Nvidia's AI Stock Valuation Not a Bubble, Says DBS CIO

Key Takeaways

• DBS CIO Hou Wey Fook says Nvidia's 17x forward earnings and 70% projected growth show AI stocks are not in bubble territory.
• Seattle will ban AI-driven price discrimination at grocery chains with 20+ locations after investigations found up to a 23% price gap for identical items.
• Meta's AI agent Muse builds profiles of people in users' lives without their consent, updating hourly with names, relationships, and topics.
• Thomson Reuters posted 9% revenue growth and 8% organic growth in Q2, with 82% recurring revenue, as it pushes fiduciary-grade AI tools.
• AM Intelligence plans to invest over $20 billion in AI data centers, ordering 20,000 NVIDIA Rubin GPUs for facilities in India and Malaysia.
• 55% of consumers want AI to speed up checkout at convenience stores, and 76% would share more data for faster service.
• 48% of financial advisors say AI improved efficiency, but 56% still cite heavy admin work as a barrier to productivity.
• Router-gradient sensitivity reduces refusals more than activation frequency in 24 of 25 test conditions for safety-critical MoE language models.
• The CuBEs study found standard AI benchmarks miss important behavioral differences across 12 cultures, especially in non-Western political contexts.
• Hou Wey Fook recommends a barbell strategy pairing technology stocks with investment-grade fixed income yielding above 5%.

Nvidia valuation shows AI stocks are not a bubble says DBS CIO

DBS Group Chief Investment Officer Hou Wey Fook says Nvidia's valuation shows AI stocks are not in bubble territory. Nvidia trades at 17 times its 12-month forward earnings. The company's earnings are projected to grow 70% next year. Hou compared this to Cisco Systems which traded at about 100 times earnings before the dot-com crash. He still recommends a barbell approach to balance portfolio risk.

Nvidia valuations show AI rally is not a bubble DBS says

DBS Group Chief Investment Officer Hou Wey Fook says Nvidia's 17 times forward earnings and 70% projected growth show AI stocks are not a bubble. He compared Nvidia to Cisco Systems which traded at 100 times earnings before the dot-com crash. Hou recommends a barbell strategy pairing technology stocks with investment-grade fixed income. Hedge funds and gold serve as risk diversifiers in the middle.

DBS says Nvidia valuation shows AI rally is not a bubble

DBS Group CIO Hou Wey Fook says Nvidia's 17 times forward earnings and 70% projected earnings growth prove AI stocks are fairly valued. He contrasts this with Cisco Systems which traded at 100 times earnings before the dot-com crash. Hou advises pairing technology with investment-grade fixed income yielding above 5%. He calls this a barbell strategy to control portfolio volatility with hedge funds and gold in the middle.

Nvidia valuation gets surprising take from top bank

DBS Group Chief Investment Officer Hou Wey Fook says Nvidia's valuation is hard to compare to the dot-com bubble. He noted Cisco Systems traded at about 100 times earnings before that crash while Nvidia trades at 17 times forward earnings. Hou believes there are still tailwinds for AI and semiconductors. He recommends a barbell strategy combining growth tech with investment-grade fixed income.

Nvidia valuation gives AI bubble fears a new counterargument

DBS Group Chief Investment Officer Hou Wey Fook does not believe AI stocks have reached bubble territory. He cites Nvidia's 17 times 12-month price-to-earnings multiple and strong projected profit growth as key evidence. Hou argues that rising share prices do not always signal a bubble when profits are growing fast. Nvidia remains a crucial test for the larger AI industry going forward.

CIOs can better measure AI ROI with these three steps

CIOs should use a portfolio approach to evaluate all AI projects together. They should run focused pilots before scaling and track both costs and business benefits. A strong framework includes input metrics, health metrics, and outcome metrics to show real return on AI investment.

CIOs should shift AI spending toward business value

CIOs are moving from simply controlling AI costs to deciding where AI investments create business results. Instead of tracking tokens, they should link AI use to revenue, margin, and customer outcomes. The best approach starts small by measuring one or two key business metrics tied directly to AI activities.

Consumers want faster AI checkout at convenience stores

A PAR survey found 33% of consumers would switch stores for faster checkout. 55% want AI to speed up checkout, and 76% would share more data for that benefit. Parents are more pro-AI, and trust depends on transparency and access to real employees.

New method finds safety-critical experts in sparse MoE LLMs

A study tested router-gradient sensitivity as a way to find safety-critical experts in sparse Mixture-of-Experts language models. Gradient selection reduced refusals more than activation frequency in 24 of 25 test conditions. The approach worked across five MoE architectures without degrading model outputs.

Advisors adopt AI fast but still lose time to admin work

The 2026 Morningstar study found AI use among advisors rose, with 48% saying it improved efficiency. Still, 56% cite heavy admin work as a barrier, and advisors spend 53% of their week on client work but want 63%. AI helps most with communication tasks, not investment decisions.

Meta Muse AI profiles people in your life without their consent

Meta's AI agent Muse creates detailed profiles of people in users' lives even if those people never use the app. Researcher Karan Joshi found that Muse updates these profiles hourly with names, relationships, birthdays, and recurring topics. The profiles include information about how close users are to each person and what their relationship needs. Muse launched on September 8 and quickly became the top free app on Apple's US App Store. Meta says the data comes from public information and what users choose to share.

Thomson Reuters may be the next AI winner with trusted data

A financial analysis suggests Thomson Reuters is well positioned for the next phase of the AI boom because it owns trusted professional data. The company provides information tools like Westlaw and Checkpoint to lawyers, accountants, and tax professionals. Second-quarter revenue grew 9 percent with 8 percent organic growth, and recurring revenue made up 82 percent of total revenue. Thomson Reuters has launched CoCounsel Legal and its own Thomson LLM, aiming for what it calls fiduciary-grade AI. The stock recently traded at $136.61, down about 42 percent from its 52-week high.

CuBEs study shows culture affects how AI models behave

Researchers created CuBEs, or Culturally-situated Behavior Evaluations, to test how large language models behave across different cultures. The study examined 13 open and closed source AI models across 12 distinct cultures. They found that standard tests that ignore culture miss important behavioral differences. For example, political bias tests that work for Western politics missed different issues like religious and colonial politics in non-Western cultures. The findings show that AI evaluations need cultural context to be accurate for global users.

Seattle bans AI-driven price changes at grocery stores

Seattle Mayor Katie Wilson is set to sign the Fair and Transparent Pricing ordinance on Monday after a 7-2 city council vote. The law bans large grocery chains with at least 20 locations worldwide from using personal data, browsing history, or AI to set different prices for customers. Stores must post prices available to all shoppers. The ordinance was partly inspired by 2025 investigations that found up to a 23 percent price gap for the same item at the same Safeway store. Union leaders support the move to stop what they call surveillance pricing.

AM Intelligence plans over $20 billion for AI data centers

AM Intelligence announced plans to invest over $20 billion in the next 15 months to expand AI data center capacity. The company placed orders for 20,000 NVIDIA Rubin GPUs for facilities in India and Malaysia, bringing its total GPU commitments to about 29,000. The expansion will increase data center capacity to 100 Megawatts initially, with plans for an additional 300 MWs. GPU delivery is scheduled for the second quarter of 2027. AM Intelligence aims to eventually reach 1 GW of compute-as-a-service capacity globally.

Sources

• Nvidia’s valuation shows AI stocks are far from a bubble, DBS CIO says
• Nvidia’s Valuations Show AI Rally Isn’t a Bubble, DBS Says
• DBS says Nvidia's valuation shows the AI rally isn't a bubble
• Nvidia Valuation Gets Surprising Take From Top Bank
• Nvidia's Valuation Is Giving AI Bubble Fears a New Counterargument
• 3 best practices for CIOs to assess AI ROI
• The real AI budget question: Where should we spend more?
• PAR: Consumers Value Fast AI Checkout
• Frequency Is Not Sensitivity Identifying Safety-Sensitive Experts in Sparse MoE LLM
• Advisors Adopt AI Fast, but Still Aren’t Getting Their Time Back, Says 2026 Morningstar Investor Perspectives Advisor Study
• Meta's Muse Is Collecting Information on Everyone in Your Life
• The Next AI Winners May Own Trusted Data: I’d Watch This Canadian Stock
• CuBEs: Culturally-Situated Behavioral Evaluations and the Limitations of Culture-Blind LLM Judges
• Wilson to sign ordinance banning AI, data-driven price changes at grocery stores
• AM Intelligence To Invest Over $20 Billion In AI Data Centres