Nvidia Chips Drive AI While Google Forecasts Trillion Revenue

Investors are looking beyond the current AI stock excitement for strong returns in 2026, shifting focus from infrastructure spending to profit generation. US equity valuations appear high, with the market concentrated on a few large tech companies. Opportunities for steady returns may emerge in international markets, including Europe, Japan, emerging economies, and the UK, as noted by Helen Jewell of BlackRock. Analysts remain positive on semiconductor companies, viewing chips as crucial for AI growth. Bank of America recommends large-cap picks such as Nvidia, Broadcom, Lam Research, KLA, Analog Devices, and Cadence Design Systems. Jefferies specifically highlights Broadcom for its custom chip business. Bank of America also identifies Alphabet (Google) as a top AI leader, projecting AI will generate over $1 trillion in new revenue within five years. The regulatory landscape for AI is seeing significant developments. On December 22, 2025, Erich Mische criticized President Trump's executive order on AI, arguing it attempts to prevent states from enacting child protection laws against AI harm, similar to Section 230. In contrast, New York Governor Kathy Hochul signed the Responsible AI Safety and Education (RAISE) Act, requiring AI developers to test models for safety and disclose product details, positioning New York as a leader in AI regulation. Ethical and legal challenges are also prominent. Adobe faces a proposed class-action lawsuit for allegedly using pirated books, including those from the controversial Books3 collection, to train its SlimLM AI model. Separately, SageMaster, an AI-powered crypto and forex trading platform, has drawn global warnings from financial authorities like Luxembourg's CSSF and Austria's FMA due to its unauthorized financial services and its founders' past ties to the OneCoin fraud. AI continues to advance in diverse applications and foundational development. In healthcare, AI shows promise in transforming early-stage Non-Small Cell Lung Cancer (NSCLC) diagnosis and treatment by simplifying data analysis, understanding genetic information, and predicting mutations. Concurrently, Tether Data's AI division, QVAC, released QVAC Genesis II on December 22, 2025, a massive synthetic dataset adding 107 billion tokens, designed to teach AI models "Option-Level Reasoning" for better thinking and explanation. The role of human wisdom alongside AI is also a key discussion point. Jonathan J. Sanford, president of the University of Dallas, emphasized on December 22, 2025, that while AI automates tasks, it cannot replace the wise leaders needed for strategic planning and ethical decision-making, advocating for a strong liberal arts education. In a technical update, Linux 7.0 is dropping support for AMD's unreleased NPU2 chip, an early AI accelerator, to reduce ongoing maintenance costs for its open-source driver.

Key Takeaways

• Investors will shift focus from AI infrastructure spending to profit generation in 2026, seeking returns beyond current AI stock hype.
• Bank of America recommends large-cap semiconductor companies like Nvidia, Broadcom, Lam Research, KLA, Analog Devices, and Cadence Design Systems for AI growth.
• Alphabet (Google) is identified as a top AI leader, with AI expected to generate over $1 trillion in new revenue in five years.
• Erich Mische criticized President Trump's executive order on AI for potentially hindering state-level child safety protections against AI harm.
• New York enacted the Responsible AI Safety and Education (RAISE) Act, requiring AI developers to test models for safety and disclose product details, positioning it as a leader in AI regulation.
• Adobe faces a class-action lawsuit for allegedly training its SlimLM AI model using pirated books from the Books3 dataset.
• SageMaster, an AI crypto/forex trading platform, received global warnings from financial authorities due to unauthorized services and founders' ties to the OneCoin fraud.
• AI is poised to transform early-stage Non-Small Cell Lung Cancer (NSCLC) diagnosis and treatment through improved biomarker testing and personalized care.
• Tether Data's QVAC released QVAC Genesis II, a synthetic AI training dataset totaling 148 billion tokens, featuring "Option-Level Reasoning" for enhanced AI thinking.
• Linux 7.0 is dropping support for AMD's unreleased NPU2 AI chip to reduce maintenance costs, while other AMD NPUs remain supported.

Investors Seek Returns Beyond AI Hype in 2026

Investors are looking for strong returns in 2026 beyond the current excitement around artificial intelligence stocks. US equity valuations appear high, and the market is heavily focused on a few large tech companies. Opportunities may be found in international markets like Europe, Japan, emerging economies, and the UK. European stocks could benefit from economic growth and defense spending, while Japan's market may see gains from inflation and corporate changes. Emerging markets and UK companies also offer potential for steady returns, according to Helen Jewell of BlackRock.

AI Stock Focus Shifts to Profits for 2026

On December 21, 2025, the focus for AI stocks shifted from spending on infrastructure to how companies make money and generate returns. Analysts like Bank of America and Jefferies remain positive on semiconductor companies, seeing chips as crucial for AI growth. Bank of America recommends large-cap picks including Nvidia, Broadcom, Lam Research, KLA, Analog Devices, and Cadence Design Systems. Jefferies highlights Broadcom for its custom chip business. Bank of America also sees Alphabet as a top AI leader, expecting AI to create over $1 trillion in new revenue in five years.

Trump AI Order Threatens Child Safety Protections

On December 22, 2025, Erich Mische criticized President Trump's executive order on artificial intelligence. Mische argues the order tries to stop states from making their own laws to protect children from AI harm. This move is similar to Section 230 of the Communications Decency Act, which has protected tech companies from being responsible for harmful content online for decades. Congress recently voted against a similar ban on state AI regulation, showing bipartisan support for states to act. Mische, a CEO of a suicide prevention organization, believes this order puts profit over child safety.

SageMaster AI Trading Platform Faces Global Warnings

SageMaster is an AI-powered platform for crypto and forex trading that uses a multi-level marketing model through its BE Club ecosystem. While it promises easy trading, financial authorities like Luxembourg's CSSF and Austria's FMA have issued warnings. They state SageMaster TechHost WorldWide Inc. is not authorized to provide financial services. The company is registered in Anguilla, an offshore location, and its key figures, Monir Islam, Moynul Islam, and Ehsaan Islam, have past ties to the OneCoin fraud. Warnings about SageMaster have been issued in many countries since 2020, raising serious concerns for users.

AI Transforms Lung Cancer Diagnosis and Treatment

Artificial intelligence has the power to change how doctors test for biomarkers in early-stage Non-Small Cell Lung Cancer, or NSCLC. AI can make data analysis easier, help understand complex genetic information, and predict important mutations from scans. Despite progress, challenges remain in diagnosing NSCLC, such as low screening rates and difficulty getting enough tissue for testing. AI could help create personalized treatments, improve care coordination, and use predictive tools to enhance patient outcomes and lower cancer recurrence.

AI Needs Human Wisdom for True Leadership

On December 22, 2025, Jonathan J. Sanford, president of the University of Dallas, stated that while AI changes business by automating tasks, it cannot create the wise leaders needed for complex decisions. He argues that leaders must possess wisdom for critical operations like strategic planning and crisis management. Sanford suggests that a strong liberal arts education, focusing on timeless texts, logic, philosophy, ethics, history, and literature, can develop the human strengths AI lacks. This education helps people understand nuance, make ethical choices, and guide technology for the common good.

Linux 7.0 Drops Support for Unreleased AMD NPU2

Linux 7.0 will no longer support AMD's NPU2, an early artificial intelligence chip that was never released in any product. The AMD engineer Lizhi Hou confirmed NPU2 is obsolete and was never publicly available. The AMDXDNA accelerator driver in the Linux kernel currently supports other AMD NPUs like NPU1, NPU4, NPU5, and NPU6. Removing the NPU2 code will help reduce ongoing maintenance costs for the open-source driver.

Tether Releases Massive AI Training Dataset

On December 22, 2025, Tether Data's AI division, QVAC, released QVAC Genesis II, a huge new synthetic dataset for training artificial intelligence. This update adds 107 billion tokens, bringing the total to 148 billion across 19 educational subjects, including chemistry and computer science. The dataset uses a new "Option-Level Reasoning" method to help AI models learn to think, reason, and explain, rather than just mimic. Tether's CEO, Paolo Ardoino, stated this release moves beyond just data volume to focus on structure, reasoning, and clarity.

Adobe Sued Over Pirated Books in AI Training

Adobe faces a proposed class-action lawsuit for allegedly using pirated books to train its SlimLM artificial intelligence model. Oregon author Elizabeth Lyon claims her copyrighted works were included without permission or payment. The lawsuit states that SlimLM was trained using the SlimPajama-627B dataset, which contains material from Books3, a controversial collection of nearly 200,000 copyrighted books. This case adds to the increasing legal challenges in the AI industry regarding the origin of training data and copyright rules.

NY Enacts Strict AI Safety and Prison Laws

New York Governor Kathy Hochul signed several important bills into law, including new rules for state prisons and a significant AI safety act. The prison bill requires more cameras after two inmate deaths, aiming to improve safety and accountability. She also signed the Responsible AI Safety and Education Act, or RAISE Act, which makes AI developers test their models for safety and share details about their products. This new law is tougher than California's and positions New York as a leader in AI regulation while federal rules are still missing.

Sources

• Which bullish 2026 equity plays don’t require AI euphoria?
• AI Stocks News Today (Dec. 21, 2025): Nvidia, Broadcom and Alphabet Set the Tone for 2026 as Wall Street Weighs Bubble Risks
• Trump's order on artificial intelligence is a potential disaster
• SageMaster: AI-powered trading platform thrives on subscriptions and recruitment, amid warnings from authorities (ANALYSIS)
• Artificial Intelligence and the Future of NSCLC | Targeted Oncology - Immunotherapy, Biomarkers, and Cancer Pathways
• AI is revolutionizing business but it can’t form the leaders we need
• Linux 7.0 To Remove Support For AMD's Never-Released Ryzen AI NPU2
• Tether Unveils Open Dataset for AI Training
• Adobe Faces Proposed Class-Action Lawsuit Over Alleged Use of Pirated Books in AI Training
• New rules for NY prisons, artificial intelligence as Gov. Hochul signs stack of bills into law