Nvidia and ASML highlight two very different roles in the AI hardware supply chain. Nvidia reported $96.2 billion in revenue for Q2 2026, while ASML brought in $10.6 billion in the same period. Despite the gap, ASML remains the only supplier of EUV lithography equipment, making it essential to chip manufacturing worldwide. Nvidia keeps higher profit margins at 74 cents per dollar, compared with ASML's 49 cents.
Applied Materials and TSMC show another pair of contrasting revenue paths. TSMC earned $46.7 billion in Q3 2026, while Applied Materials reported $9.1 billion in Q2 2026. TSMC's revenue has grown steadily from $26.8 billion in Q4 2024, while Applied Materials has stayed flatter over the same period. TSMC also opened its first Arizona fabrication plant, achieving production yields that match its facilities in Taiwan.
AI is making its presence felt in politics in worrying ways. North Carolina's election has seen multiple campaigns use AI-generated images and attack ads, including fake fliers targeting a Supreme Court candidate and ads mocking religious groups. In Pennsylvania, a Republican gubernatorial candidate aired an ad featuring an AI-generated version of Democratic Gov. Josh Shapiro. Pennsylvania Senator Lisa Boscola plans to introduce legislation banning AI deepfakes targeting candidates, following Delaware's example.
Google is facing a criminal probe over its AI data center project in Finland. The company cleared about 500 acres of forest near Otanmäki, prompting an investigation backed by Finland's Licensing and Supervision Authority and local police. Google views Finland as a strong location because of its cool climate and government support for AI projects, but the project has caused local problems including wolf displacement and potential electricity bill increases.
On the fraud front, India's festive e-commerce season is expected to see 966 million online orders in 2026, up from 760 million in 2025. Fake orders and fraudulent transactions are rising alongside that growth, with 7.1% of attempted digital transactions flagged as suspected fraud in 2025, well above the global average of 3.8%. E-commerce companies are turning to AI and machine learning to analyze order history, device data, and delivery patterns to identify risky transactions.
MIT researchers and the MIT-IBM Watson AI Lab launched the LLM Election Observatory, a dashboard that monitors how popular AI chatbots answer questions about the 2026 midterm elections in real time. The tool aims to give voters more accurate and unbiased information by tracking how chatbot responses shift over time.
In other AI developments, LG Energy Solution unveiled Enswer, an AI analytics agent for battery energy storage system trading, at Smart Energy Plus 2026 in Seoul. The agent works within the company's AEROS Trading software, helping traders by searching market data, bidding history, and regulations. Meanwhile, a new study finds that AI financial assistants struggle with underspecified money questions, often answering a different question than intended while still producing plausible-looking numbers.
Key Takeaways
- Nvidia reported $96.2 billion in Q2 2026 revenue compared to ASML's $10.6 billion, though ASML remains the sole supplier of EUV lithography equipment.
- TSMC earned $46.7 billion in Q3 2026, significantly outpacing Applied Materials' $9.1 billion in Q2 2026, with TSMC's Arizona plant now matching Taiwan yields.
- North Carolina's election is the first since generative AI became widespread, with multiple campaigns using AI-generated images and attack ads.
- A Republican candidate in Pennsylvania aired an AI-generated ad of Gov. Josh Shapiro, prompting proposed legislation to ban AI deepfakes targeting candidates.
- Google faces a criminal investigation in Finland over clearing 500 acres of forest for an AI data center near Otanmäki.
- India's festive e-commerce orders are projected to grow from 760 million in 2025 to 966 million in 2026, with 7.1% of transactions flagged as suspected fraud.
- MIT and the MIT-IBM Watson AI Lab launched the LLM Election Observatory to monitor how AI chatbots answer election-related questions in real time.
- LG Energy Solution unveiled Enswer, an AI analytics agent for battery energy storage system trading, at Smart Energy Plus 2026 in Seoul.
- A new study identifies 'frame collapse' in AI financial assistants, where models answer the wrong question while still producing real numbers.
- Binance's $6.3 billion security breach highlights growing concerns that AI could accelerate cyberattacks on crypto platforms.
ASML and Nvidia revenue gap reveals different AI chip roles
ASML makes the machines that build the world's most advanced chips, while Nvidia sells the AI chips themselves. Nvidia brings in much more revenue, reaching $96.2 billion in Q2 2026 compared to ASML's $10.6 billion. ASML remains the only supplier of EUV lithography equipment, making it essential to the entire chip industry. Nvidia keeps higher profit margins at 74 cents per dollar versus ASML's 49 cents. Both companies play important but different roles in the AI supply chain.
Applied Materials and TSMC show different revenue growth paths
Applied Materials designs and manufactures wafer fabrication equipment for chip production. Taiwan Semiconductor Manufacturing Company operates as a global foundry that makes and tests integrated circuits for other tech firms. TSMC consistently earns more revenue, reaching $46.7 billion in Q3 2026 while Applied Materials reported $9.1 billion in Q2 2026. TSMC also opened its first Arizona fabrication plant that matches production yields in Taiwan. Applied Materials opened a new R&D center in California this year.
Applied Materials and TSMC revenue trends compared over eight quarters
Applied Materials and Taiwan Semiconductor Manufacturing both serve the AI chip industry but show different revenue patterns. TSMC revenue grew steadily from $26.8 billion in Q4 2024 to $46.7 billion in Q3 2026. Applied Materials revenue stayed flatter, ranging from $6.8 billion to $9.1 billion over the same period. TSMC announced an industry initiative for advanced photomasks and reached production yields in Arizona matching its Taiwan facilities. Investors will watch whether the revenue gap between the two companies continues to grow.
Five scary AI uses in North Carolina politics this election
North Carolina's election is the first since generative AI became widely available. Campaign materials used AI in alarming ways, including fake fliers showing Supreme Court candidate Sarah Stevens. U.S. Senate candidate Michael Whatley's campaign went silent after an AI-related incident. The North Carolina Senate Majority Fund used an AI-generated image of a Democrat in an ad. State Sen. Mike Lee's campaign ran an AI ad that mocked Quakers. Jennifer Balkcom's campaign turned a real Democrat quote into an AI-produced attack ad.
AI-generated ad depicting Gov. Shapiro airs in Pennsylvania despite criticism
Republican gubernatorial candidate Stacey Garrity paid for an ad featuring an AI-generated version of Democratic Gov. Josh Shapiro. The ad shows Shapiro admitting to fake scandals and discloses AI use about 40 seconds in. Viewers recognized the ad as artificial, with many expressing disgust. Federal rules prevent TV stations from rejecting ads paid for by legally qualified candidates. Pennsylvania Senator Lisa Boscola plans to introduce legislation banning AI deepfakes targeting candidates during elections. Delaware already has a similar law prohibiting deepfakes within 90 days of Election Day.
India's festive e-commerce boom brings surge in fake orders and fraud
India's festive online shopping orders are expected to grow from 760 million in 2025 to 966 million in 2026, a 27.1% increase. Fake orders, account misuse, and fraudulent refund claims rise during this season, making it harder for companies to spot fraud. In 2025, 7.1% of attempted digital transactions in India were suspected to be fraudulent, compared with a global average of 3.8%. E-commerce firms are now using AI and machine learning to analyze order history, device data, and delivery patterns to flag risky transactions. Logistics platform ShipMaxx360 said customer, payment, and delivery data is often spread across multiple platforms, adding to the challenge.
Beginner guide to building AI agents in 2026
A good AI agents course should start with a basic execution loop where the model receives a goal, uses tools, and decides what to do next. Learners should first build a simple command-line project with clear success conditions, limited tool arguments, and rules to stop endless loops. After that, students can practice the same workflow using frameworks like LangGraph, smolagents, or Microsoft Agent Framework. The course should also teach evaluation by creating test sets, expected outcomes, cost limits, and review logs. Comparing results before accepting new prompts or models helps ensure reliable performance.
LG Energy Solution reveals AI agent Enswer for battery trading
LG Energy Solution publicly unveiled Enswer, an AI analytics agent for managing BESS (Battery Energy Storage System) trading, at Smart Energy Plus (SEP) 2026 in Seoul. The agent works inside the company's ESS software AEROS Trading, which uses battery and market data for bidding, verification, and settlement. Enswer acts as an analytical assistant that searches for market data, bidding history, regulations, and news to provide evidence-based strategies. It can replace repetitive tasks performed by traders, according to Executive Vice President Kim Hyeon-tae.
Google faces criminal probe over AI data center deforestation in Finland
A criminal investigation has been launched into Google's AI data center in Finland after the company cleared about 500 acres of forest near Otanmäki. The investigation is backed by Finland's Licensing and Supervision Authority (LVV) and local police. Google sees Finland as a good location because of its cool climate, affordable power, and government support for AI projects. However, the project has caused local problems including wolf displacement, pollution, and potential 40% electricity bill increases. Finnish government representatives have not yet commented on the case.
Study reveals AI financial assistants fail on unclear money questions
A new study examines how AI financial assistants handle underspecified questions about money and investments. The research identifies a problem called frame collapse, where models answer a different question than the one intended while still producing real numbers. The study focuses on Indian mutual funds, where product variants and tax rules are formally defined. Models often used ambiguous figures under the wrong label or assumed outdated tax rules when investor facts were missing. The authors propose controls to help financial assistants avoid these failures.
Binance hack reveals AI could increase crypto attack risks
Binance, the world's largest crypto exchange, suffered a security breach affecting about $6.3 billion in assets, with roughly $3 billion frozen. The attack started from a compromised engineer's laptop, exposing weaknesses in how assets were stored rather than flaws in the blockchain itself. Binance used internet-connected hot wallets that prioritized user experience over security. Analysts warn that AI could make cyberattacks on crypto platforms more dangerous by helping attackers find vulnerabilities faster, though defenders are also using AI.
MIT launches AI voting information watchdog dashboard
MIT researchers and the MIT-IBM Watson AI Lab created a dashboard called the LLM Election Observatory. It monitors in real time how popular AI chatbots answer questions about the 2026 midterm elections. The tool is designed to help voters get accurate and unbiased information about candidates and issues. The dashboard updates continuously so users can see how chatbot responses change over time.
3 inverse ETFs doubled as AI stocks fell in October
Three leveraged inverse ETFs posted huge gains in October as AI infrastructure and quantum computing stocks declined. The Tradr 2X Short CBRS Daily ETF gained 126%, the Tradr 2X Short IREN Daily ETF gained 135%, and the T-REX 2X Inverse CRWV Daily Target ETF advanced 96%. These funds profit from drops in stocks like Cerebras Systems, IREN Ltd, and CoreWeave Inc. However, the daily-reset leveraged design makes them risky and better suited for short-term trading than long-term investing.
Sources
- ASML vs. Nvidia: Who Really Captures the AI Chip Boom?
- Applied Materials vs. Taiwan Semiconductor Manufacturing: Comparing Revenue Growth Trends Between These Artificial Intelligence Companies
- Applied Materials vs. Taiwan Semiconductor Manufacturing: Comparing Revenue Growth Trends Between These Artificial Intelligence Companies
- The 5 Scariest Uses of AI in N.C. Politics This Election
- Here's why an AI-generated campaign ad depicting Gov. Shapiro is allowed on the air in Pennsylvania
- As festive online shopping surges, how are e-commerce firms using AI to detect fake orders, prevent fraudulent refunds?
- AI Agents Course Guide for Beginners in 2026
- LG Energy Solution to manage BESS trading with AI agent 'Enswer,' unveiled publicly for first time at SEP
- Google hit with criminal probe over unapproved AI data center deforestation — 816 acres of forest destroyed in Finland as wolves flee into local villages, facility threatens 40% power bill hikes
- Right Numbers, Wrong Questions: Evaluating LLMs on Underspecified Financial Queries
- Binance hack shows AI could make crypto attacks 'more dangerous', analysts warn
- Q&A: Hey AI, who should I vote for this year?
- These 3 ETFs Doubled Investors' Money in October as AI Stocks Crashed
Comments
Please log in to post a comment.