Nvidia Refocuses AI Hardware While Google DeepMind Develops Safety Tools

Global AI investments are expanding significantly beyond just computer chips in 2025, reaching new sectors like telecommunications and satellite technology. These industries are crucial for supporting the massive data demands of AI applications, including fiber optics and low-Earth orbit satellites. Goldman Sachs projects that tech giants will spend over $500 billion on AI infrastructure by 2026. Reflecting this broader push, the National Institute of Standards and Technology (NIST) is investing $20 million to establish two new centers with Mitre Corp., aiming to boost AI in US manufacturing and cybersecurity. In a strategic shift, Nvidia is stepping back from competing directly with cloud giants like Amazon Web Services. CEO Jensen Huang's DGX Cloud service, intended to offer AI chips to businesses, struggled to gain traction. Nvidia has now restructured its cloud team to focus on internal research and development, reinforcing its core business of powerful AI hardware. Meanwhile, Google DeepMind released Gemma Scope 2, a new suite of open tools designed to enhance AI safety by helping researchers understand how Gemma 3 language models process information and exhibit unexpected behaviors like "jailbreaking" or "hallucinating." The integration of AI also presents new challenges and opportunities across various industries. Cybersecurity experts are warning about "Grokking," a scam where cybercriminals exploit X's AI tool, Grok, to spread malicious links through promoted videos. Users are advised to avoid clicking links from Grok in such threads. In the news sector, leaders from CNN, The New York Times, and Reuters discussed how AI will transform content creation, envisioning journalists as developers and news archives as smart databases, potentially rendering traditional websites obsolete. Companies are actively preparing their workforces for this AI-driven future; PwC, for instance, is training new US associates in tax, assurance, and advisory on AI tools through interactive, in-person sessions. However, the promise of improved work-life balance with AI remains uncertain for many. Fortune 500 CEOs, including Nvidia's Jensen Huang and Palantir's Alex Karp, suggest that the intense AI race demands significant dedication, implying less social life for those seeking early success. Even dating apps like Match Group and Bumble are investing millions in generative AI features, such as Tinder's "Chemistry" pilot, to combat "app fatigue" and create more meaningful connections. The National Hockey League (NHL) also weighed in, launching its "AI Can't NHL" ad campaign, which highlights the unique human skills and instincts in hockey that artificial intelligence cannot replicate.

Key Takeaways

  • Goldman Sachs projects tech giants will spend over $500 billion on AI infrastructure by 2026.
  • AI investments are expanding beyond computer chips into sectors like telecommunications and satellite technology.
  • Nvidia is refocusing on its core AI hardware business, stepping back from cloud services like DGX Cloud after struggling against competitors such as Amazon Web Services.
  • Google DeepMind released Gemma Scope 2, a new suite of tools to enhance AI safety by understanding Gemma 3 language models and identifying unexpected behaviors.
  • Cybercriminals are exploiting X's AI tool, Grok, to spread malicious links through promoted videos, a scam dubbed "Grokking."
  • NIST is investing $20 million to establish two centers with Mitre Corp. to boost AI in US manufacturing and cybersecurity.
  • PwC is training new US associates in tax, assurance, and advisory departments on AI tools through in-person immersion sessions.
  • Dating apps, including Match Group and Bumble, are investing millions in generative AI features, such as Tinder's "Chemistry" pilot, to improve user matching and combat "app fatigue."
  • CEOs like Nvidia's Jensen Huang and Palantir's Alex Karp suggest that AI may not lead to improved work-life balance, emphasizing the high dedication required in the AI era.
  • The National Hockey League (NHL) launched its "AI Can't NHL" ad campaign, highlighting the unique human skills and instincts in hockey that artificial intelligence cannot replicate.

AI Investments Expand Beyond Chips in 2025

In 2025, AI investments are expanding past just computer chips into new sectors. Experts like Scott Kupor from Diameter Capital Partners highlight growth in telecommunications and satellite technology. These industries support the massive data needs of AI applications, including fiber optics and low-Earth orbit satellites. Goldman Sachs projects tech giants will spend over $500 billion on AI infrastructure by 2026. This shift shows AI's impact reaches many industries that help it grow and scale.

Nvidia Shifts Focus Back to AI Hardware Not Cloud

Nvidia Corp. is changing its strategy and stepping back from competing with cloud giants like Amazon Web Services. CEO Jensen Huang had planned a cloud service called DGX Cloud in 2023 to offer AI chips to businesses. However, the service struggled to gain users against established companies. Nvidia has now restructured its cloud team to focus more on internal research and development. This move helps Nvidia strengthen its main business of making powerful AI hardware like GPUs.

Cybercriminals Use X AI Grok to Spread Bad Links

Cybersecurity experts warn about a new scam on X, formerly Twitter, called "Grokking." Scammers use X's AI tool, Grok, to trick users into clicking dangerous links. They post promoted videos with hidden malicious links, then ask Grok where the video is from. Grok then replies with the bad link, making it seem trustworthy. Clicking these links can install malware or steal personal information. Experts advise users never to click links from Grok in promoted video threads and to report suspicious posts.

Google DeepMind Unveils Gemma Scope 2 for AI Safety

Google DeepMind researchers released Gemma Scope 2, a new set of tools to understand how Gemma 3 language models work. This open suite helps AI safety teams see how these models process information at every level. It uses sparse autoencoders and transcoders to trace model behavior, especially when models act unexpectedly like "jailbreaking" or "hallucinating." Gemma Scope 2 covers all Gemma 3 models, from small to large, and helps researchers study complex behaviors for better AI safety.

News Leaders Discuss AI Future for News Content

At the NewsTECHForum, leaders from CNN, The New York Times, Reuters, and Hacks/Hackers discussed the future of news. They talked about "Liquid Content" and how AI will change news products. Their vision includes journalists becoming developers and news archives turning into smart databases. They also believe that traditional websites might eventually become outdated.

NHL Launches New Ad Campaign AI Can't NHL

The National Hockey League (NHL) released a new marketing ad called "AI Can't NHL." This spot is part of their "The Next Golden Era Is Now" campaign. It features young stars like Charlie McAvoy, Cale Makar, Jake Oettinger, and Quinn Hughes. The ad shows how human skills and instincts in hockey are unique and cannot be copied by artificial intelligence. The new ad debuted on December 23, 2025, during NHL games on TNT.

PwC Trains New Staff on AI Tools for Future Work

PwC, a major accounting firm, is training its new US associates on artificial intelligence tools. These AI immersion sessions are being piloted in tax, assurance, and advisory departments. The goal is to give new employees hands-on experience and prepare them for changing roles with AI. Margaret Burke and Krishnan Chandrasekhar lead this effort, believing AI will elevate associates' work. PwC chose in-person, interactive training to foster collaboration and problem-solving with AI tools.

NIST Invests 20 Million to Boost AI in Manufacturing

The National Institute of Standards and Technology, NIST, is investing 20 million dollars to boost AI in manufacturing and cybersecurity. This federal agency will create two new centers in partnership with the nonprofit Mitre Corp. These centers aim to help US companies use AI to make products more efficiently and strengthen national security. Acting NIST Director Craig Burkhardt stated their goal is to remove barriers to American AI innovation. The centers will be called the AI Economic Security Center for U.S. Manufacturing Productivity and the AI Economic Security Center to Secure U.S. Critical Infrastructure from Cyberthreats.

Top CEOs Say AI Will Not Bring More Work Life Balance

Many workers hope AI will improve their work-life balance in 2026, especially Gen Z who value it highly. However, several Fortune 500 CEOs warn that AI may not lead to more free time. Nvidia CEO Jensen Huang works seven days a week, driven by the intense AI race. Zoom CEO Eric Yuan believes "work is life, life is work," though he puts family first in conflicts. Palantir CEO Alex Karp suggests that early success often means less social life. These leaders imply that the AI era demands even more dedication, not less.

Dating Apps Use AI to Find Your Perfect Match

Dating apps are investing millions in AI to make users love them again and fight "app fatigue." Companies like Match Group and Bumble are struggling with declining users and stock prices. They hope new generative AI features will lead to better matches and fewer swipes. Match Group CEO Spencer Rascoff calls AI a "platform shift" that will change everything. Tinder is piloting a "Chemistry" feature that suggests daily matches based on values and photos, aiming for more meaningful connections.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI Investments Telecommunications Satellite Technology Data Needs AI Infrastructure Fiber Optics Low-Earth Orbit Satellites Tech Giants Nvidia AI Hardware GPUs Cloud Services R&D X AI Grok Cybersecurity AI Safety Google DeepMind Gemma Scope 2 Language Models Gemma 3 Model Behavior Jailbreaking Hallucinating AI in News Media Industry NHL Marketing AI Limitations Human Skills Sports PwC AI Training Workforce Development AI Tools Accounting Firm NIST AI in Manufacturing National Security US Companies AI Innovation Critical Infrastructure Work-Life Balance AI Impact CEOs Workforce Gen Z Dating Apps Generative AI Match Group Bumble Personalization AI Features

Comments

Loading...