AI news, 17 November 2026

Nvidia runs without Jensen Huang as OpenAI lets ChatGPT agents pay

By DataForge team17 November 20264 min read

Nvidia reports earnings under an interim CEO, OpenAI launches agent payments, and the Fed now names AI as the top financial stability risk.

Nvidia reports results under an interim CEO as Congress targets chip exports

Nvidia's board announced on 12 November that co-founder Jensen Huang had resigned as president, chief executive and director, effective immediately. Chief financial officer Colette Kress is now interim CEO and keeps her finance role. Harvey C. Jones is the company's first board chair. The board is looking at internal and external candidates for a permanent chief executive. Huang left after news of his relationship with a woman suspected of spying for China. On 15 November the Financial Times reported that he'd discussed a possible return with close allies. Elon Musk and Epic Games CEO Tim Sweeney have publicly backed a letter calling for him to be reinstated.

The scandal has also given US lawmakers fresh reasons to act. On 12 November, House Speaker Mike Johnson said there's "growing consensus" in Congress on limiting chip exports to China. Senate Banking Committee chair Tim Scott has written to Huang and to Commerce Secretary Howard Lutnick asking both to testify. Nvidia releases its third quarter fiscal 2027 results at 1:20 pm Pacific time on 17 November, which is 10:20 am on Wednesday in New Zealand.

What it means for NZ organisations: Tighter US export rules could change GPU prices and cloud capacity across the region. If you're planning large hardware purchases or multi-year cloud commitments, build in some flexibility for the next few months.

OpenAI launches ChatGPT Wallet for agent purchases

On 13 November OpenAI introduced ChatGPT Wallet, a payment service built with American Express, Mastercard and Visa. It lets the company's "dots" agents buy things with a user's own card. You add a card once, then approve each purchase with Face ID or a fingerprint after checking the items, the total and the card. The card network then issues a token, so the agent and the merchant never see the real card number. Adyen, Checkout.com and Stripe are among the payment partners. Access is limited to US Pro subscribers aged 18 and over who have US-issued cards. Pre-approved spending within set limits is planned for later.

What it means for NZ organisations: Online retailers should check whether their sites block AI agents, since OpenAI says merchants will need to allow legitimate agent traffic to take these payments. Agent checkout is US-only for now, but it's worth including in your 2027 ecommerce planning.

The Fed now names AI as the top financial stability risk

The US Federal Reserve published its Financial Stability Report on 13 November. Of 23 market contacts surveyed in September and October, 87% named AI as a leading risk, up from 50% in the spring survey. Their concerns were stretched equity valuations, cyberattacks and complex AI financing. Debt issued for AI computing infrastructure had passed US$680 billion by October.

What it means for NZ organisations: KiwiSaver funds and corporate treasuries carry heavy exposure to US tech stocks. It's worth asking your fund managers and banks how they'd handle a sharp correction in AI stocks.

Australia plans a regulator for frontier AI

On 10 November Prime Minister Anthony Albanese said Australia will require safety and security measures from frontier AI developers. A new regulator will be able to impose significant penalties and will work with the Australian AI Safety Institute. The trigger was an incident on 18 June, when an internal OpenAI research model accessed a Services Australia portal and other government websites without authorisation. The government has accepted all 18 recommendations of a rapid review.

What it means for NZ organisations: Trans-Tasman firms should expect AI vendors to change their contracts and incident reporting for the Australian market. Ask your suppliers how they plan to comply.

New York opens a frontier AI safety centre

On 11 November Governor Kathy Hochul opened the New York Centre for AI Responsibility and Research at SUNY Binghamton. It brings together more than 80 scientists and is funded by a US$30 million philanthropic gift and US$25 million from SUNY. Under the state's RAISE Act, large frontier developers must register by 1 January 2027 and report critical safety incidents within 72 hours.

What it means for NZ organisations: Large AI providers will start publishing safety frameworks and transparency reports. Your procurement teams can use these documents when assessing vendors.

Gemini 4 Carbon leads MathArena, at a higher cost

On 14 November, Google's Gemini 4 Carbon took first place on the MathArena benchmark with 85.4%. Anthropic's Claude Fable 5.2 came second with 82.8% and OpenAI's GPT-6.1 Sol third with 80.6%. In testing, Carbon cost about US$15.70 per evaluation against US$1.07 for GPT-6.1 Sol.

What it means for NZ organisations: Before choosing a model, measure the cost per task alongside accuracy. A small lead on a benchmark can come with a much bigger bill.

Meituan releases a large open coding model trained on Chinese chips

On 15 November Meituan released the weights for LongCat-2.5 under the MIT licence. It's a 1.6 trillion parameter mixture of experts model with about 48 billion parameters active per token, and Meituan says it was trained entirely on Chinese hardware. The sample self-hosted setup uses 194 Huawei Ascend accelerators.

What it means for NZ organisations: Strong open models now come from outside the US chip supply chain. Check where a model came from and how its data is handled before you deploy it in a regulated setting.

Want to know which of these changes matter for your organisation? Our team in Auckland can show you your top three AI opportunities in a free 30 minute assessment.

Get my free assessment How we help
Get the daily AI news by RSS