AI Stocks Lift US Markets as Anthropic Targets $1.4 Trillion IPO Value

AI stocks helped lift U.S. markets on Wednesday, with the S&P 500 up roughly 0.9% to 1.3% across the reports and on pace for its third gain in four days. The articles describe the same market move but give differing estimates of the index’s rise.

Advertising technology executive Jean-Charles Brunn, CEO of Criteo, expects AI platforms eventually to adopt advertising. He said investors are questioning whether AI will deliver on its promises, much as they doubted earlier technologies, and reported strong recent growth in demand for Criteo’s online advertising services. Amazon, Google and Facebook are among the brands using them.

Mercor raised $750 million in a Series D round, according to an article dated November 10, 2026. CEO Foody says the company is expanding into enterprise services to evaluate AI agents, assess their costs and improve performance. Mercor says it spends three times as much on AI inference as on employee salaries and organizes human expertise to help companies measure and train AI systems.

Research into three open Mixture-of-Experts model families found that neuron sets associated with the same domain overlap by 58% to 74%, versus 15% to 24% across different domains. The researchers’ RAMP-based Pruning method retained 91% to 98% of original performance at 75% neuron retention across six models. At 50% retention, it cut resident GPU memory by 36% to 43% and increased batched prefill throughput by 17% to 39%.

SpaceX is developing a custom Nvidia AI rack, named Kessel, for terrestrial data centers and possible orbital computing, according to SemiAnalysis. The plan is to buy chip packages from Nvidia while sourcing other parts independently, bypassing Nvidia’s L6 compute boards. Kessel would retain 72 Oberon GPU packages, 18 compute trays and nine switch trays, but replace the standard rack’s copper cable cartridges with a central passive circuit-board midplane. Its 84-layer M10 board compares with 156 layers of M9 material planned for Nvidia’s earlier Kyber design.

Yale clerical workers represented by UNITE HERE Local 34 plan to seek workplace AI protections in contract negotiations; their current agreement expires in January. The union says Yale’s Clarity chatbot, introduced in fall 2024, can make errors employees must check, and alleges the University failed to provide 90 days’ notice before its introduction. Yale says it has not replaced jobs with AI and welcomes discussion of workplace technology. Connecticut Lt. Gov. Susan Bysiewicz warned AI could affect women workers and good middle-class jobs in New Haven.

Baird cut Nvidia from outperform to neutral and lowered its price target from $500 to $200, which the analyst said still implied an 18% gain from Tuesday’s close. Tristan Gerra cited uncertainty about CEO Jensen Huang’s future amid a board investigation and the possibility that major cloud companies will slow AI spending. He said a possible H200 export ban to China could prompt a $5 billion write-down, expects tighter controls tied to Southeast Asian data-center loopholes, and raised possible Department of Justice action involving Nvidia partner Megaspeed. He estimates hyperscaler capital spending could grow 40% year over year in 2027.

An AI app called CrimeRadar falsely reported a possible shooting at a Fleet Farm in Grand Chute, Wisconsin, on November 4; police said the radio call concerned an abandoned vehicle. A police official said the app likely misread dispatch audio, though the cause was unclear. The alert prompted calls from the public and the business. Police clarified that there had been no shooting, said the department is not affiliated with such apps, and urged people to check official sources before sharing reports.

Anthropic began its IPO road show on November 10, 2026, with a proposed price range that could value the company at up to $1.4 trillion. At the midpoint, it expects to raise more than $60 billion, and plans an additional $26 billion share sale to large investors including Google, AMD, Microsoft, Broadcom and Nvidia. Its filing warns that advanced AI could cause serious harm if safeguards fail, including risks from misaligned models and groups of autonomous agents. The offering will test investor demand amid questions about whether heavy AI spending can generate profits.

Notion said a breach at AI provider Fireworks AI exposed chat transcripts from a small number of workspaces, drawn from the Insights dataset used by administrators to review team AI use. Notion learned of the exposed dataset on November 2, cut off traffic, deleted its data from Fireworks’ systems and stopped training runs by November 3. Fireworks confirmed customer data had been accessed on November 8, after which Notion disabled Insights. The incident followed an October Fireworks breach that exposed API keys and tokens for a limited number of accounts.

Key Takeaways

  • The S&P 500 rose on Wednesday as AI stocks rebounded, though reports put the gain between 0.9% and 1.3%.
  • Criteo CEO Jean-Charles Brunn expects AI platforms to adopt advertising over time; Amazon, Google and Facebook use Criteo’s advertising services.
  • Mercor raised $750 million in Series D funding and is offering enterprise evaluation and improvement services for AI agents.
  • RAMP-based Pruning retained 91%–98% of model performance at 75% neuron retention; at 50% retention it cut GPU memory use by 36%–43% and boosted batched prefill throughput by 17%–39%.
  • SpaceX is developing its Kessel Nvidia AI rack with 72 Oberon GPU packages and a custom circuit-board midplane for Earth-based data centers and potential orbital computing.
  • Yale’s clerical workers union plans to seek workplace AI protections, raising concerns about the Clarity chatbot; Yale says it has not cut jobs in favor of AI.
  • Baird downgraded Nvidia to neutral and cut its target to $200, citing leadership uncertainty, possible slower AI spending and export-control risks.
  • CrimeRadar mistakenly reported a shooting in Grand Chute after apparently misreading police radio audio; police said the incident involved an abandoned vehicle.
  • Anthropic’s planned IPO could value it at up to $1.4 trillion; the company expects to raise more than $60 billion at the midpoint and plans a $26 billion share sale to investors including Google, AMD and Nvidia.
  • A Fireworks AI breach exposed some Notion AI chat transcripts; Notion disabled its Insights feature after removing its data from the vendor’s systems.

AI stock rebound pushes the S&P 500 higher

A rebound in AI industry stocks helped lift the U.S. market on Wednesday. The S&P 500 rose 1.3% and was on track for its third gain in four days.

AI stock rebound lifts the U.S. market

A rebound in AI industry stocks helped lift the U.S. market on Wednesday. The S&P 500 rose 1% and was on track for its third gain in four days.

AI stocks help Wall Street continue its gains

A rebound in AI industry stocks helped lift the U.S. market on Wednesday. The S&P 500 rose 0.9% and was on track for its third gain in four days.

Criteo CEO expects AI platforms to turn to advertising

Criteo CEO Jean-Charles Brunn told AFP on Tuesday that many AI platforms will eventually adopt advertising. He said investors are asking whether AI can deliver on its promises, while past technologies also faced early doubts before people saw their value. Brunn said demand for Criteo’s online advertising services has grown significantly in recent years. He named Amazon, Google, and Facebook among the major brands that use the company’s services.

Mercor raises $750 million for its AI training platform

Mercor raised $750 million in a Series D funding round, according to the article dated November 10, 2026. CEO Foody said the company is bringing its capabilities to enterprises to evaluate AI agents, measure their costs, and improve their performance. He said Mercor spends three times as much on AI inference as it does on employee salaries. Foody said Mercor organizes human expertise to help enterprises measure, train, and improve AI systems as they spread across the economy.

Neuron-level analysis reveals how MoE models specialize

Researchers used Routing-Aware Magnitude Profiling (RAMP) to study how neurons specialize in three open Mixture-of-Experts model families. Neuron sets from the same domain had 58 to 74 percent overlap, compared with 15 to 24 percent across different domains. Neuron importance changed between pre-training checkpoints, even as expert rankings stayed more stable, while post-training largely preserved neuron rankings. The researchers also introduced RAMP-based Pruning, or RAMPr, which retained 91 to 98 percent of original performance at 75 percent neuron retention across six models. At 50 percent retention, it reduced resident GPU memory by 36 to 43 percent and raised batched prefill throughput by 17 to 39 percent.

SpaceX plans custom Nvidia AI rack for Earth and orbit

SpaceX is developing a custom Nvidia AI rack called Kessel for data centers on Earth and computing in orbit, according to SemiAnalysis. SpaceX plans to buy chip packages from Nvidia and source other parts itself, bypassing Nvidia’s L6 compute boards. Kessel would replace the copper cable cartridges in Nvidia’s standard Vera Rubin NVL72 rack with a central passive printed circuit board midplane. It would keep 72 Oberon GPU packages, 18 compute trays and nine switch trays, arranged back-to-back around the board. The design uses an 84-layer M10 circuit board, compared with 156 layers of M9 material planned for Nvidia’s earlier Kyber design, and could be easier to manufacture.

Yale clerical workers seek AI protections in next contract

UNITE HERE Local 34, Yale’s clerical workers’ union, plans to make limits on workplace AI a priority in contract talks with the University. Its current contract expires in January and does not specifically protect union work from AI, according to steward Jen Ongley. The union’s October report criticized pressure to use AI for clerical tasks and cited Clarity, a chatbot introduced at Yale in fall 2024 for students, faculty and staff. Ongley said Clarity can produce errors that employees must check, and she repeated the union’s claim that Yale failed to give the required 90-day notice before introducing it. Yale spokesperson Tina Posterli said the University has not cut jobs in favor of AI and welcomed talks about workplace technology; Lt. Gov. Susan Bysiewicz also warned that AI could affect women workers and good middle-class jobs in New Haven.

Baird cuts Nvidia rating amid AI spending and leadership concerns

Baird lowered Nvidia’s rating from outperform to neutral on Wednesday and cut its price target from $500 to $200, which still suggests an 18 percent gain from Tuesday’s close. Analyst Tristan Gerra cited uncertainty about CEO Jensen Huang’s future during a board investigation and the risk that AI spending by major cloud companies could slow. He predicted a possible ban on H200 chip exports to China could lead to a $5 billion write-down, and he expects tighter export controls related to Southeast Asia data center loopholes. Gerra also raised the possibility of Department of Justice action against Nvidia partner Megaspeed. He estimated AI hyperscaler capital spending could rise only 40 percent year over year in 2027 and said Nvidia shares may stay in a trading range while concerns remain.

AI scanner app falsely flags abandoned car as shooting

An AI-powered app called CrimeRadar reported a possible shooting at Fleet Farm in Grand Chute, Wisconsin, on Nov. 4, but police said the call was about an abandoned vehicle. Cpl. Dylan Davis said the app likely misread radio dispatch audio, though police do not know exactly how the error happened. He said such apps use AI to turn public safety radio communications into text and may build reports from keywords even when they misunderstand what was said. The false alert led to calls from the public and the business, and police said they clarified that no shooting had occurred. The department, which is not affiliated with the apps, has created policies on AI and urged people to check official channels before sharing reports.

Anthropic targets up to $1.4 trillion valuation in planned IPO

Anthropic set an IPO price range that could value the company at up to $1.4 trillion. At the midpoint, it expects to raise more than $60 billion, and it plans to sell another $26 billion in shares to large investors including Google, AMD, Microsoft, Broadcom and Nvidia. The San Francisco company, led by CEO Dario Amodei and President Daniela Amodei, began its road show on November 10, 2026. Its filing warned that advanced AI could cause serious harm if safeguards fail, including risks from misaligned models and groups of autonomous agents. The planned share sale will test investor interest amid concerns about whether heavy AI spending will produce profits.

Fireworks breach exposed some Notion AI chat transcripts

Notion said a security breach at its AI provider, Fireworks AI, exposed AI chat transcripts from a small number of workspaces. The transcripts came from the Insights dataset, which lets workspace administrators see how teams use AI. Notion learned on November 2 that Fireworks had reported an exposed dataset, then cut off traffic, deleted its data from the vendor’s systems and stopped training runs by November 3. Fireworks confirmed on November 8 that customer data had been accessed, and Notion disabled the Insights feature. The incident followed an earlier Fireworks breach in October that exposed API keys and tokens for a limited number of accounts.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI stocks U.S. stock market Criteo AI advertising Mercor AI agents AI model pruning Mixture-of-Experts GPU optimization SpaceX Nvidia AI data centers workplace AI AI labor protections AI export controls Anthropic IPO AI safety AI app misinformation cybersecurity Notion

Comments

Loading...