to the current lightweight model, check the latency and quality. Adjust the temperature and make max tokens if necessary.  

• If you move from O4 Mini to the current O-class replacement, make sure tool use behaviors and output formats are as expected.  

Risks & Good Practice 

• Expectation drift: Legacy models may have had a distinct feel, capture style via instructions/examples  

• Run quick acceptance tests again on your key prompts to check accuracy, format, and latency.  

• Keep a record of your model settings, pin versions if possible, and save notes for rolling back prompts if needed.  

FAQs 

What models are being retired?  

GPT-4o, GPT-4.1, GPT-4.1 Mini, and OpenAI o4 mini leave ChatGPT on 13 February 2026. GPT-5 (instant/thinking) retirement was previously announced.  

Why is OpenAI retiring from these models?  

We are making this change to improve your experience, improve performance and reliability, and focus on newer models.  

Will there be changes to the API?  

There are no API changes right now. Please keep an eye on the official deprecation notices for updates about snapshots and Azure timelines.  

How will this affect users?  

Most users do not need to do anything. The default model will switch to the latest versions, which offer better performance and stability.  

How can I preserve a legacy modes tone?  

To match the previous model’s behavior, use custom instructions and include a short style block with examples of tone, phrasing, and dos and don’ts.

Source:   OpenAI Retires Older GPT Models (Feb 2026) 

We are adding AI to more of our products and services to boost creativity and productivity. At the same time, we want to help people understand how content is created and changed. Everyone needs to have this information, so we are investing in tools and new solutions like SynthID to make it available.  

We also know that working with others in the industry is essential to increase overall transparency online, as content travels between platforms. That’s why we joined the coalition of content provenance and validity (C2PA) as a steering committee member earlier this year.  

Today, we want to share how we are helping to develop the least C2PA provenance technology and bring it to our products.  

Improving Content Technology to Make Credentials More Secure 

Provenance technology can show if a photo was taken with a camera, edited with software, or made by generative AI. This information helps users make better choices about the content they see, such as photos, videos, and audio, and builds trust and media literacy.  

As a steering committee member of the C2PA, I have worked with other members to improve technology that attaches provenance information to content. In the first half of this year, Google helped develop the latest version (2.1). This version is more secure against tampering because it imposes stricter requirements for verifying content history. By strengthening these protections, we help ensure the attached data is accurate and not misleading.  

Adding The C2PA Standard To Our Products 

In the next few months, we will add this new version of content credentials to some of our main products:  

• Search: If an image has C2PA metadata, people can use our “About this image” feature to check if it was created or edited with AI tools. This feature provides people with greater context about images they find online and is available in Google Images, Lens, and Circle for search.  

• Ads: We are starting to add C2PA metadata to our ad systems. Over time, we plan to expand this and use C2PA signals to help guide our policy enforcement.  

We are also looking at ways to share C2PA information with viewers on YouTube when content is recorded with a camera. We will provide more updates about this later in the year.  

We will ensure that our implementations validate content. We will ensure our system checks content against the upcoming C2PA trust list, which helps platforms confirm its origin. For example, if the data indicates an image was captured with a specific camera model, the trust list can help verify that information. Implementing content provenance technology today, we’ll continue to bring it to more products over time.  

Working With Others In The Industry 

Setting up and displaying content provenance remains a difficult challenge, and it varies by product or service. There is no single solution for all online content, so working with others in the industry is key to developing lasting and compatible solutions. That’s why we encourage hardware providers to consider using C2PA’s content credentials.  

Our work with the C2PA supports our wider efforts to be transparent and develop AI responsibly. For example, we are expanding the SynthID, a watermarking tool from Google DeepMind to more generative AI tools and different types of media. We have also joined other groups focused on AI safety and research. We introduced a secure AI framework (SAIF) and coalition. We are also making progress on the voluntary commitments made at the White House last year.

Source: How we’re increasing transparency for gen AI content with the C2PA 

Intel Corporation (NASDAQ: INTC) has reached a major milestone in the semiconductor industry. It’s 18A (1.8 NM class) process node has started high-volume manufacturing at the new Fab 52 facility in Arizona. This marks the completion of CEO Pat Gelsinger’s 5 Nodes in 4 Years plan, making Intel the first company to produce 2NM-class technology at scale. By late December 2025, the 18A node is being used in the first production run of the Panther Lake processor family, a key product aimed at strengthening Intel’s position in the growing AI/PC market.  

Starting volume production at the US$30 billion Fab 52 is a major step for the US semiconductor industry. Despite Wall Street skepticism and technical obstacles along the way, recent internal reports show that manufacturing yields have improved and become more stable. Earlier this year, Intel’s 18A process lagged behind Taiwan Semiconductor Manufacturing Co.’s, but it has improved by about 7% each month since then. Yields reached 60-65% in November, and Intel expects to reach its 70% goal by the end of 2025. This progress supports both Intel’s own products and its foundry customers.  

The Architecture of Leadership: RibbonFET and PowerVia 

The 18A node is not just a smaller transistor; it brings the biggest changes in semiconductor design in over ten years. Two key technologies are at its core: RibbonFET and PowerVia. RibbonFET is Intel’s version of gate-all-around (GAA) transistors, replacing the older FinFET design. With the gate surrounding all four sides of the transistor channel, RibbonFET offers better control, reduces power loss, and allows for higher current. This leads to a 25% improvement in performance per watt compared to earlier generations, which is important for AI workloads that need high efficiency.  

PowerVia works alongside RibbonFET, making it Intel’s first use of backside power delivery in the industry. Normally, power and signal lines are placed together on the front of a chip, which can cause voltage drops and make routing more difficult. PowerVia moves the power network to the back of the silicon wafer, separating it from the signal lines. This change reduces voltage drop by 10% and opens up more space for signal routing, resulting in a 0.72x area reduction compared to the Intel 3 node. With these two innovations, Intel has moved ahead of competitors who are not expected to use backside power until they are 2nm or smaller nodes in 2026.  

Experts say that the steady improvements in 18A yields demonstrate the value of Intel’s use of ASML (NASDAQ: ASML) twin-scan NXE:3800 low-NA EUV lithography systems. At first, some questioned Intel’s decision to use refined low-NA EUV instead of high-NA EUV for 18A, but the current production ramp shows this decision has worked. By improving its process with existing equipment, Intel has started high-volume manufacturing before TSMC’s N2 (2NM) node, which is not expected to reach similar volumes until mid or late 2026.  

Shifting the Competitive Landscape: Intel Foundry vs. The World 

The successful launch of 18A at Fab 52 is already affecting the global foundry market. TSMC has long dominated advanced manufacturing, working with companies like Apple and NVIDIA. Now Intel’s progress is attracting major foundry customers. Microsoft and Amazon have decided to use the 18A node for their customers’ custom AI chips, striving to diversify their supply chains and rely less on Taiwanese manufacturing.  

Now, Samsung and TSMC face more competition. Samsung was the first to use GAA at 3nm, but had yield problems that limited its success. Intel’s 60 to 65% yields on a more advanced 1.8nm class node make it attractive to customers concerned about Samsung’s reliability. For TSMC, Intel is now a direct competitor in the foundry business, not just a CPU designer. If Intel continues to improve yields by 7% each month, it could provide a cost-effective alternative to TSMC’s N2 node when it reaches volume production.  

The Panther Lake production ramp is also an important internal test for Intel, as it accounts for 70% of the Panther Lake Dye Area in-house. On 18A, Intel is cutting back on large pavements to outside foundries. This approach, called the IDM 2.0 strategy, aims to boost Intel’s profit margins, which have been under strain throughout heavy investment. If Panther Lake meets its performance goals in the market this month, it will show the industry that Intel’s manufacturing is back on track.  

Geopolitics and the AI infrastructure era 

The importance of 18A production at Fab 52 is especially clear in the context of global technology politics as the U.S. government works to bring key technology manufacturing back to the country through the CHIPS and Science Act. Intel’s Arizona facility is a leading example of advanced domestic production. The 18A node has already been selected for the Department of Defense’s Secure Enclave program, ensuring the next generation of U.S. defense and intelligence hardware is made in America. This gives Intel an advantage that is about national security as much as it is about technology.  

The 18A node comes at a key time for AI. Today’s AI/PC trend needs processors that can run detailed neural networks on the device without draining the battery. The efficiency improvements from RibbonFET and PowerVia are designed to address these needs. By being first to produce two NM-class chips, Intel is giving the industry the hardware needed for the next generation of AI applications. This could help Intel regain ground in the laptop and workstation markets after years of competition from ARM-based chips.  

This milestone also marks the end of a period of doubt for Intel. Many saw the five-nodes-in-four-years promise as just marketing, not a real engineering goal. By delivering 18A in volume by the end of 2025, Intel has regained trust from investors and partners. This success is similar to the tick-tock era when Intel led the industry, showing the company has moved past the 10nm and 7nm delays that lasted nearly ten years.  

The road to 14A and high NA EUV 

Looking forward, the success of 18A sets the stage for Intel’s next big step: the 14A (1.4nm) node. While 18A used improved low-NA EUV, 14A will be the first to use ASML’s high-NA EUV lithography on a large scale. Intel has already received the first high-NA machines at its Oregon research site, and the 18A ramp at Fab 52 will help perfect the next generation of chip manufacturing.  

In the short term, the industry is watching the launch of Clearwater Forest, the 18A-based Xeon processor planned for early 2026. While Panther Lake is aimed at consumers, Clearwater Forest will test how well 18A works in the important data center market. If Intel can deliver better performance per watt in servers, it could stop losing market share to AMD.  

There are still challenges, especially in scaling the 18A process to meet the different needs of many foundry customers, each with their own design rules. Still, Intel’s current progress suggests it could regain the manufacturing crown by 2026. Analysts say that if yields reach 70% by early 2026, Intel’s foundry could become profitable sooner than expected, changing the economics of the semiconductor industry.  

Another Chapter for Silicon 

Starting Volume production at Fab52 is more than a company milestone. It shows that the semiconductor industry is still full of fast, disruptive innovation. Intel’s 18A node combines advanced transistor design with a new power delivery system, setting a new standard for silicon chips. As Panther Lake chips reach consumers this month, people will experience the 1.8nm era for the first time.  

The main points are clear:  

• Intel has managed its toughest technical transition.  

• The US is back in advanced manufacturing.  

• The competition for AI hardware is heating up.  

The next few months will be important as Intel shifts from stabilizing yields to optimizing them for customers worldwide.  

For the tech industry, the message is clear: Intel is back, and it is not merely talk; it is becoming a reality in Arizona. As 2025 draws to a close, the question is no longer whether Intel can build the future, but how quickly it can grow.

Source: Intel’s 18A Node Hits Volume Production at Fab 52 as Yields Stabilize for Panther Lake Ramp 

NVIDIA has officially released the public beta of Project G-Assist, an AI assistant designed for GeForce RTX PC owners, which is available in the NVIDIA app. With this tool, you can easily tune, control, and optimize your gaming PC settings using voice or text commands.  

Key Features and Capabilities 

Project G-Assist runs directly on your device, so you don’t need a subscription, and you get lower latency.  

• Performance Monitoring and Tuning: You can ask the AI to show or graph your frame rates, latency, GPU temperature, and usage. It can also optimize your game settings and apply GPU overlocks.  

• Voice and text commands: the Assistant understands natural language, so you can quickly make changes by saying things like “improving my power efficiency”.  

• Context-aware help: G-Assist can use screen snapshots to give you in-game tips, such as boss strategies and where to find items.  

• Peripheral control: The AI can control devices from partners such as Corsair, Logitech, MSI, and Nano Leaf, allowing you to adjust lighting or fan speeds.  

• Community Plugins: You can add custom plugins to the assistant through a dedicated hub to get new features.  

At Gamescom, NVIDIA is rolling out the first major update to Project G-Assist. This experimental on-device AI assistant lets users tune their NVIDIA RTX systems using voice and text commands.  

The update launches a new AI model that uses 40% less VRAM, improves tool calling intelligence, and expands G-Assist to support all RTX GPUs with 6 GB or more VRAM, including laptops. There is also a new G-Assist plugin hub that lets users easily find and download plugins for additional features.  

NVIDIA also announced a new path-tracing particle system for the NVIDIA RTX Remix modding platform, set to launch in September. This system will add fully simulated physics, dynamic shadows, and realistic reflections to visual effects.  

NVIDIA also announced the winners of the NVIDIA and ModDB RTX Remix Mod Contest. You can see the winning and finalist RTX Mods in the RTX Remix GE Force article.  

G-Assist: Now Smarter and Broadly Accessible on More RTX PCs! 

Modern PCs are powerful, but getting the most out of them often means dealing with a complicated mix of settings across software, GPU tools, and control panels.  

Project G-Assist is a free on-device AI assistant designed to make things easier. It functions as a central command center, giving users quick access to functions that were once hidden in menus via voice or text commands. Users can ask assistants to:  

• Run Diagnostics to Optimize Game Performance.  

• Display or chart frame rates, latency, and GPU temperatures.  

• Adjust GPU or even peripheral settings, such as keyboard lighting.  

The G-Assist update delivers a much more efficient AI model that is faster and uses 40% less memory while still giving accurate responses. This improvement allows G-Assist to run on all RTX GPUs with 6GB or more VRAM, including laptops.  

It’s easy to get started:  

1. Install the latest game-ready driver (580.97 and above) from the NVIDIA App.  

2. Open the NVIDIA app, go to Settings > About and opt in to beta and experimental features/early access, then relaunch the app; it should be on version 11.0.5.  

3. In the NVIDIA app, go to Home and then scroll down to discover and download the G-Assist 0.1.17 update.  

4. Press Alt + G to activate.  

Another G-Assist update coming in September will add support to laptop-specific commands, including features like NVIDIA Battery Boost and Battery OPS.  

Introducing the G-Assist Plugin Hub with Mod.io 

NVIDIA is working with Mod IO to launch the G-Assist Plugin Hub. This hub lets users easily find G-Assist plugins and discover and download community-made ones.  

With the latest update, users can now ask G-Assist which new plugins are available in the Hub and install them using natural language, all thanks to a modern Mod.io plugin.  

Project G-Assist is a big step towards bringing AI into PC gaming. It makes system optimization easier and provides real-time in-game help.

Source: New Lightweight AI Model for Project G-Assist Brings Support for 6GB NVIDIA GeForce RTX and RTX PRO GPUs 

Recent reports from 2025 and early 2026 indicate that Microsoft has launched an aggressive talent-acquisition and retention strategy for its AI divisions, Microsoft AI and Core AI. This approach acts as a pilot program for the company.  

The company offers a fast-track hiring process, high pay, and an independent work environment to compete with companies like Meta and OpenAI.  

Important details of this project include the following:  

1. High Compensation Packages 

To keep AI talent from leaving for competitors, Microsoft is offering compensation packages much larger than those for traditional software roles at the company.  

• Base salaries: AI software engineers can earn up to $377,611.  

• Total compensation: Some employees receive multi-million-dollar packages that include large stock awards.  

• Sign-on & stock: Top roles may include stock awards of up to $1.9 million at hire and annual stock grants of up to $1.47 million.  

• Salary differential: On average, AI employees earn $120,000 more per year than peers in other divisions, such as Azure.  

2. Fast Track Hiring and Startup Culture 

• 24-Hour Offers: Microsoft uses a fast-track hiring process to secure key talent within 24 hours.  

• Customized compensation: a compensation modeler and dedicated consultants create tailored, high value offers for candidates.  

• Operational freedom: led by Mustafa Suleyman. The new AI division operates as a self-contained center of innovation. It is often located away from the main Redmond campus, such as in Mountain View, California, to create a startup-like environment.  

3. Targeted Emphasis On Critical Talent 

• Internal incentives: Managers are encouraged to give financial rewards to employees working on internal AI tools to help keep them at the company.  

• Targeted Roles: These high-compensation packages are primarily for specialists in machine learning, AI infrastructure, and Co-Pilot tools.  

4. Broader AI Initiatives 

• Training Ecosystem: In addition to hiring, Microsoft is investing in training. The company plans to train 500,000 people in India by 2026 through the India AI mission, which includes new AI productivity labs and a center of excellence.  

This aggressive strategy is a direct reaction to the ongoing competition for AI talent. It is meant to help Microsoft keep its lead in AI technology.  

What You Need to Know 

• A leaked Microsoft spreadsheet shows that employees in the AI Division are paid much more than those in Azure and Cloud.  

• An average AI software engineer earns up to $377,611, which is $120,000 more than the average salary in the Azure division.  

• Microsoft is now focusing on more AI projects, such as Copilot, while other important departments appear to be lower priorities.  

In recent years, there have been hints and rough estimates about Microsoft employees’ pay. A leaked payment guideline shows that the highest-ranking employees get a base pay between $231,700 and $361,500, a hiring bonus of up to $1.2 million, and $1 million in annual stock awards. The lowest-ranking employees receive $42,500 with no extra compensation.  

A few months before the payment guideline leak, a poll of Microsoft employees showed that more than half would leave for a better offer at a rival company. Many said that not receiving a raise hurts their performance and morale, leading them to seek other opportunities.  

According to a spreadsheet obtained by Business Insider, employees in Microsoft’s AI department are earning higher salaries. That is, it is not surprising, since Microsoft has invested billions in OpenAI’s technology and is integrating it across its products.

Source: Leaked document: Microsoft pays, on average, $120K more per year for an AI software engineer compared to someone working in Azure 

Samsung Electronics America has launched the Galaxy Book6 Ultra, Galaxy Book6 Pro, and Galaxy Book6X at CES 2026. These are the most advanced Galaxy Book laptops so far, offering strong performance, AI-powered productivity, and a slim, well-designed build.  

At Samsung, we believe true innovation starts with getting the fundamentals right, Said Won-Joon Choi, President and Chief Operating Officer (COO) and Head of the RD Office, Mobile experience (MX) business at Samsung Electronics. Performance defines the PC experience. With the Galaxy Book6 series, we combine unsurpassed speed and power with dependable AI to deliver the exceptional productivity and creativity capabilities users expect from Samsung.  

Engineered For Unmatched Performance 

The Galaxy Book6 Series combines advanced hardware and impressive visuals and audio in a slim, portable design. Powered by Intel Core Ultra Series 3 processors, the first client SoCs built on Intel 18A, these laptops offer fast CPU, GPU, and NPU1 performance for quick processing, fluid multitasking, and responsive AI features.  

• The new Intel Core Ultra Series 3 processors are built on the Intel 18A 1.8 nm class node, with up to 16 power-efficient cores, providing over 60% faster CPU performance than the previous generation. The Integrated NPU delivers up to 50 TOPs, enabling AI tasks like image cleanup, translation, and smart search to run quickly without the cloud.  

• The Galaxy Book6 Ultra features the latest NVIDIA GeForce RTX 5060 LTE laptop GPU, enabling fast AI image generation, smooth video playback and editing, and captivating games for creative and entertainment needs.  

To get the best performance, strong hardware needs optimized cooling. Samsung’s new cooling system keeps the laptops running efficiently and quietly. The improved vapor chamber and air flow help remove heat while keeping the Galaxy Book6 Ultra and Pro models quiet.  

• For the first time, the Pro series includes a vapor chamber in the Galaxy Book6 Pro, and the Galaxy Book6 Ultra has a larger one. This helps spread heat more evenly, keeping the devices cool and responsive during heavy use.  

• The Galaxy Book6 Ultra and Pro feature larger fins attached to the vapor chamber, increasing surface area and helping the fan remove heat more effectively. The disc design improves cooling efficiency by 35% compared to the previous generation.  

• The Galaxy Book6 Ultra uses a new dual-path outlet, fan, and heat sink to absorb and release heat from the processor and other parts, helping prevent overheating and slowdowns.  

• The redesigned fan has an improved angle for faster heat release, and a larger inlet grill lets more heat escape. Samsung’s unique blade spacing also helps keep fan noise low.  

Samsung designed the Galaxy Book6 to offer strong performance and long battery life for all-day use. Better power management means users can stay productive and connected longer wherever they are.  

• The Ultra and Pro models feature Samsung’s longest-lasting Galaxy Book battery to date. The Galaxy Book6 Ultra can play video for up to 30 hours, about 5 hours longer than the previous model.  

• The Galaxy Book6 Ultra charges quickly, restoring up to 63% of the battery in just 30 minutes. This means users can recharge during a short break or keep working for hours without interruption.  

Since users spend all day looking at the screen, the Galaxy Book6 Ultra and Pro now feature much better displays. They show vivid, high-contrast images with clear detail in any lighting, cutting glare while keeping rich detail.  

• The advanced Dynamic AMOLED 2x Touch Screen displays reach up to 1,000 nits of HDR brightness for clear contrast and vivid colors indoors or outdoors, and 500 nits for everyday tasks. Also, look sharp. Users can easily use the touch screen to edit content directly.  

• Vision Booster adjusts the display for outdoor use by analyzing light and screen content, making it easier to see and keeping colors accurate, even in bright sunlight. Anti-reflective technology also reduces glare for a clearer, more comfortable view wherever you are.  

• True Bright 1300 certification means the display looks bright and clear, while True Black 0.0005 nits allow for deep blacks. This makes the screen great for creative work, entertainment, and productivity.  

• The screen’s refresh rate adapts from 30Hz to 120Hz, providing smooth motion for animation, gaming, and video streaming.  

• Corning Gorilla Glass with DXC makes the screen tougher against drops, scratches, and cuts. Front surface reflection is reduced by up to 75% compared to regular glass. This holds the display clear in bright light and maintains Gorilla Glass’s well-known durability.  

A great PC needs both strong visuals and quality sound. The Galaxy Book6 Ultra and Pro feature speakers designed for balanced audio, making voices clear in meetings and classes and delivering rich bass for movies and games.  

• The Galaxy Book6 Ultra has six speakers with Dolby Atmos, including four force-cancelling woofers and two tweeters. This set-up delivers clear, powerful sounds for movies, games, and music.  

• The Galaxy Book6 Ultra uses a back-to-back woofer design to cancel vibrations, reduce distortion and keep sound clear, and the laptop steady even at high volumes.  

• Both the Galaxy Book6 Ultra and the Galaxy Book6 Pro 16 have up-to-firing tweeters for clearer calls and dialogue, and side-firing woofers for deeper, more dynamic bass in music and entertainment.  

Slim And Balanced All the Way 

Samsung’s design combines strong performance with careful craftsmanship, creating slim, premium laptops for everyday use. Every detail is designed for balance and ease of use.  

• By updating components such as the vapor chamber, fan, display, bezels, and hinge, the Galaxy Book6 Ultra and Galaxy Book6 Pro 16 are slimmer and easier to carry. The Galaxy Book6 Ultra is now 15.4mm thick, 1.1mm thinner than the Galaxy Book 4 Ultra. The Galaxy Book6 Pro 16 is 11.9mm thick, 0.6mm thinner than the Galaxy Book 5 Pro 16.  

• The Galaxy Book6 series features an asymmetrical design with clean lines, curved corners, and a centered logo, giving it a unified and premium look.  

• The two-tone keyboard and haptic touch trackpad are centered for visual balance and comfort. This setup helps users type naturally, navigate smoothly, and reduce typing errors.  

• The inside of the Galaxy Book6 series is organized and balanced. Samsung’s new PCB layout spreads components out to reduce space and weight, enabling a slimmer design, stable performance, and greater durability.  

All Day Performance and Networking with Galaxy AI 

To support all-day productivity with AI, you need speed, stability, and lasting performance. The Galaxy Book6 uses powerful computing and Galaxy AI to provide fast, easy-to-use tools that work both on the device and in the cloud, helping users stay productive all day.  

• With Ai Select, users can tap any phrase on the touchscreen to get instant information and analysis while browsing, shopping, or viewing content. Ai Cutout lets users quickly remove backgrounds from images, making it easy to create visuals for presentations, online stores, or marketing.  

• Intelligent Search lets users describe what they are looking for in plain language and quickly find it, whether it’s a vacation photo, a setting, or a file from two weeks ago.  

• Note Assist helps with writing and note-taking by summarizing text and translating notes into different languages. This is useful for organizing ideas after meetings or working with teams around the world.  

• Storage Share lets users easily access photos and files across Galaxy devices. For example, users can open photos from their Galaxy phone and edit them on the Galaxy Book6’s bigger screen without cables or external drives.  

• Link to Windows/Phone lets users view phone apps, messages, and notifications on the Galaxy Book6’s larger screen for easier multitasking. For example, users can take a photo on their phone, edit it with a Generative Edit, and finish editing on their PC. They can also use Live Translate during calls or chats and review translation on the laptop’s larger display.  

• Nearby Devices lets users quickly see and connect to other devices with drag-and-drop and check their features. For example, during a meeting, users can link their Galaxy Book6 to a Galaxy Tab or smartphone to share files, mirror screens, or control devices, making collaboration and productivity easier.  

• Delta Control lets users move a single cursor across the Galaxy Book6 phone and tablet, making it easy to copy, paste, and drag and drop between devices. For example, users can drag a photo from their phone or text from their Galaxy Tab straight into a presentation on their PC, helping them work faster.  

• The second screen allows users to expand their workspace right away. For example, they can view research on a Galaxy Tab while editing a document on the Galaxy Book6.  

Advanced Security Backed by Samsung Quality and Care 

Samsung stands behind the Galaxy Book6 series, emphasizing durability and consistent performance. Samsung Nux offers strong hardware-based security, and Windows 11 Secure Core PC features that provide additional protection. Each Galaxy Book6 is tested for durability and quality. Samsung Care Plus 23 provides coverage for accidental damage, repairs, and replacements, so users can work and create with confidence.  

Availability 

You can reserve the Galaxy Book6 series now on Samsung.com in a stylish grey color. Choose the model that fits your needs:  

• Galaxy Book6 Ultra starts at $2,499.99  

• Galaxy Book6 Pro at $1,599.99  

• Galaxy Book6 at $1,099.99  

The Galaxy Book6 Enterprise Edition, designed for managed IT environments, will be available starting in late spring 2026.  

For more details about the Galaxy Book6 series, visit Samsung Newsroom or Samsung.com.

Source: Engineered for Perfection: Galaxy Book6 Delivers Advanced Performance and Al-Powered Productivity in a Sleek New Design