Nvidia faces calls for a Justice Department investigation after national security experts and AI policy leaders raised allegations of chip smuggling to China. An NBC News video published November 19, 2026, covered the proposed probe and its relevance to US-China competition in AI. The articles provide no details substantiating the allegations. Separately, Nvidia shares rose as investors anticipated a potential $1.5 trillion market debut for Anthropic, which could attract retail investors and help revive technology stocks. Applied Optoelectronics, an optical networking supplier, has moved in the opposite direction: its shares are down 80% from their May 13 record close, with further warning signs emerging Wednesday.
Microsoft CEO Satya Nadella said buyouts offered to 60,000 employees give them a choice as the company reorganizes around AI. He said the aim is not to replace software engineers but to change how they work, with people directing AI agents. Nadella cited the 11 developers building Cowork, an AI agent in Microsoft 365 apps, as an example of AI being used to create a product. He expects agent management to spread across roles, while acknowledging the transition will take time and may be difficult for some workers.
Two initiatives focus on making AI systems more useful and accountable. Researchers’ open-source Agent Trust Infrastructure framework verifies agents’ identity, integrity and permissions across organizations using public-key infrastructure, zero-trust architecture and remote attestation. In tests across eight organizations and six agent frameworks, it contained 84% of attacks with all safeguards enabled, versus 43% without them, adding about 174 microseconds per request. Guidepoint, meanwhile, launched Guidepoint Labs in New York on November 18, 2026, to provide expert human data for AI training and evaluation. Its network spans more than 2 million advisors in 300-plus industries and 150 countries; services include training data, expert-built tasks, agent trajectories, evaluation, red teaming, and multilingual and multimodal data. The company says 80% of its experts are willing to participate.
Perplexity introduced its Orchestrator API, a single endpoint for frontier and open-weight models from providers including OpenAI, Anthropic, Google and xAI. It can fail over to another provider when a deployment is unavailable, with requests billed at the responding model’s published token rate. Its preview `perplexity/auto` feature lets developers set routing choices. In company tests, routing around refusals increased completed tasks from 38 of 63 to 41 of 63 in two configurations. Separately, Aslan president Jesse Rome says AI can take on sales tasks that do not require a salesperson’s distinctive skills, leaving people to focus on relationships and customer needs. Aslan+ offers training, role-play, call preparation and feedback; a survey of 1,000 B2B buyers found 90% cited industry understanding as the top factor in choosing a seller.
Meta and the Richland Parish Chamber of Commerce brought hands-on AI training to small businesses in Richland Parish, Louisiana. Meta community engagement manager Sharon Gilmore said AI can help with tasks such as invoicing and answering email. Well Aesthetics and Wellness CEO Kasey Thompson said the workshop gave her new ideas for using AI and could free up more time with clients. At the policy level, President Donald Trump told the APEC Economic Leaders’ Meeting in Shenzhen that the United States opposes global rules controlling or restricting AI. He called the technology “almost completely good,” while recognizing risks that need careful management, and argued for domestic vigilance rather than multilateral limits.
AI demand is also reflected in Asia Pacific financing. Equity issuance outside Japan rose 121% in the third quarter of 2026, although the number of offerings declined 2.1% over nine months. High Technology generated $123.9 billion, or 41% of regional equity proceeds, amid AI-related funding demand. Major transactions included SK Hynix’s $26.5 billion NASDAQ ADR offering and Alibaba’s $10.2 billion Hong Kong placement. Regional debt issuance reached a record $4.16 trillion in the first nine months, while China raised $159.4 billion in equity proceeds.
Key Takeaways
- National security and AI policy leaders urged the Justice Department to investigate Nvidia over alleged chip smuggling to China; the articles provide no further evidence or details.
- Nvidia shares rose as investors anticipated Anthropic’s potential $1.5 trillion market debut; Applied Optoelectronics shares fell 80% from their May 13 record close.
- Microsoft offered buyouts to 60,000 employees as it reorganizes around AI; Satya Nadella says employees will increasingly direct AI agents rather than be replaced by them.
- Agent Trust Infrastructure contained 84% of attacks in tests with safeguards enabled, compared with 43% without, at about 174 microseconds of added latency per request.
- Guidepoint Labs launched with access to more than 2 million advisors across 300-plus industries and 150 countries, offering human data and expert services for AI training and evaluation.
- Perplexity’s Orchestrator API routes requests among models from OpenAI, Anthropic, Google and xAI, with failover and a preview routing feature; its tests showed more tasks completed when routing around refusals.
- Aslan+ applies AI to sales training, role-play, preparation and feedback; 90% of 1,000 surveyed B2B buyers said industry understanding is the top factor in choosing a seller.
- Meta and the Richland Parish Chamber are offering local businesses hands-on AI training, including potential uses for invoicing and email.
- At the APEC summit, Donald Trump opposed global rules restricting AI and called for domestic vigilance in managing risks.
- Asia Pacific equity issuance outside Japan rose 121% in Q3 2026; high technology accounted for $123.9 billion, or 41%, of regional equity proceeds.
Leaders call for probe into Nvidia over alleged China chip smuggling
National security experts and AI policy leaders are urging the Justice Department to investigate Nvidia over alleged chip smuggling to China. NBC News AI reporter Jared Perlo discusses the letter and why the proposed investigation matters in the AI race between the United States and China.
Leaders call for probe into Nvidia over alleged China chip smuggling
National security experts and AI policy leaders are urging the Justice Department to investigate Nvidia over alleged chip smuggling to China. NBC News AI reporter Jared Perlo discusses the letter and why the proposed investigation matters in the AI race between the United States and China.
Leaders call for probe into Nvidia over alleged China chip smuggling
A video published by NBC on November 19, 2026, is titled “National security and AI leaders urge Justice Department to investigate Nvidia.” The page also displays a poll asking which recent case most influenced views of Ghana’s judiciary, open from September 1 to November 30, 2026.
Applied Optoelectronics shares fall 80% from record high as warning signs emerge
Shares of optical networking supplier Applied Optoelectronics have fallen 80% since reaching a record closing high on May 13. The highly volatile stock showed further warning signs on Wednesday as the artificial intelligence trade failed to gain momentum. The article does not provide further details about the signals.
Nvidia shares rise as investors look to Anthropic IPO for AI market boost
Nvidia shares rose on Thursday as technology investors awaited Anthropic’s $1.5 trillion market debut. Anthropic could become a new AI favorite and draw retail investor interest away from Nvidia. The article says its IPO could help revive the technology trade.
Nadella says Microsoft buyouts will help workers adapt to AI
Microsoft CEO Satya Nadella said the company offered buyouts to 60,000 employees so they could choose whether to stay as Microsoft reorganizes around AI. He said the goal is not to replace software engineers, but to change how they work, with employees directing AI agents. Nadella pointed to the 11 developers who build Cowork, an AI agent in Microsoft 365 apps, saying they use AI to create the product. He expects employees in many roles to manage AI agents, but said the change will take time and may be difficult for some workers.
ATI offers a security framework for AI agents across organizations
Researchers proposed Agent Trust Infrastructure, or ATI, a security framework that helps verify the identity, integrity, and permissions of AI agents working across organizations. ATI uses public key infrastructure, zero trust architecture, and remote attestation, with separate organizational certificate authorities instead of one shared authority. In tests involving eight organizations and six agent frameworks, ATI contained 84% of attacks when all safeguards were enabled, compared with 43% without them. The system added about 174 microseconds of overhead per request and is available as an open-source package.
AI can support sales teams while keeping relationships personal
Jesse Rome, president of Aslan, says salespeople need to act as solution consultants and trusted partners as buyers gain information and bring more decision-makers to calls. Aslan’s survey of 1,000 B2B buyers found that 90% named understanding their industry as the top factor in choosing a seller. Rome described Aslan+, an AI-powered platform that supports training, role-play, call preparation, and feedback. He says AI can handle parts of the job that are not unique to a salesperson, while people focus on personal relationships and customer needs.
Guidepoint launches expert services for AI training and evaluation
Guidepoint announced Guidepoint Labs in New York on November 18, 2026, to provide expert human data for AI training, evaluation, and reinforcement learning environments. The unit draws on more than 2 million advisors across 300-plus industries and 150 countries, with 15,000-plus recruited each month. Its services include training data, expert-built tasks and grading guides, agent trajectories, evaluation and red teaming, and multilingual and multimodal data. Guidepoint Labs executive Grant Cho said 80% of experts are willing to take part in these AI training projects.
Trump opposes global rules for artificial intelligence at APEC summit
US President Donald Trump said the United States opposes global rules that control or restrict artificial intelligence. Speaking at the 33rd Asia-Pacific Economic Cooperation Economic Leaders’ Meeting in Shenzhen, China, he described the technology as “almost completely good” but said it has risks that need careful management. Trump called the United States the world leader in “superintelligence,” which he also referred to as “SI.” He argued for domestic vigilance rather than multilateral limits.
AI investment drives Asia Pacific’s $302 billion equity surge
Asia Pacific equity issuance outside Japan rose 121% in the third quarter of 2026, even though the number of offerings fell 2.1% over nine months. High Technology raised $123.9 billion, or 41% of regional equity proceeds, as AI-related investment boosted demand for funding. Major deals included SK Hynix’s $26.5 billion NASDAQ ADR offering and Alibaba’s $10.2 billion Hong Kong placement. Debt issuance also reached a record $4.16 trillion in the first nine months, while results varied across markets, with China raising $159.4 billion in equity proceeds.
Perplexity launches API for model routing and failover
Perplexity launched its Orchestrator API, which provides one endpoint for frontier and open-weight AI models from providers including OpenAI, Anthropic, Google and xAI. It can switch to another provider for the same model when a deployment is unavailable, and bills each request at the answering model’s published token rate. The `perplexity/auto` routing feature is in preview and lets developers set routing choices. In Perplexity’s tests, routing around refusals increased completed tasks from 38 out of 63 to 41 out of 63 in two configurations.
Meta brings hands-on AI training to Richland Parish businesses
Meta and the Richland Parish Chamber of Commerce brought the Community Accelerator program to small businesses in Richland Parish, Louisiana. The program offers hands-on training in AI tools that may help businesses work more efficiently and grow. Sharon Gilmore, Meta’s community engagement manager for its Hyperion Data Center, said AI can help with tasks such as invoicing and answering emails. Kasey Thompson, CEO of Well Aesthetics and Wellness, said the workshop showed her new ways to use AI and could help her spend more time with clients. Meta is encouraging local businesses to attend its workshops and learn about its platforms.
Sources
- National security and AI leaders urge Justice Department to investigate Nvidia
- National security and AI leaders urge Justice Department to investigate Nvidia
- National security and AI leaders urge Justice Department to investigate Nvidia
- This Wildly Volatile AI Stock Is Down 80% From Its Peak—and Flashing Warning Signals
- Nvidia Stock Is Rebounding on Hope Anthropic IPO Will Rescue the AI Trade
- Microsoft CEO Satya Nadella explains why he asked 60K employees to quit
- Agent Trust Infrastructure (ATI): A Federated Identity, Trust, Attestation, and Accountability Architecture for Cross-Organization Autonomous AI Agents
- How Sales Teams Can Use AI Without Losing the Human Side of Selling
- Guidepoint Unveils Guidepoint Labs to Power the Next Frontier of AI Training and Evaluation
- Trump rejects 'globalist regulation' of artificial intelligence at APEC summit
- AI investment fuels Asia Pacific’s $302bn equity surge
- Perplexity Launches Orchestrator API for Failover and Model Routing
- Meta brings AI training to Richland Parish small businesses
Comments
Please log in to post a comment.