OpenAI on Friday launched a limited US-only preview of its latest and most powerful AI model series for a select group of partners at the request of the US government, the company said.&nbsp;The rollout comes two weeks after the White House ordered OpenAI rival Anthropic to block all foreign nationals from accessing its Fable 5 and Mythos 5 models, citing national security concerns. Anthropic later suspended access to both models, saying it could not reliably comply with the restriction on foreign nationals.&nbsp;Earlier this month, President Donald Trump signed an executive order establishing a voluntary federal review of national security risks in advanced AI models before their public release.<h2>OpenAI limits rollout to trusted US partners</h2>The company said it briefed the US government on the capabilities of its new AI models before the launch. At the government's request, the initial preview has been restricted to a select group of trusted partners whose identities have been shared with authorities.&nbsp;OpenAI said the participating organisations are based in the United States, although overseas employees at those companies and entities will also be able to access the new models.&nbsp;The GPT 5.6 series includes three models. Sol is OpenAI's new flagship model, Terra is designed for everyday tasks, and Luna is a faster, lower-cost option.&nbsp;Once widely available, OpenAI said Terra will cost half as much as its predecessor GPT 5.5 as the company competes with Anthropic and Google for customers.<h2>US eases restrictions on Anthropic</h2>Meanwhile, the US government on Friday lifted its block on Anthropic's Claude Mythos 5 AI model, allowing the company to release it to more than 100 US institutions, including major companies and government agencies.&nbsp;The decision came in a letter sent to Anthropic on Friday afternoon and marked a major de-escalation in the dispute between the Trump administration and the AI company. Two weeks earlier, the administration imposed export controls on Mythos and Fable 5 after warnings from Amazon and other companies that the models could be "jailbroken" for malicious purposes.

The move follows a White House order issued two weeks ago directing another AI company, Anthropic, to block foreign nationals from accessing its Fable 5 and Mythos 5 models over national security concerns.

OpenAI on Friday launched a limited US-only preview of its latest and most powerful AI model series for a select group of partners at the request of the US government, the company said. The rollout comes two weeks after the White House ordered OpenAI rival Anthropic to block all foreign nationals from accessing its Fable 5 and Mythos 5 models, citing national security concerns. Anthropic later suspended access to both models, saying it could not reliably comply with the restriction on foreign nationals. Earlier this month, President Donald Trump signed an executive order establishing a voluntary federal review of national security risks in advanced AI models before their public release. OpenAI limits rollout to trusted US partners The company said it briefed the US government on the capabilities of its new AI models before the launch. At the government's request, the initial preview has been restricted to a select group of trusted partners whose identities have been shared with authorities. OpenAI said the participating organisations are based in the United States, although overseas employees at those companies and entities will also be able to access the new models. The GPT 5.6 series includes three models. Sol is OpenAI's new flagship model, Terra is designed for everyday tasks, and Luna is a faster, lower-cost option. Once widely available, OpenAI said Terra will cost half as much as its predecessor GPT 5.5 as the company competes with Anthropic and Google for customers. US eases restrictions on Anthropic Meanwhile, the US government on Friday lifted its block on Anthropic's Claude Mythos 5 AI model, allowing the company to release it to more than 100 US institutions, including major companies and government agencies. The decision came in a letter sent to Anthropic on Friday afternoon and marked a major de-escalation in the dispute between the Trump administration and the AI company. Two weeks earlier, the administration imposed export controls on Mythos and Fable 5 after warnings from Amazon and other companies that the models could be "jailbroken" for malicious purposes.

The move follows a White House order issued two weeks ago directing another AI company, Anthropic, to block foreign nationals from accessing its Fable 5 and Mythos 5 models over national security concerns.

After Anthropic, US now limits OpenAI's latest powerful AI model to domestic users

OpenAI has said it will provide the European Union with access to its new cybersecurity-focused AI model, GPT-5.5-Cyber, while Anthropic has not yet extended similar access for its Mythos model. The announcement marks a key development in how advanced AI tools are being shared with governments and institutions. According to OpenAI, the model will be available in a limited preview to selected partners, including businesses, cyber agencies and EU bodies.<h2>What access the EU will get</h2>European organisations, including the EU AI Office and cybersecurity authorities, will be able to test and review GPT-5.5-Cyber. The aim is to allow regulators and experts to closely monitor how the model works and how it could be used in real-world security scenarios. Officials said this access would help address concerns around safety and ensure that the deployment of such tools is properly understood. According to EU spokesperson Thomas Regnier, discussions between OpenAI and the European Commission are ongoing, with further meetings planned.<h2>Why Anthropic’s Mythos is not yet available</h2>While OpenAI has moved ahead with access, Anthropic has not yet provided the EU with preview access to its Mythos model. The EU has confirmed that discussions with Anthropic are taking place, but they are still at an earlier stage. Officials said that while several meetings have been held, no agreement similar to OpenAI’s current arrangement has been reached so far.&nbsp;This difference highlights varying approaches among AI companies when it comes to sharing advanced cybersecurity tools with regulators.&nbsp;<h2>Growing focus on AI in cybersecurity</h2>The move comes at a time when AI tools are increasingly being used in cybersecurity. Models like GPT-5.5-Cyber are designed to help identify vulnerabilities, analyse threats and support defence systems. Recent developments, including the release of Mythos, have raised concerns about how powerful such tools could be, especially if used without proper oversight.By granting access to the EU, OpenAI is positioning its model as a tool that can be monitored and used responsibly by trusted institutions.&nbsp;<h2>What OpenAI has said</h2>OpenAI officials have said that cybersecurity should not be managed by companies alone. The company’s Head of OpenAI for Countries, George Osborne, stated that collaboration with governments and institutions is important to ensure safety. He added that making such tools available to trusted partners can help improve security and support public safety efforts.&nbsp;<h2>What this means for Europe</h2>Access to GPT-5.5-Cyber could allow European institutions to better understand how AI can be used in defence systems. It may also help shape future policies around AI and cybersecurity. At the same time, the delay in accessing Anthropic’s Mythos shows that not all AI tools are being shared at the same pace.<h2>What happens next</h2>Further discussions between the EU and AI companies are expected in the coming weeks. The outcome of these talks could influence how advanced AI systems are regulated and deployed across Europe. As cybersecurity risks continue to grow, access to such tools may become an important part of how governments prepare for future threats

OpenAI has said it will provide the European Union with access to its new cybersecurity-focused AI model, GPT-5.5-Cyber, while Anthropic has not yet extended similar access for its Mythos model.


OpenAI has said it will provide the European Union with access to its new cybersecurity-focused AI model, GPT-5.5-Cyber, while Anthropic has not yet extended similar access for its Mythos model. The announcement marks a key development in how advanced AI tools are being shared with governments and institutions. According to OpenAI, the model will be available in a limited preview to selected partners, including businesses, cyber agencies and EU bodies. What access the EU will get European organisations, including the EU AI Office and cybersecurity authorities, will be able to test and review GPT-5.5-Cyber. The aim is to allow regulators and experts to closely monitor how the model works and how it could be used in real-world security scenarios. Officials said this access would help address concerns around safety and ensure that the deployment of such tools is properly understood. According to EU spokesperson Thomas Regnier, discussions between OpenAI and the European Commission are ongoing, with further meetings planned. Why Anthropic’s Mythos is not yet available While OpenAI has moved ahead with access, Anthropic has not yet provided the EU with preview access to its Mythos model. The EU has confirmed that discussions with Anthropic are taking place, but they are still at an earlier stage. Officials said that while several meetings have been held, no agreement similar to OpenAI’s current arrangement has been reached so far. This difference highlights varying approaches among AI companies when it comes to sharing advanced cybersecurity tools with regulators. Growing focus on AI in cybersecurity The move comes at a time when AI tools are increasingly being used in cybersecurity. Models like GPT-5.5-Cyber are designed to help identify vulnerabilities, analyse threats and support defence systems. Recent developments, including the release of Mythos, have raised concerns about how powerful such tools could be, especially if used without proper oversight. By granting access to the EU, OpenAI is positioning its model as a tool that can be monitored and used responsibly by trusted institutions. What OpenAI has said OpenAI officials have said that cybersecurity should not be managed by companies alone. The company’s Head of OpenAI for Countries, George Osborne, stated that collaboration with governments and institutions is important to ensure safety. He added that making such tools available to trusted partners can help improve security and support public safety efforts. What this means for Europe Access to GPT-5.5-Cyber could allow European institutions to better understand how AI can be used in defence systems. It may also help shape future policies around AI and cybersecurity. At the same time, the delay in accessing Anthropic’s Mythos shows that not all AI tools are being shared at the same pace. What happens next Further discussions between the EU and AI companies are expected in the coming weeks. The outcome of these talks could influence how advanced AI systems are regulated and deployed across Europe. As cybersecurity risks continue to grow, access to such tools may become an important part of how governments prepare for future threats

OpenAI has said it will provide the European Union with access to its new cybersecurity-focused AI model, GPT-5.5-Cyber, while Anthropic has not yet extended similar access for its Mythos model.

OpenAI pushes cyber AI into Europe but Anthropic still not releasing Mythos

OpenAI has rolled out an update for its AI video generator Sora, introducing new tools that let users create reusable characters, stitch clips together, and explore trending videos through leaderboards.<h2>Character cameos: Turning anything into a digital avatar</h2>The biggest addition to the Sora 2 update is the introduction of “character cameos,” a feature that lets users turn almost any image or person into a reusable avatar for AI-generated videos.&nbsp;As per OpenAI’s official release notes, this feature will allow users to create an original persona that can appear in multiple videos. These cameos can be based on people, pets, drawings, or even toys.&nbsp;“Each character comes with its own permissions,” OpenAI explained. “You can keep it private, share it with friends, or make it public for anyone to use.” Users can also assign each character a display name and handle, tagging them to make cameo appearances in videos.&nbsp;The feature builds on an earlier version of Sora that allowed users to make AI-generated likenesses of themselves for use in videos. However, it remains unclear how OpenAI plans to verify if uploaded images represent real or fictional people, especially given ongoing debates about the ethical use of AI-generated likenesses.&nbsp;

&nbsp;<h2>New tools: clip stitching and community leaderboards</h2>OpenAI has also added a video stitching function, allowing users to combine several short clips into longer, multi-scene videos. This gives creators greater flexibility to tell full stories or create more cinematic experiences directly within the Sora app.&nbsp;The update further introduces leaderboards that highlight the most remixed videos and the most popular cameo characters. The feature aims to encourage creativity and community participation by spotlighting top user-generated content.&nbsp;As of now, it is available to users in the United States, Canada, Japan, and South Korea can sign up for Sora without an invitation code to try the new tools.&nbsp;<h2>Trademark dispute and wider availability</h2>The update comes shortly after Cameo, a celebrity video platform, filed a trademark infringement lawsuit against OpenAI for using the term “cameo” in Sora’s features. OpenAI has not commented publicly on the legal dispute.&nbsp;Sora continues to attract attention in the fast-growing AI video space, competing with tools like Runway, Pika, and Synthesia. With these new creative updates, OpenAI is aiming to make Sora more social and user-driven, offering tools that blend AI video generation with community engagement.&nbsp;&nbsp;

OpenAI makes AI video creation more personal and community-driven. OpenAI’s Sora 2 update adds “character cameos,” letting users turn photos or drawings into reusable AI avatars. Making Sora more social, creative, and interactive than ever.

OpenAI has rolled out an update for its AI video generator Sora, introducing new tools that let users create reusable characters, stitch clips together, and explore trending videos through leaderboards. Character cameos: Turning anything into a digital avatar The biggest addition to the Sora 2 update is the introduction of “character cameos,” a feature that lets users turn almost any image or person into a reusable avatar for AI-generated videos. As per OpenAI’s official release notes, this feature will allow users to create an original persona that can appear in multiple videos. These cameos can be based on people, pets, drawings, or even toys. “Each character comes with its own permissions,” OpenAI explained. “You can keep it private, share it with friends, or make it public for anyone to use.” Users can also assign each character a display name and handle, tagging them to make cameo appearances in videos. The feature builds on an earlier version of Sora that allowed users to make AI-generated likenesses of themselves for use in videos. However, it remains unclear how OpenAI plans to verify if uploaded images represent real or fictional people, especially given ongoing debates about the ethical use of AI-generated likenesses.

OpenAI makes AI video creation more personal and community-driven. OpenAI’s Sora 2 update adds “character cameos,” letting users turn photos or drawings into reusable AI avatars. Making Sora more social, creative, and interactive than ever.

Create your own AI character: OpenAI rolls out big 'Sora 2' update

OpenAI launched its new ChatGPT Atlas browser. “This is an AI-powered web browser built around ChatGPT,” OpenAI Chief Executive Officer Sam Altman said during a livestreamed event on Tuesday. He said AI “represents a rare, once-a-decade opportunity” to rethink the browser.&nbsp;He posted on X&nbsp; Our new AI-first web browser, ChatGPT Atlas, is here for macOS.&nbsp;Please send feedback! Availability on other platforms to follow.&nbsp;

<blockquote class="twitter-tweet">Our new AI-first web browser, ChatGPT Atlas, is here for macOS. Please send feedback! Availability on other platforms to follow.&mdash; Sam Altman (@sama) <a href="https://twitter.com/sama/status/1980690768391201180?ref_src=twsrc%5Etfw">October 21, 2025</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>

<h2>A new AI-driven browser experience</h2>As per OpenAI, this new browser allows users to open a ChatGPT sidebar in any window to summarise pages, analyse data, compare products, or complete online tasks such as booking a trip or making purchases.&nbsp;In its agent mode, ChatGPT Atlas can interact directly with websites on behalf of users, performing complex tasks from start to finish.&nbsp;The browser is currently available globally on macOS and will soon roll out to Windows, iOS, and Android, OpenAI said in its official announcement.&nbsp;<h2>Competing in a crowded AI browser market</h2>The launch of ChatGPT Atlas places OpenAI in direct competition with a growing number of AI browsers such as Perplexity’s Comet and Opera’s Neon. These browsers integrate AI tools capable of summarising content, filling forms, and drafting code to make browsing more efficient.&nbsp;Shares of Alphabet, the parent company of Google, dropped 1.6 per cent in afternoon trading following the announcement, reflecting market reaction to the new competition.&nbsp;Google has already begun integrating its Gemini AI assistant into Chrome for U.S. users, with plans to extend support to its iOS app soon.&nbsp;<h2>A challenge to Chrome’s dominance</h2>ChatGPT Atlas represents OpenAI’s effort to expand beyond its core chatbot technology. Since the debut of ChatGPT in 2022, the company has sought new ways to grow amid increasing competition from Google, Anthropic, and other AI startups.&nbsp;By embedding AI directly into the browsing experience, OpenAI aims to create a more interactive and intelligent internet tool one that goes beyond traditional search engines.&nbsp;A recent Bloomberg report noted that this shift comes at a time when Big Tech companies are heavily investing in generative AI, reshaping the online information ecosystem.&nbsp;In a related development, a U.S. federal judge ruled in September that Google can continue paying partners to promote its search engine, acknowledging that generative AI now poses a serious challenge to traditional search dominance.

OpenAI has launched ChatGPT Atlas, a new AI-powered web browser for macOS that blends ChatGPT directly into browsing. It helps users summarise pages, compare products, and complete tasks online. A Windows and mobile rollout will follow soon.

OpenAI launched its new ChatGPT Atlas browser. “This is an AI-powered web browser built around ChatGPT,” OpenAI Chief Executive Officer Sam Altman said during a livestreamed event on Tuesday. He said AI “represents a rare, once-a-decade opportunity” to rethink the browser. He posted on X Our new AI-first web browser, ChatGPT Atlas, is here for macOS. Please send feedback! Availability on other platforms to follow.

OpenAI has launched ChatGPT Atlas, a new AI-powered web browser for macOS that blends ChatGPT directly into browsing. It helps users summarise pages, compare products, and complete tasks online. A Windows and mobile rollout will follow soon.

OpenAI launches ChatGPT Atlas browser to challenge Google Chrome and Perplexity

Chipmaker AMD said on Monday that it will supply artificial intelligence chips to Sam Altman-led OpenAI. The deal could potentially give the ChatGPT creator the option to buy up to roughly 10% of AMD. This deal has sent out a message that OpenAI, a pioneer&nbsp;in consumer-facing artificial&nbsp;intelligence, has confidence in the chips developed by Advanced Micro Devices (AMD).&nbsp;AMD executive vice president Forrest Norrod told Reuters that this deal was&nbsp;certainly transformative for the entire industry. After the announcement, AMD's shares witnessed a massive jump of 25 per cent in premarket trading.&nbsp;Per the deal,&nbsp;OpenAI will build a one-gigawatt facility based on its forthcoming MI450 series of chips beginning next year.&nbsp;AMD expects that it will earn revenue in to tune of $100 billion over the next four years from OpenAI and others.&nbsp;"Other people are going to come along with it because this is really the pioneer, a pioneer in the industry that has a lot of influence over the broader ecosystem," AMD strategy chief Mat Hein told Reuters.&nbsp;AMD will help OpenAI build AI infrastructure.<h2>What's the AMD-OpenAI shares deal?</h2>AMD has issued a warrant that gives OpenAI the ability to buy up to 160 million shares of the company over the course of the deal. These shares will be sold in tranches based on milestones that the two companies intend to achieve.&nbsp;The first tranche of the shares will be vested in OpenAI during the second half of 2026, after the initial shipment of MI450 chips. AMD is valued at a whopping $267.23 billion. OpenAI has a valuation of $500 billion.&nbsp;OpenAI has taken numerous steps to ensure the supply of chips in consonance with the company's AI-visioned future. In September, Nvidia announced an investment of up to $100 billion.&nbsp;The company is also attempting to design its own AI chips with which it has partnered with Broadcom.

AMD-OpenAI deal: After the announcement of the deal, AMD's shares witnessed a massive jump of 25 per cent in premarket trading.

Chipmaker AMD said on Monday that it will supply artificial intelligence chips to Sam Altman-led OpenAI. The deal could potentially give the ChatGPT creator the option to buy up to roughly 10% of AMD. This deal has sent out a message that OpenAI, a pioneer in consumer-facing artificial intelligence, has confidence in the chips developed by Advanced Micro Devices (AMD). AMD executive vice president Forrest Norrod told Reuters that this deal was certainly transformative for the entire industry. After the announcement, AMD's shares witnessed a massive jump of 25 per cent in premarket trading. Per the deal, OpenAI will build a one-gigawatt facility based on its forthcoming MI450 series of chips beginning next year. AMD expects that it will earn revenue in to tune of $100 billion over the next four years from OpenAI and others. "Other people are going to come along with it because this is really the pioneer, a pioneer in the industry that has a lot of influence over the broader ecosystem," AMD strategy chief Mat Hein told Reuters. AMD will help OpenAI build AI infrastructure. What's the AMD-OpenAI shares deal? AMD has issued a warrant that gives OpenAI the ability to buy up to 160 million shares of the company over the course of the deal. These shares will be sold in tranches based on milestones that the two companies intend to achieve. The first tranche of the shares will be vested in OpenAI during the second half of 2026, after the initial shipment of MI450 chips. AMD is valued at a whopping $267.23 billion. OpenAI has a valuation of $500 billion. OpenAI has taken numerous steps to ensure the supply of chips in consonance with the company's AI-visioned future. In September, Nvidia announced an investment of up to $100 billion. The company is also attempting to design its own AI chips with which it has partnered with Broadcom.

AMD-OpenAI deal: After the announcement of the deal, AMD's shares witnessed a massive jump of 25 per cent in premarket trading.

AMD gives OpenAI option to acquire 10 per cent stake in multi-year deal

The artificial intelligence boom, fuelled by the overvaluation of AI stocks and heavy promotion by experts, industry insiders, and investors like Bill Gates, is beginning to crack. Soaring computing costs are not matching returns on investment (ROI). Experts now warn that we are living in an era much like the dot-com bubble of the late 1990s, which burst at the start of this millennium. Will the AI crash hit in 2026? Who will survive it? Here's what you should know.Also read: <a href="https://www.wionews.com/business-economy/nvidia-to-resume-h20-ai-chip-sales-to-china-following-us-policy-reversal-1752578149833">Nvidia to resume H20 AI chip sales to China following US policy reversal</a><h2>AI boom is starting to crack – AI stocks are already falling</h2>Between 2020 and 2025, the AI sector attracted enormous media attention and capital. The frenzy drew comparisons to the dot-com boom. But recent developments suggest that this period of irrational exuberance may soon face a sharp reality check.The stock of AI chip giant Nvidia — the poster child of the AI hardware boom — dropped by nearly 7.4 per cent over two consecutive days. This same stock had surged over 3,000 per cent in recent years. The fall reflects a wave of investor scepticism. CoreWeave, one of Nvidia’s cloud computing partners, lost 33 per cent of its value in just two days, wiping out $24 billion in market capitalisation.<h2>What triggered fears of an AI industry collapse?</h2>A mix of disappointing AI product updates and poor business returns has raised concerns. The underwhelming response to GPT-5, the latest iteration of ChatGPT, is one example. But the bigger shock came from a recent MIT study that found 95 per cent of enterprise generative AI projects had delivered no measurable return on investment.<h2>What the MIT study revealed about AI startups</h2>The MIT study analysed 300 enterprise AI initiatives, conducted 150 executive interviews, and surveyed 350 employees. It found that only 5 per cent of new AI deployments had delivered tangible business value. Many initiatives were misaligned with actual business needs, prioritising flashy features over cost-saving or efficiency—supposedly AI’s core strengths. Internally developed systems failed more often than vendor-supplied solutions, suggesting poor planning, integration, and execution in many companies.<h2>Sam Altman and others warn of overvaluation</h2>Recently, Sam Altman, CEO of OpenAI, publicly compared today’s AI fervour to the dot-com bubble. He warned of "overexuberant investment."Torsten Sløk of Apollo Global Management added that AI valuations are “more extreme than during the dot-com era,” with vastly greater speculative capital at risk.Also read: <a href="https://www.wionews.com/world/-chatgpt-hallucinates-don-t-over-rely-on-it-openai-ceo-sam-altman-warns-users-1751366853229">‘ChatGPT ‘hallucinates’, don’t overrely on it,’ OpenAI CEO Sam Altman warns users</a>&nbsp;<h2>Valuations far exceed real earnings</h2>AI companies are being valued at levels far above their actual revenue. OpenAI is reportedly valued at $157 billion — roughly 40 times its estimated annual income. Anthropic is targeting a 50× valuation. Deutsche Bank has warned that these levels are unsustainable and a sharp correction is likely.Also read: <a href="https://www.wionews.com/photos/how-will-openai-s-gpt-5-change-the-way-you-use-chatgpt-1754649832228">How will OpenAI’s GPT-5 change the way you use ChatGPT?</a><h2>Computing and infrastructure costs could make ai startups unviable</h2>AI model training requires immense computing resources. Models like GPT-4 cost around $100 million to train. Since 2016, training costs have doubled approximately every year.Environmental and infrastructure costs are also climbing. GPT-4 consumed around 50 GWh of energy during training. Water usage by AI data centres is projected to exceed 6.6 billion cubic metres per year by 2027—more than half of the UK’s annual water use.&nbsp;<a href="https://www.wionews.com/videos/shared-chatgpt-links-publicly-visible-in-google-search-1754072068804/embed">https://www.wionews.com/videos/shared-chatgpt-links-publicly-visible-in-google-search-1754072068804/embed</a><h2>Will the AI bust mirror the dot-com crash?</h2>A Financial Times analysis suggests that while a market shake-up is likely, today’s Big Tech firms—Amazon, Google, Microsoft—are built on strong earnings and operational capacity. They are better positioned to survive a downturn. The firms most at risk are small or mid-sized startups without deep financial reserves or real-world utility.<h2>Real-world use, not novelty, will determine survival</h2>As Bret Taylor, chairman of OpenAI, put it, only companies that deliver "real-world use" will persist. He noted that after the dot-com crash of 2000–2001, only those offering genuine utility, like Google and Amazon, survived.The AI boom may not be over, but it is entering a phase of correction and realism. The most optimistic view is that any “apocalypse” will ultimately consolidate the sector, leaving only a few serious players with sustainable models and deep pockets. &nbsp;

The Artificial Intelligence or AI industry is seeing a looming bust, similar to the dot-com crash. A recent study found that 95 per cent of generative AI pilots failed to deliver return on investment. Who will survive the crash?

The artificial intelligence boom, fuelled by the overvaluation of AI stocks and heavy promotion by experts, industry insiders, and investors like Bill Gates, is beginning to crack. Soaring computing costs are not matching returns on investment (ROI). Experts now warn that we are living in an era much like the dot-com bubble of the late 1990s, which burst at the start of this millennium. Will the AI crash hit in 2026? Who will survive it? Here's what you should know. Also read: Nvidia to resume H20 AI chip sales to China following US policy reversal AI boom is starting to crack – AI stocks are already falling Between 2020 and 2025, the AI sector attracted enormous media attention and capital. The frenzy drew comparisons to the dot-com boom. But recent developments suggest that this period of irrational exuberance may soon face a sharp reality check. The stock of AI chip giant Nvidia — the poster child of the AI hardware boom — dropped by nearly 7.4 per cent over two consecutive days. This same stock had surged over 3,000 per cent in recent years. The fall reflects a wave of investor scepticism. CoreWeave, one of Nvidia’s cloud computing partners, lost 33 per cent of its value in just two days, wiping out $24 billion in market capitalisation. What triggered fears of an AI industry collapse? A mix of disappointing AI product updates and poor business returns has raised concerns. The underwhelming response to GPT-5, the latest iteration of ChatGPT, is one example. But the bigger shock came from a recent MIT study that found 95 per cent of enterprise generative AI projects had delivered no measurable return on investment. What the MIT study revealed about AI startups The MIT study analysed 300 enterprise AI initiatives, conducted 150 executive interviews, and surveyed 350 employees. It found that only 5 per cent of new AI deployments had delivered tangible business value. Many initiatives were misaligned with actual business needs, prioritising flashy features over cost-saving or efficiency—supposedly AI’s core strengths. Internally developed systems failed more often than vendor-supplied solutions, suggesting poor planning, integration, and execution in many companies. Sam Altman and others warn of overvaluation Recently, Sam Altman, CEO of OpenAI, publicly compared today’s AI fervour to the dot-com bubble. He warned of "overexuberant investment." Torsten Sløk of Apollo Global Management added that AI valuations are “more extreme than during the dot-com era,” with vastly greater speculative capital at risk. Also read: ‘ChatGPT ‘hallucinates’, don’t overrely on it,’ OpenAI CEO Sam Altman warns users Valuations far exceed real earnings AI companies are being valued at levels far above their actual revenue. OpenAI is reportedly valued at $157 billion — roughly 40 times its estimated annual income. Anthropic is targeting a 50× valuation. Deutsche Bank has warned that these levels are unsustainable and a sharp correction is likely. Also read: How will OpenAI’s GPT-5 change the way you use ChatGPT? Computing and infrastructure costs could make ai startups unviable AI model training requires immense computing resources. Models like GPT-4 cost around $100 million to train. Since 2016, training costs have doubled approximately every year. Environmental and infrastructure costs are also climbing. GPT-4 consumed around 50 GWh of energy during training. Water usage by AI data centres is projected to exceed 6.6 billion cubic metres per year by 2027—more than half of the UK’s annual water use. https://www.wionews.com/videos/shared-chatgpt-links-publicly-visible-in-google-search-1754072068804/embed Will the AI bust mirror the dot-com crash? A Financial Times analysis suggests that while a market shake-up is likely, today’s Big Tech firms—Amazon, Google, Microsoft—are built on strong earnings and operational capacity. They are better positioned to survive a downturn. The firms most at risk are small or mid-sized startups without deep financial reserves or real-world utility. Real-world use, not novelty, will determine survival As Bret Taylor, chairman of OpenAI, put it, only companies that deliver "real-world use" will persist. He noted that after the dot-com crash of 2000–2001, only those offering genuine utility, like Google and Amazon, survived. The AI boom may not be over, but it is entering a phase of correction and realism. The most optimistic view is that any “apocalypse” will ultimately consolidate the sector, leaving only a few serious players with sustainable models and deep pockets.

What we're covering here: &nbsp;The Artificial Intelligence or AI industry is seeing a looming bust, similar to the dot-com crash. A recent study found that 95 per cent of generative AI pilots failed to deliver return on investment. Who will survive the crash?

AI APOCALYPSE is coming: The bubble of overvalued stocks, underperforming apps will burst soon, but who will survive?

Elon Musk says Grok 4 Heavy is smarter than OpenAI’s new GPT-5Elon Musk has posted on X that his company’s AI model, Grok 4 Heavy, is still more intelligent than OpenAI’s latest model, GPT-5.&nbsp;In a post on X, Musk said:&nbsp;“Bottom line though: Grok 4 Heavy was smarter 2 weeks ago than GPT-5 is now and G4H is already a lot better.”

<blockquote class="twitter-tweet">Bottom line though: Grok 4 Heavy was smarter 2 weeks ago than GPT5 is now and G4H is already a lot better. Let that sink in. <a href="https://t.co/BrggsEwnuz">https://t.co/BrggsEwnuz</a>&mdash; Elon Musk (@elonmusk) <a href="https://twitter.com/elonmusk/status/1953509618493403404?ref_src=twsrc%5Etfw">August 7, 2025</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>

&nbsp;He Also posted another X post in which he said -&nbsp;“Grok 5 will be out before the end of this year and it will be crushingly good.”

<blockquote class="twitter-tweet">Grok 5 will be out before the end of this year and it will be crushingly good <a href="https://t.co/W0SP9bu1GH">https://t.co/W0SP9bu1GH</a>&mdash; Elon Musk (@elonmusk) <a href="https://twitter.com/elonmusk/status/1953515379688714697?ref_src=twsrc%5Etfw">August 7, 2025</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>

&nbsp;<h2>What’s new in GPT-5?</h2>While Musk made his claims, OpenAI said GPT-5 is a major upgrade over previous versions of ChatGPT. During its launch livestream, the company highlighted improvements in coding, maths, writing, health advice, and visual perception.&nbsp;The new model can generate apps, websites, and games within minutes, using only natural language prompts. GPT-5 is also the first OpenAI model that can automatically choose between a reasoning mode or efficiency mode based on the question asked.&nbsp;<h2>Availability and usage limits</h2>OpenAI has made GPT-5 available for free to all users, but with a usage limit. Once a free user reaches that limit, the system will switch to GPT-5 Mini, a smaller but capable version of the model.&nbsp;Plus plan users get higher limits, while Pro plan subscribers can use GPT-5 Pro, which is faster and more powerful.

Elon Musk says Grok 4 is still smarter than GPT-5, even after OpenAI’s latest launch. While GPT-5 brings big upgrades in coding, reasoning, and visuals, Musk teases that Grok 5 will be “crushingly good.”

Elon Musk says Grok 4 Heavy is smarter than OpenAI’s new GPT-5 Elon Musk has posted on X that his company’s AI model, Grok 4 Heavy, is still more intelligent than OpenAI’s latest model, GPT-5. In a post on X, Musk said: “Bottom line though: Grok 4 Heavy was smarter 2 weeks ago than GPT-5 is now and G4H is already a lot better.”

Elon Musk says Grok 4 is still smarter than GPT-5, even after OpenAI’s latest launch. While GPT-5 brings big upgrades in coding, reasoning, and visuals, Musk teases that Grok 5 will be “crushingly good.”

Elon Musk says Grok 4 is still smarter than OpenAI’s GPT-5

OpenAI GPT-5 upgrades bring faster reasoning, fewer errors, and a huge context window for better healthcare and education support. It helps patients understand medical info, aids teachers with personalised lessons, and offers safer, clearer responses. 

How will OpenAI GPT-5 transform healthcare and education support?

openai-gpt-5

Openai Gpt 5, News on Openai Gpt 5, stories on Openai Gpt 5, articles on Openai Gpt 5, Openai Gpt 5 stories at Wion, Openai Gpt 5 news, Openai Gpt 5 latest news

openai gpt-5