US tech stocks took a beating this week as investors grew nervous about the massive scale of AI spending. The Nasdaq tumbled 2.6% on Monday, with the tech sector down 3.4%, while the S&P 500 fell 1.4% and the Dow dropped 1.1%. The selloff was fueled by fears that AI could hurt the broader economy, combined with Middle East tensions that pushed oil up 4.5%. The volatility index jumped 6.31%, signaling real anxiety in the market.
Meta took the hardest hit among mega-caps after disclosing a staggering $825 billion in future AI infrastructure commitments in a SEC filing. That figure includes $355.7 billion in leases, $454.5 billion in purchase commitments, and $14.7 billion in cloud obligations. Shares fell over 7%, and despite record quarterly revenue of $64.6 billion, Meta's free cash flow turned negative at $2.1 billion. Accenture dropped 3%, while EPAM, AeroVironment, SoundHound AI, and Meta itself also slid in the October 23 session.
Friday's close pushed the S&P 500 and Nasdaq to their lowest levels in five weeks. Telecom stocks like AT&T, Verizon, and T-Mobile fell between 3% and 4.7% on fears of competition from SpaceX's telecom ambitions. Meanwhile, Crown Castle surged 12.6% and American Tower gained 6.9%. The selloff deepened after Meta's near-8% drop on Thursday over AI spending concerns. Earnings season looms with Microsoft, Alphabet, and Amazon reporting next week.
The bond market is flashing warning signs about AI infrastructure spending. Long-duration bonds for companies including Nvidia, Oracle, and CoreWeave have fallen sharply, with Nvidia's 2060 notes dropping below $0.63 on the dollar. Short sellers profited roughly $6 billion in paper gains from AI data center stocks in October, with Nebius shorts gaining $2.94 billion and CoreWeave shorts near $1.7 billion. CoreWeave now faces a $76 billion funding gap, needing about $85 billion to meet contracts while holding only $9 billion in coverage.
On the product and adoption front, Palo Alto Networks launched Cortex XDR One, a platform combining endpoint detection with protection for autonomous AI agents, targeting risks like prompt injection and data poisoning. In education, Kenyan startup StudyLabAI has signed up over 90,000 learners across 100 countries, turning student materials into interactive study tools with a 24/7 AI tutor. Meanwhile, a Citi Wealth report found that 51% of family offices now pick AI as their top direct investment target, with 89% reporting positive returns in 2026.
AI is also creeping into investment workflows, though humans still call the shots. Fixed-income investors use AI to summarize research and monitor developments, but adoption is uneven and decisions remain firmly in human hands. For CIOs, open-weight models from Meta and Google like Gemma are narrowing the performance gap with frontier models at lower cost. On the hardware side, Apple's M6 Mini, AMD's Strix Halo, and Nvidia's DGX Spark each offer different trade-offs for local AI workloads, with memory capacity emerging as the key factor.
Key Takeaways
- The Nasdaq fell 2.6% and the tech sector dropped 3.4% as investors feared AI's economic impact and Middle East tensions pushed oil up 4.5%.
- Meta disclosed $825 billion in future AI infrastructure commitments, sending shares down over 7% despite record quarterly revenue of $64.6 billion.
- Short sellers earned roughly $6 billion in paper profits from AI data center stocks in October, with CoreWeave facing a $76 billion funding gap.
- Nvidia's 2060 bonds fell below $0.63 on the dollar, and the bond market is raising alarms about AI infrastructure debt obligations.
- Palo Alto Networks launched Cortex XDR One, a platform protecting autonomous AI agents from prompt injection, data poisoning, and unauthorized API calls.
- Kenyan startup StudyLabAI signed up over 90,000 learners in 100+ countries without outside funding, offering AI tutors built from student materials.
- Citi Wealth found 51% of family offices now target AI as their top direct investment, with 89% reporting positive returns in 2026.
- Fixed-income investors are adopting AI for routine tasks like research summaries, but human decision-making remains firmly in place.
- Open-weight models from Meta and Google are narrowing the performance gap with frontier models at lower infrastructure cost.
- McLaren named Airwallex its official AI finance partner; Airwallex raised $320 million at an $11 billion valuation.
Nasdaq Falls 2.6% as Investors Worry About AI Impact
The Nasdaq index dropped 2.6% on Monday because investors feared AI could hurt the economy. Tech stocks fell the most, with the sector down 3.4%. The S&P 500 fell 1.4% and the Dow Jones dropped 1.1%. Oil prices rose 4.5% and gold gained 1.2%. Concerns about AI and Middle East tensions drove the declines.
US AI Stocks Fall at Open; Accenture Drops Over 3%
AI-related stocks fell across the US market on October 23. The Dow Jones dropped 1%, the S&P 500 lost 0.47%, and the Nasdaq slipped 0.36%. Accenture fell 3.03%, EPAM declined 1.9%, and AeroVironment dropped 1.88%. SoundHound AI fell 1.73% and Meta slipped 1.6%. The Volatility Index rose 6.31%.
Wall Street Dips After AI Selloff and SpaceX Telecom Fears
The S&P 500 and Nasdaq hit their lowest levels in five weeks on Friday. Telecom stocks like AT&T, Verizon, and T-Mobile fell between 3% and 4.7% due to fears of competition from SpaceX. Cell tower operators Crown Castle rose 12.6% and American Tower gained 6.9%. The Nasdaq fell nearly 3% on Thursday after Meta dropped almost 8% over AI spending concerns. Earnings season remains a focus with Microsoft, Alphabet, and Amazon reporting next week.
Meta's $825 Billion AI Commitments Rattle Markets
Meta disclosed $825 billion in future AI infrastructure commitments in a SEC filing. This includes $355.7 billion in leases, $454.5 billion in purchase commitments, and $14.7 billion in cloud obligations. Meta shares fell over 7% after the news. Quarterly revenue hit a record $64.6 billion, but free cash flow turned negative at $2.1 billion. Meta plans to raise funds by creating a new non-voting Class C stock starting January 19, 2027.
Short Sellers Profit $6 Billion on AI Data Center Stocks
Short sellers earned about $6 billion in paper profits from AI data center stocks in October. CoreWeave shorts made close to $1.7 billion while Nebius shorts gained $2.94 billion. CoreWeave fell 36% and Nebius fell 33% from late September to late October. CoreWeave needs about $85 billion to meet contracts but only has $9 billion covered, leaving a $76 billion funding gap. Its unsecured bonds yielded around 15%, signaling high default risk.
AI slowly enters investment workflows but humans still decide
Fixed-income investors are starting to use AI to summarize research, review documents, and monitor market developments. Adoption is uneven, with some firms using AI daily while others still test it or restrict it due to data security concerns. AI mainly speeds up routine tasks rather than replacing analysis. Investment decisions remain firmly in human hands, with analysts still checking AI outputs against established sources.
Palo Alto Networks launches Cortex XDR One for AI agents
Palo Alto Networks launched Cortex XDR One, a platform combining endpoint detection and response with protection for autonomous AI agents. The product targets risks like prompt injection, data poisoning, and unauthorized API calls that come from AI agents accessing internal systems. It uses small language models on the sensor to inspect threats even when devices are offline. The technology came from Palo Alto Networks' acquisition of Koi, a startup focused on agentic security.
Kenyan startup StudyLabAI offers AI tutors built on student material
Kenyan ed-tech startup StudyLabAI built a platform that turns students' own slides, lecture recordings, and notes into interactive study tools. It has signed up over 90,000 learners across more than 100 countries without raising outside funding. The platform offers narrated video lessons, adaptive quizzes, flashcards, and a 24/7 AI tutor grounded in each student's course material. A teachers' edition launched in January 2026 lets educators create video lessons and track student progress.
CIOs should experiment with open-weight AI models like Gemma
Open-weight AI models from Meta and Google are challenging enterprises' reliance on a small set of closed model providers. A CIO's experiments showed that the performance gap between open-weight and frontier models is narrowing faster than expected. These models can handle tasks like standup reports and log monitoring at lower cost without heavy infrastructure needs. Enterprise agreements and standards like ISO 42001 help address compliance concerns when using either open or closed models.
Bond market warns about AI infrastructure spending
The bond market is sending warning signals about the massive capital being spent on AI data center build-outs. Long-duration bonds for companies like SpaceX, Nvidia, Oracle, and CoreWeave have fallen sharply, with Nvidia's 2060 notes dropping below $0.63 on the dollar. Rising interest rate risks and historical patterns of tech bubbles are fueling skepticism about whether AI stocks can meet their debt obligations. Bond traders question whether Wall Street is paying attention to these signs.
AI Becomes Top Direct Investment Target for Family Offices
Citi Wealth's 2026 Global Family Office Report, based on 351 executives across 41 countries, found that 51% of family offices choose AI as their top direct investment destination. Nearly 89% of family offices reported positive portfolio returns in 2026, with 46% increasing their equity exposure. Inflation is now the top macroeconomic concern at 63%, up from 37% in 2025. Direct investments most commonly range from $1 million to $5 million, with healthcare and real estate also in high demand.
Best Local AI Hardware Options for Beginners in 2026
AI Master reviewed popular local AI devices including Apple's M6 Mini, AMD's Strix Halo, and Nvidia's DGX Spark. Apple's M6 Mini offers efficiency on a 2nm process but is limited by 16GB of memory. AMD's Strix Halo provides 128GB of memory at a competitive price but lacks Nvidia's CUDA support. Nvidia's DGX Spark excels at initializing AI models but is expensive with slower decode speeds. Memory capacity and bandwidth are the most important factors when choosing hardware for local AI workloads.
Airwallex Expands McLaren Partnership with AI Finance Focus
Airwallex has been named McLaren's official AI finance partner, expanding on an existing technology relationship. The deal includes Airwallex branding on the team kit sleeves, cars, and driver overalls. Airwallex recently raised $320 million at an $11 billion valuation to develop AI-native financial tools including T:0 and Airi. The partnership aims to connect Airwallex's cross-border payment services with its future AI-driven finance products in a real-world operating environment.
Sources
- Nasdaq Drops 2.6% on AI Investment Concerns
- US AI Stocks Fall Early; Accenture Drops 3.03%
- Wall Street dips after AI-led selloff; SpaceX worries hit telecoms
- Meta's $825 Billion AI Investment Commitments Spark Concerns as Massive Debt Rattles Markets
- Short Sellers Just Made $6 Billion Betting Against AI Data Center Stocks
- AI Enters the Investment Workflow
- Palo Alto Networks launches Cortex XDR One for AI agents
- Kenyan startup StudyLabAI gives students an AI tutor built on their own course material
- Open-weight AI models are changing the equation. Here’s what CIOs should experiment with next
- The Bond Market Is Sending a Dire Warning to Wall Street About the AI Infrastructure Build-Out, but Is Anyone Paying Attention?
- AI Top Direct Investment for Family Offices
- Local AI for Beginners in 2026: What Hardware Should You Buy?
- Airwallex turns McLaren deal into an AI finance story
Comments
Please log in to post a comment.