A sharp selloff has put the scale and concentration of AI investing under scrutiny. The S&P 500 lost $5.6 trillion in market value in one week and ended more than 10% below its early-October record. Its 10 largest stocks now account for about 40% of the index, compared with 29% at the dot-com bubble peak. Bank of America’s monthly survey found fund managers’ net overweight in technology dropped to 12% from 37%—the biggest one-month reduction since June 2006. Some 52% expect an AI hyperscaler to cut capital spending within six months, while 38% identify hyperscaler bonds as the likeliest source of a credit event. Advisers recommend reviewing fund exposure, diversifying in line with risk tolerance and avoiding panic selling; CFRA Research’s example moderate-risk allocation is 45% U.S. stocks, 15% international stocks, 35% bonds and 5% cash.
Anthropic’s November 19, 2026, IPO raised $81 billion, described in one article as the largest public offering in history, but its shares closed at $88.91 on November 21, down 8.5% from Thursday and only modestly above the $83 IPO price as the tech selloff continued. The company reportedly has $230 billion in cash and plans to spend $900 billion on computing infrastructure by 2045. Nvidia, Microsoft and Google are among its backers or suppliers, tying their prospects to expectations for AI growth. The reporting flags risks from weaker profits, higher energy and equipment costs, and public opposition to data centers. Co-founder Dario Amodei says Anthropic may slow model development for safety, though the article questions how much room companies will have to do so as economic reliance on AI grows. A separate article says Tristan Thompson called his Anthropic investment a “home run” after the company’s $81.3 billion stock debut, but supplies no details about his investment or remarks.
Other companies are pitching into AI infrastructure and software demand. Fabrinet makes optical packaging and hardware for high-performance computing and networking, with 800G transceiver programs already in production. It is trying to turn existing capacity of 5.8 billion and planned expansion to 9.8 billion into steady sales while addressing shortages of lasers, memory and ASICs. Its major customers include Cisco, NVIDIA, Nokia and Amazon; analysts cited project $10.1 billion in revenue and $1.1 billion in earnings by 2029. Separately, Oracle says financing is in place for its infrastructure deal with OpenAI after November stock and bond sales, and that Nvidia rack handovers at Shackelford remain on track.
Chip companies offer different exposure to the sector: GlobalWafers is moving factories in Texas, Italy and Japan toward production; Microchip Technology expects data-center revenue to rise from about $591 million in 2025 to roughly $1 billion in 2026; and analysts forecast JCET Group earnings growth of about 28.1% a year. In AI data management, AvePoint shares have gained more than 20%. Third-quarter revenue rose 19% to $131.1 million, SaaS revenue climbed 26% to $105.5 million, and annual recurring revenue increased 27% to $495.1 million. AvePoint forecast 2026 revenue of $511.3 million to $513.3 million and expanded its strategic partnership with Microsoft on November 10.
Agentic AI is also creating opportunities for platforms serving small businesses. Zeta Global uses customer data to automate marketing and reports 197 large AI-focused customers, averaging $1.8 million in revenue per customer. Toast is developing Toast IQ tools to manage restaurant workflows such as scheduling and bookkeeping, alongside its payments, ordering, payroll and back-office services. Klaviyo offers AI-assisted customer communications and marketing for small online businesses. Cognition, meanwhile, is discussing a funding round of at least $2 billion at a possible $75 billion valuation, above its $48 billion September valuation; talks remain incomplete. Its Devin agents can plan, write, test and deploy code with limited human help. Reported run-rate revenue rose from $492 million in May to nearly $900 million in September, and Cognition says enterprise use grew more than tenfold between early 2026 and May. Customers include NVIDIA, GE Aerospace, Citi, Mercedes-Benz and Modal; the company says it can combine models from different providers.
One article addresses a separate market forecast: ChatGPT estimates XRP could reach $5 by December 31, 2029, from a starting price of 89 cents, assigning that base case a 55% probability. It gives a 30% chance to $1.50 and a 15% chance to $15. The forecast assumes circulation grows from 63.5 billion XRP to 74.6 billion by the end of 2029, with 300 million coins added monthly for 37 months. XRP was around 90 cents after declines of 22.4% in two weeks and 32.1% in 30 days; the November 20 close was its lowest in a year. ChatGPT projects a sustained move above $1 in the second quarter of 2027 and a break above the $3.65 all-time high in the second quarter of 2029, while assigning a 12% chance to Standard Chartered’s $19.60 target.
Key Takeaways
- The S&P 500 lost $5.6 trillion in a week, while its 10 largest stocks represent about 40% of the index; investors are being urged to review AI and tech concentration rather than panic sell.
- Bank of America found fund managers’ net technology overweight fell from 37% to 12%, its biggest monthly cut since June 2006; 52% expect an AI hyperscaler to reduce capital spending within six months.
- Anthropic raised $81 billion in its November 19 IPO, but shares closed at $88.91 on November 21, down 8.5% from Thursday and near the $83 IPO price.
- Anthropic reportedly holds $230 billion in cash and plans $900 billion in computing infrastructure spending by 2045; Nvidia, Microsoft and Google are among its backers or suppliers.
- Fabrinet has 800G transceiver programs in production and plans to expand capacity from 5.8 billion to 9.8 billion, while dealing with laser, memory and ASIC shortages; customers include NVIDIA and Amazon.
- Oracle says financing is in place for its OpenAI infrastructure deal after November stock and bond sales, with Nvidia rack handovers at Shackelford on track.
- Microchip Technology expects data-center revenue to grow from about $591 million in 2025 to roughly $1 billion in 2026; GlobalWafers is advancing factories in three countries, and JCET earnings are forecast to grow about 28.1% annually.
- AvePoint reported third-quarter revenue of $131.1 million, up 19%, and forecast 2026 revenue of $511.3 million to $513.3 million after expanding its partnership with Microsoft.
- Cognition is discussing at least $2 billion of funding at a possible $75 billion valuation; its Devin coding agents support customers including NVIDIA, GE Aerospace, Citi and Mercedes-Benz.
- Small-business AI examples include Zeta Global’s automated marketing, Toast IQ workflow tools and Klaviyo’s AI-assisted customer communications. ChatGPT’s XRP forecast places a 55% probability on $5 by the end of 2029, based on stated circulation assumptions.
Three AI chip stocks investors are watching after Nvidia's selloff
A sharp selloff in Nvidia and other AI stocks has raised doubts about the chip sector, but the article examines three companies that may offer different ways to invest in it. GlobalWafers makes silicon wafers and is moving new factories in Texas, Italy and Japan toward production. Microchip Technology makes embedded control chips and expects its data center revenue to grow from about US$591 million in 2025 to roughly US$1 billion in 2026. JCET Group provides chip assembly, packaging and testing, and the article says analysts forecast its earnings to grow about 28.1% per year.
AI infrastructure excitement lifts attention on Fabrinet
Fabrinet has drawn renewed investor attention as interest in AI infrastructure rises following Anthropic's IPO and Coherent's US$2 billion share repurchase plan. The company makes optical packaging and hardware for high performance computing and networking, including 800G transceiver programs already in production. Its near term challenge is turning its existing 5.8 billion capacity and planned expansion toward 9.8 billion into steady sales while managing shortages of lasers, memory and ASICs. Fabrinet also relies on major customers including Cisco, NVIDIA, Nokia and Amazon, so changes in their programs could affect revenue and earnings. Analysts cited in the article project US$10.1 billion in revenue and US$1.1 billion in earnings by 2029.
Three stocks linked to the agentic AI boom in small businesses
Agentic AI tools can handle tasks such as administration and marketing, creating new opportunities for platforms serving solo business owners and microbusinesses. Zeta Global Holdings uses customer data to automate marketing and has 197 large AI focused customers, whose average revenue per customer reached US$1.8 million. Toast provides restaurants with payments, ordering, payroll and back office tools, and its Toast IQ products are being developed to manage workflows such as scheduling and bookkeeping. Klaviyo provides AI assisted customer communication and marketing tools for small online businesses. The article presents these three companies as examples of stocks connected to the trend.
AI exposure in 401(k) plans raises concerns during market swings
The S&P 500 lost $5.6 trillion in market value during the week and finished more than 10% below its early October record, while AI stocks showed no clear recovery. Many retirement savers may own more technology and AI stocks than they realize because market weighted funds such as S&P 500 and Nasdaq funds hold large amounts of the biggest companies. The top 10 stocks make up about 40% of the S&P 500, compared with 29% at the dot-com bubble peak. Experts advise investors to review their funds, consider diversification based on their age and risk tolerance, and avoid panic selling. CFRA Research describes a moderate risk portfolio as 45% U.S. stocks, 15% international stocks, 35% bonds and 5% cash.
Bank of America survey finds sharp drop in fund managers' tech bets
Bank of America's latest monthly survey found that fund managers' net overweight position in technology fell to 12% this month from 37% the month before. It was the largest one month cut in tech exposure since June 2006. The survey also found that 52% of managers expect an AI hyperscaler to cut capital spending in the next six months, while 38% see AI hyperscaler bonds as the likeliest source of a credit event. The article notes that AI investment has supported company earnings and economic growth, while AI related companies made up nearly 49% of the S&P 500's market value by mid October. It advises retirement savers to review their investments and understand their exposure to technology and AI.
Anthropic’s huge IPO raises questions about AI’s risks
Anthropic raised $81 billion in what the article calls the largest initial public offering in history, on November 19, 2026. The company has $230 billion in cash and plans to spend $900 billion on computing infrastructure by 2045, making strong revenue growth important to its plans. Its backers and suppliers include Nvidia, Microsoft and Google, linking many industries to the same expectations for AI growth. The article warns that weaker profits, rising energy and equipment costs, or public resistance to data centers could harm investment and markets. Anthropic co-founder Dario Amodei says the company may slow model development for safety, but the article questions how much freedom companies will have to do so as the economy depends more on AI.
Tristan Thompson calls his Anthropic AI investment a home run
The article title says Tristan Thompson revealed an investment in Anthropic AI after the company’s $81.3 billion stock debut. The supplied article content provides no further details about the investment or Thompson’s comments.
Cognition discusses raising $2 billion or more at a $75 billion value
Cognition is discussing a funding round of at least $2 billion at a possible $75 billion valuation, according to a November 20, 2026, Bloomberg News report cited by people familiar with the matter. The reported value is above the company’s $48 billion valuation in September, and the talks have not resulted in a completed investment. Cognition says its Devin software agents can plan, write, test and deploy code with limited human help. Its reported run-rate revenue rose from $492 million in May to almost $900 million by September, and the company says enterprise use grew more than tenfold between early 2026 and May. Customers named by Cognition include NVIDIA, GE Aerospace, Citi, Mercedes-Benz and Modal; the company says its independent approach lets it combine different models rather than rely on one provider.
Anthropic shares fall after debut as AI stocks face a selloff
Anthropic’s highly anticipated stock market debut failed to lift a technology sector facing a broader selloff, the article reported on November 21, 2026. Investors had viewed the company, the first large machine-learning model maker to go public, as a sign of the health of the AI trade. Its shares closed at $88.91, down 8.5% from Thursday’s close and not far above the $83 IPO price. Declines among other major technology stocks showed that the listing did little to ease the sector’s gloom.
Oracle says funding is in place for its OpenAI infrastructure deal
Oracle says its infrastructure deal with OpenAI is fully funded, according to the article’s headline and description. The description says vendor financing is in place following Oracle’s November stock and bond sales. It also says Nvidia rack handovers at Shackelford remain on track. The supplied article content does not provide further details about the deal.
AvePoint gains as companies seek safer AI data management
AvePoint, Inc. (NASDAQ:AVPT) shares have risen more than 20% as companies look for ways to secure and manage data used by artificial intelligence systems. The company helps organizations protect and recover cloud data, control access to sensitive information, and oversee AI agents. Third-quarter revenue rose 19% to $131.1 million, while software as a service revenue increased 26% to $105.5 million. Annual recurring revenue grew 27% year over year to $495.1 million for a second straight quarter of faster growth. AvePoint expects 2026 revenue of $511.3 million to $513.3 million and announced an expanded strategic partnership with Microsoft on November 10, 2026.
ChatGPT sees XRP reaching $5 by the end of 2029
ChatGPT forecasts that XRP could reach $5 by December 31, 2029, about 460% above the forecast’s 89-cent starting price. XRP was trading around 90 cents after falling 22.4% in two weeks and 32.1% in 30 days, and its November 20 close was its lowest in a year. The forecast gives a 55% chance to a $5 base case, a 30% chance to a $1.50 low case, and a 15% chance to a $15 high case. The estimates assume 74.6 billion XRP in circulation by the end of 2029, compared with 63.5 billion now, as 300 million new coins are added monthly for 37 months. ChatGPT projects XRP could rise above $1 consistently in the second quarter of 2027 and pass its $3.65 all-time high in the second quarter of 2029, while giving a 12% chance of reaching Standard Chartered’s $19.60 target for 2029.
Sources
- 3 AI Chip Stocks Retail Investors Are Watching After The Nvidia Selloff
- Fabrinet Stock Gets An AI Infrastructure Boost
- 3 Stocks Riding The Agentic AI Boom For Solopreneur Platforms
- Your 401(k) is full of AI. How to deal with the volatility
- ‘End of boom & AI euphoria’: Bank of America says money managers cut tech bets by the most since 2006. Are you ready?
- Anthropic: The ever more dizzying gamble of AI
- Tristan Thompson Reveals His “Home Run” Investment in Anthropic AI After Its $81.3 Billion Stock Debut
- Potential Cognition AI Round Would Raise $2B-Plus at $75B Valuation
- Anthropic’s post-debut stumble delivers reality check for AI trade
- Oracle Says OpenAI Infrastructure Deal Is Fully Funded
- Is AvePoint, Inc. (NASDAQ:AVPT) an Emerging AI Infrastructure Play Under $25?
- ChatGPT AI Predicts XRP's Price for the End of 2029
Comments
Please log in to post a comment.