Nvidia Reports $96.2 Billion in Second Quarter Revenue

Nvidia is at the center of two connected questions: who can access its AI chips and whether demand for them will keep growing. The company has told data-center customers it plans to stop supplying operators that rent more than half their GPUs to Chinese firms; the Financial Times reported that future contracts may require at least 51% of services to go to non-Chinese customers. Nvidia says the proposed Remote Access Security Act (RASA) could expose both the company and buyers to legal action by extending US controls to cloud access. The House passed the bill in January, but it still needs Senate approval. Separately, the Commerce Department has announced interim rules aimed at blocking Chinese companies from remotely accessing advanced AI chips. Reports say Moonshot AI sought Nvidia chips through data centers in Europe, and the Trump administration accused it of accessing chips through East Asian facilities. Another article says Alibaba and ByteDance prefer training large language models in Southeast Asia.

Investors are also watching Nvidia’s results for clues about AI infrastructure demand and energy needs. The company reported $96.2 billion in second-quarter revenue, up 106% year over year, including $89 billion from data centers, and forecast $108 billion for the third quarter. Its fiscal 2027 third-quarter results were due after the market close on Nov. 17, with interim CEO Colette Kress leading the call; another report said its fourth-quarter outlook was due Nov. 18. Investors are focused on customer spending and the next-generation Vera Rubin platform. Strong data-center growth could benefit Constellation Energy, which has 20-year power agreements with Amazon and Alphabet, Bloom Energy, which reported $1.24 billion in third-quarter revenue and makes on-site solid oxide fuel cells, and Energy Transfer, which transports natural gas to power plants.

Companies are also building controls around workplace AI. Commvault and Zscaler formed a partnership to identify AI tools, enforce usage rules, detect risky activity and help recover data, applications and identities after incidents. Coalition added OpenAI, Anthropic, Gemini and Cursor Enterprise integrations to its Wirespeed security platform. The connections bring administrative and access activity into Wirespeed, which can flag risky settings, possible data exposure and account-takeover signals, and combine those alerts with information from more than 100 security and IT tools.

In education, Instructure launched Instructure Foundry, a venture for AI products, and is developing Project Athena, a study coach that works with Canvas. Using course materials, assignments and upcoming assessments, Athena can create quizzes, study guides and coaching prompts. Instructure says student conversations will not train external AI models; the tool connects through a read-only link under existing student permissions. Hinds Community College joined after an initial pilot, with access expected to extend to its 7,000 students. Google, meanwhile, acquired Howie, an AI assistant for small-business email and calendar tasks such as scheduling meetings, finding times in group conversations and sending follow-ups. Howie’s team will continue supporting existing customers as it works with Google Labs on scheduling and calendar products.

AI investment and adoption remain uneven. Ramp data shows the top 1% of US businesses by AI spending paid a record $10,199 per employee in October 2026, while overall AI adoption reached 56.8%. The article also says Anthropic’s valuation could reach $2 trillion by December, citing a 27.5% YES price. Separately, investor Michael Burry remains bearish on Nvidia, Palantir and other AI stocks because of valuations and downturn risk, while adding to positions in Adobe, Microsoft, Alphabet, Salesforce and Palo Alto Networks. He described software’s long-term prospects as promising, citing fundamentals and growth potential.

At the TomorrowX conference in Austin, Coinbase CEO Brian Armstrong predicted AI agents could handle most transactions within a decade and estimated they might facilitate $3 trillion to $5 trillion in transaction volume by 2030. He argued that crypto, including stablecoins and Coinbase’s Base network, could support agent payments. Armstrong also said prediction markets and derivatives accounted for 25% of Coinbase’s trading-fee revenue in the third quarter, prediction markets had reached a $200 million annualized run rate, and shares traded on Base had exceeded $5 billion in volume. He urged public companies to consider tokenizing equity.

South Africa’s Macrocomm is developing utility-management systems that use data processing and AI to improve data accuracy, connect datasets, flag unusual activity and support decisions by municipalities. The article argues for local control of technology used in essential services while maintaining global partnerships. In October 2026, Finance Minister Enoch Godongwana announced an additional R3.5 billion for smart-meter rollout over the Medium-Term Expenditure Framework, on top of R2 billion already allocated. It cites improved revenue collection in Sol Plaatje, Kannaland and Bela-Bela as evidence of smart metering’s value. In a separate discussion of AI, the article says the Associated Press stylebook now advises journalists not to describe AI as thinking, feeling, wanting or understanding, and Pope Leo XIV’s first encyclical, Magnifica Humanitas, rejects the claim that AI systems possess consciousness.

Key Takeaways

  • Nvidia plans to stop supplying data-center operators that rent more than 50% of their GPUs to Chinese firms; future contracts may require at least 51% of service to go to non-Chinese customers.
  • Nvidia warns that the proposed Remote Access Security Act could create legal risks over remote access to chips. The House passed the bill, but it still needs Senate approval; Commerce has also announced interim restrictions.
  • Nvidia reported $96.2 billion in second-quarter revenue, including $89 billion from data centers, and forecast $108 billion for the third quarter. Investors are watching customer spending and the Vera Rubin platform.
  • Commvault and Zscaler will combine AI-use monitoring and policy controls with recovery services; Coalition added OpenAI, Anthropic, Gemini and Cursor Enterprise to its Wirespeed platform.
  • Instructure’s Project Athena uses Canvas course materials and assessments to create study aids, with read-only access and a commitment not to use student conversations to train external models.
  • Google acquired Howie, an AI email and calendar assistant for small businesses, and plans to use the team’s experience as Google Labs develops related tools.
  • Ramp says the top 1% of US businesses by AI spending spent a record $10,199 per employee in October 2026; AI adoption reached 56.8%.
  • Coinbase CEO Brian Armstrong estimated AI agents could drive $3 trillion to $5 trillion in transaction volume by 2030, with crypto and stablecoins supporting payments.
  • South Africa added R3.5 billion for smart-meter rollout, on top of R2 billion already allocated; Macrocomm uses AI in municipal utility-management systems.
  • Michael Burry remains bearish on Nvidia and Palantir but has added to software holdings including Adobe, Microsoft, Alphabet, Salesforce and Palo Alto Networks. The AP stylebook and Pope Leo XIV’s encyclical reject describing AI as conscious.

Nvidia may stop chip sales to data centers serving mostly Chinese firms

Nvidia has told customers it will stop selling chips to data center operators that rent more than half of their GPUs to Chinese companies, CNBC reported on November 17, 2026. The proposed Remote Access Security Act would give the US government more power to control access to cloud services, and Nvidia warned it could create legal risks for the company and its customers. The bill passed the House in January but still needs Senate approval. The Commerce Department’s Bureau of Industry and Security announced an interim rule to block Chinese companies from remotely accessing AI chips. Reports have also described Chinese startup Moonshot AI seeking Nvidia chips through data centers in Europe and East Asia.

Nvidia plans limits on chip sales to data centers serving Chinese customers

Nvidia has told customers it plans to stop supplying chips to data center operators that rent more than 50% of their GPUs to Chinese companies, according to a CNBC report on November 17, 2026. The proposed Remote Access Security Act would extend US controls to cloud computing access, and Nvidia warned that the rules could create legal risks for chipmakers and customers. The Financial Times reported that Nvidia had told data center clients to direct at least 51% of their services to non-Chinese customers under future contracts. The bill passed the US House in January but had not cleared the Senate, while the Commerce Department’s Bureau of Industry and Security announced an interim rule to prevent Chinese companies from remotely accessing AI chips. CNBC reported that Moonshot AI had sought Nvidia chips through data centers in Europe, and the Trump administration accused it of accessing chips through East Asian data centers.

Nvidia warns proposed US law could bring legal risks over China access

Nvidia has warned chip buyers that the proposed Remote Access Security Act, or RASA, could expose the company and its customers to US legal action, CNBC reported on November 17. The bill would give the US government more power to control access to cloud services, similar to its control over physical hardware. Nvidia said it would not sell chips to customers that rent more than 50% of their GPUs to Chinese companies. RASA would give the US Commerce Department tools to stop Chinese firms from using overseas data centers to access Nvidia chips. Nvidia said in July that the measure, if enacted, could have a material impact on its business.

Nvidia warns of legal risks as US tightens rules on remote AI chip access

Nvidia has been warning chip buyers that the proposed Remote Access Security Act, or RASA, could create legal risks for the company and its customers when Chinese clients access chips remotely. The article says the US House of Representatives has passed a bill, while the Commerce Department has released interim regulations aimed at preventing foreign actors from using advanced computing based on American technology. It also notes that China has domestic chipmakers, but many AI operators prefer Nvidia chips for their performance and reliability. The article mentions claims that Alibaba and ByteDance prefer training their large language models in Southeast Asia.

Nvidia earnings could shape outlook for AI power companies on Nov. 18

Nvidia’s fourth-quarter outlook on November 18 could show whether demand for AI infrastructure is accelerating or slowing. Its second-quarter revenue was $96.2 billion, up 106% from a year earlier, including $89 billion from data centers, and it forecast third-quarter revenue of $108 billion. Investors will watch customer spending and demand for Nvidia’s next-generation Vera Rubin platform. Strong results could also support companies tied to data center power needs, including Constellation Energy, Bloom Energy, and Energy Transfer. Constellation has 20-year power agreements with Amazon and Alphabet, Bloom makes on-site solid oxide fuel cells and reported $1.24 billion in third-quarter revenue, and Energy Transfer moves natural gas to power plants.

Nvidia earnings could signal what comes next for AI energy stocks

Nvidia planned to report its fiscal 2027 third-quarter results after the market closed on Nov. 17, with interim CEO Colette Kress leading the call. The company reported $96.2 billion in second-quarter revenue, up 106% from a year earlier, and forecast $108 billion for the third quarter. Investors will watch for signs that customers will keep expanding AI data centers and demand for Nvidia’s Vera Rubin platform. The article says strong spending could benefit energy suppliers, including Constellation Energy, Bloom Energy and Energy Transfer, which provide nuclear power, on-site fuel cells or natural gas pipeline services.

Commvault and Zscaler join forces to secure enterprise AI

Commvault and Zscaler formed a strategic partnership to help businesses manage risks from using AI and recover from related problems. Their work will focus on finding AI applications and tools, enforcing rules for their use, and detecting and recovering from risky activity. Zscaler will contribute tools that identify AI use and apply security policies, while Commvault will provide recovery services for data, applications and identities. The companies say the partnership responds to AI agents gaining access to business data and the need to control their actions and restore systems after an incident.

Coalition connects major AI services to its Wirespeed security platform

Coalition added integrations for OpenAI, Anthropic, Gemini and Cursor Enterprise to its Wirespeed AI security operations platform. The links bring administrative and access activity from those services into Wirespeed, where the system can flag risky settings, possible data exposure, access problems and signs of account takeover. Wirespeed can investigate alerts and escalate them, while combining AI activity with information from more than 100 security and IT tools. Coalition says customers connect through APIs, so they can add the platform without replacing their existing security tools.

Coinbase CEO predicts AI agents could drive most transactions within a decade

Coinbase CEO Brian Armstrong said at the TomorrowX conference in Austin that AI agents could handle most transactions within a decade. He estimated that they could facilitate $3 trillion to $5 trillion in transaction volume by 2030 and argued that crypto, including stablecoins and Coinbase’s Base network, could support their payments. Armstrong also said prediction markets and derivatives made up 25% of Coinbase’s trading fee revenue in the third quarter, while prediction markets reached a $200 million annualized run rate. He said shares traded on Base had passed $5 billion in volume and urged public companies to consider tokenizing their equity.

Instructure creates Foundry and develops AI study coach Athena

Instructure launched Instructure Foundry, a separate venture for developing AI products for education. Its first project, Project Athena, is an AI study coach designed to work with Canvas and use students’ course materials, assignments and upcoming assessments to create quizzes, study guides and coaching prompts. Instructure says student conversations will not train external AI models, and Athena will connect through a read-only link governed by existing student permissions. Hinds Community College is involved in the initiative after an initial pilot, and the project is expected to extend access to its 7,000 students.

Macrocomm backs South Africa’s push for homegrown AI in utilities

Macrocomm, a South African technology developer with 21 years of experience, builds utility management systems with data processing and AI tools for municipalities. Its software uses AI to improve data accuracy, connect separate datasets, spot unusual activity and risks, and support practical decisions. The article argues that South Africa should develop and control technology used for essential services, while still working with global partners. In October 2026, Finance Minister Enoch Godongwana announced an extra R3.5 billion for smart meter rollout over the Medium-Term Expenditure Framework, in addition to R2 billion already allocated. The article points to improved revenue collection in Sol Plaatje, Kannaland and Bela-Bela as evidence of smart metering’s value.

Michael Burry adds software stocks while staying bearish on Nvidia

Michael Burry, the investor known for his role in The Big Short, remains bearish on Nvidia, Palantir and other AI stocks. He is concerned about high valuations and the risk of a market downturn. In a recent interview, Burry said he had added to holdings in Adobe, Microsoft, Alphabet, Salesforce and Palo Alto Networks. He remains optimistic about the software industry’s long-term prospects, citing its strong fundamentals and potential for growth.

The Catholic Church and media say AI is not conscious

The article says the Catholic Church and legacy media largely agree that AI is not conscious. The Associated Press updated its stylebook this year, telling journalists not to describe AI systems as thinking, feeling, wanting or understanding. Pope Leo XIV made a similar argument in his first encyclical letter, Magnifica Humanitas. The article says he rejected the idea that so-called artificial intelligences possess consciousness.

Google acquires Howie to build AI tools for small businesses

Google acquired Howie, an AI assistant that helps small businesses manage email and calendars, according to a report dated November 17, 2026. Howie schedules meetings through email and text, finds available times in group conversations, creates booking links and follows up. Howie said its team will continue serving existing customers and work on scheduling and calendar management. Google Labs said it hopes to learn from Howie’s experience as it builds AI tools for small businesses. Howie CEO Petersmith said the company was built on Gmail and Google Calendar and sees Google Labs as a place to develop the product further.

Top US firms spend record $10,199 per employee on AI

Ramp data says the top 1% of US businesses by AI spending spent a record $10,199 per employee in October 2026. AI adoption reached 56.8% during the same period. The figures show that AI spending is concentrated among leading firms, which the article says may give them a competitive edge. The article also links the trend to Anthropic, saying its valuation could reach $2 trillion by December at 27.5% YES.

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.

Nvidia AI chips US-China technology restrictions Remote Access Security Act AI infrastructure demand data centers AI energy demand enterprise AI security AI governance education AI AI agents cryptocurrency stablecoins AI investment AI adoption smart meters municipal AI AI consciousness

Comments

Loading...