openai, nvidia and amd Updates

OpenAI announced a significant partnership with Cerebras Systems Inc. on January 14, 2026, a move set to add 750 megawatts of ultra low-latency AI compute to its platform. This multiyear deal, with capacity rolling out until 2028, aims to make ChatGPT the fastest AI platform globally, unlocking new applications and reaching a billion more users. Cerebras, known for its unique giant chips that process information rapidly by keeping all data on one chip, previously demonstrated its hardware running an OpenAI model 15 times faster than other systems. This collaboration challenges competitors like Nvidia and complements OpenAI's existing use of chips from Nvidia, AMD, and Broadcom. Meanwhile, Oracle faces a proposed class action lawsuit filed by bondholders on January 14 in a New York state court. The suit claims losses tied to Oracle's AI infrastructure investments, impacting investors who bought $18 billion worth of notes and bonds in September. Oracle's co-founder Larry Ellison and its banks are named as defendants, causing the company's stock to drop nearly 5%. Despite such challenges, Wall Street maintains a bullish outlook on AI. Alphabet's market cap recently hit $4 trillion after two major partnerships, and its Google stock has become the top-performing Big Tech stock, reflecting investor confidence in companies demonstrating measurable AI returns and cost-cutting initiatives like Walmart's AI-powered drone delivery. The geopolitical landscape for AI is also evolving, with major AI companies like Meta and OpenAI shifting their policies to support U.S. military efforts. OpenAI, for instance, changed its stance in January 2024, driven partly by the high costs of developing advanced AI models. Concurrently, the United Arab Emirates joined the U.S.-led Pax Silica initiative on January 14, a program designed to secure AI and semiconductor supply chains and reduce reliance on rival nations. This initiative, which includes Australia, Britain, Israel, Japan, Qatar, Singapore, and South Korea, highlights the UAE's role in logistics, capital, and energy, further solidified by a multibillion-dollar deal for data center hubs in Abu Dhabi using U.S. technology. The massive power demands of AI data centers are sparking debate, with opinion contributors Mark MacCarthy and David M. Klaus arguing that these centers should bear their own electricity costs rather than burdening average consumers. They point out that the electric grid lacks spare capacity, necessitating new, expensive infrastructure. In a regulatory development, Tennessee's 2026 legislative session includes a proposed law making it a Class A felony to knowingly train AI to encourage suicide or homicide, or to simulate a human. On the application front, McKinsey is now incorporating an AI chatbot into its graduate recruitment process, assessing candidates on their ability to collaborate and reason with AI, rather than technical expertise. In the media sector, Datavault AI Inc. and Fintech.TV announced a partnership to enhance viewer engagement. Fintech.TV will integrate Datavault AI's patented content detection system, featuring a real-time bias meter with color-coded indicators and interactive polling powered by ADIO Inaudible Tone technology. This collaboration aims to combat misinformation, promote balanced media, and tap into the fintech market, which is projected to exceed $1 trillion by the early 2030s.

Key Takeaways

• OpenAI partnered with Cerebras Systems Inc. on January 14, 2026, to add 750 megawatts of AI compute, aiming to make ChatGPT the fastest AI platform globally by 2028.
• Cerebras's unique giant chips, which previously ran an OpenAI model 15 times faster, will enhance AI inference, complementing OpenAI's use of chips from Nvidia, AMD, and Broadcom.
• Oracle faces a proposed class action lawsuit filed January 14 by bondholders claiming losses from its AI buildout, leading to a nearly 5% stock drop.
• Wall Street remains bullish on AI, with Alphabet's market cap reaching $4 trillion, and investors favoring companies demonstrating measurable returns and cost-cutting via AI.
• Major AI companies like Meta and OpenAI shifted policies in January 2024 to support U.S. military efforts, driven by high development costs and defense sector applications.
• The UAE joined the U.S.-led Pax Silica initiative on January 14 to secure AI and semiconductor supply chains, contributing logistics, capital, and industrial capacity.
• Opinion contributors argue that AI data centers should pay their own electricity costs for new infrastructure, rather than burdening average consumers.
• Tennessee's 2026 legislative session includes a proposed law making it a Class A felony to train AI to encourage suicide or homicide, or to simulate a human.
• McKinsey is using an AI chatbot in graduate recruitment to assess candidates' collaboration and reasoning skills with AI, reflecting a workforce with 20,000 AI agents.
• Fintech.TV integrated Datavault AI's real-time bias meter and interactive polling, powered by ADIO Inaudible Tone technology, to combat misinformation and engage viewers.

OpenAI partners with Cerebras for faster AI

OpenAI announced a partnership with Cerebras on January 14, 2026. This deal will add 750MW of ultra low-latency AI compute to OpenAI's platform. Cerebras uses unique giant chips to make AI models respond much faster by removing bottlenecks. This integration will make AI interactions more natural and allow real-time AI to reach more people. Sachin Katti from OpenAI and Andrew Feldman from Cerebras both expressed excitement about the collaboration. The new capacity will become available in phases until 2028.

OpenAI partners Cerebras for faster ChatGPT

OpenAI signed a multiyear deal with Cerebras Systems Inc. to use 750 megawatts of computing power. Greg Brockman, OpenAI's co-founder, stated this partnership will make ChatGPT the fastest AI platform globally. This speed aims to unlock new uses and bring AI to a billion more users. Cerebras, a semiconductor startup, uses unique large chips to process information quickly, challenging Nvidia Corp. The companies have explored working together since 2017, and Cerebras previously ran an OpenAI model 15 times faster than other hardware. Andrew Feldman, Cerebras CEO, highlighted his products' strength in the AI inference stage.

OpenAI teams with Cerebras for faster AI chips

OpenAI, the company behind ChatGPT, announced it will use chips from Cerebras, a startup in Sunnyvale, California. This agreement will add 750 megawatts of electricity consumption for Cerebras chips, which will be used for AI inference. OpenAI President Greg Brockman said this partnership will make ChatGPT the fastest AI platform in the world. Cerebras, co-founded by Andrew Feldman in 2015, designs unique giant chips that keep all data on one chip to operate faster. This deal is part of OpenAI's larger effort to expand its computing power, which also includes chips from Nvidia, AMD, and Broadcom.

Oracle sued by bondholders over AI losses

Oracle faces a proposed class action lawsuit filed by bondholders in a New York state court on January 14. The lawsuit claims losses tied to Oracle's AI buildout. Investors who bought $18 billion worth of notes and bonds issued by Oracle in September are part of the suit. Oracle's co-founder Larry Ellison and the company's banks are also named as defendants in the case.

Oracle stock falls after bondholder lawsuit

Oracle's stock dropped nearly 5% on Wednesday following news of a proposed class action lawsuit. Bondholders filed the suit in a New York state court. The lawsuit claims losses related to Oracle's investments in AI infrastructure.

Fintech TV adds AI bias meter and polling

Datavault AI Inc. and Fintech.TV announced a new partnership to enhance viewer engagement. Fintech.TV will integrate Datavault AI's patented content detection system, including a real-time bias meter and interactive polling. The bias meter will show viewers potential biases in news content using visual cues like color-coded indicators. Datavault AI's ADIO Inaudible Tone technology will power the polling, allowing viewers to respond via mobile devices without interrupting the broadcast. This collaboration aims to promote balanced media and create new revenue streams in the growing fintech market, projected to exceed $1 trillion by the early 2030s. Vince Molinari of Fintech.TV and Nathaniel Bradley of Datavault AI believe these tools will combat misinformation and improve audience participation.

Tennessee proposes new laws on AI and pregnant drivers

Tennessee's 2026 legislative session has begun, with several new bills introduced. One proposed law would make it a Class A felony to knowingly train artificial intelligence to encourage suicide or homicide, or to simulate a human. Another bill aims to allow pregnant drivers in their second or third trimester, or those with high-risk pregnancies, to receive a disabled driver placard. Other proposed laws include prohibiting children under 14 from paid content creation and imposing a tax on electric vehicle charging stations. Schools would also need to implement a system to record early warning signs in students.

AI companies now support US military efforts

Major AI companies like Meta and OpenAI have shifted their stance on military use of their tools. OpenAI, for example, changed its policy in January 2024. This change is partly due to the high costs of developing advanced AI models, with the defense sector offering a significant application area. The relationship between states and large technology companies has evolved from neoliberal free market ideals to a focus on geopolitical concerns. Previously, the "Silicon Valley Consensus" allowed tech companies a free hand with minimal regulation. Now, governments are increasingly influencing the tech sector, as seen with trade agreements and responses to foreign digital taxes.

McKinsey uses AI chatbot for graduate hiring

McKinsey is now asking graduate applicants to use an AI chatbot as part of its recruitment process. This "AI interview" focuses on how candidates collaborate with and reason using the AI tool, not on their technical AI expertise. Applicants receive a business scenario and use the AI to explore information, structure thoughts, and refine insights. CaseBasix, a company that helps candidates, reports that the AI interview assesses judgment and communication skills. McKinsey CEO Bob Sternfels stated the firm has a "workforce" of 20,000 AI agents working alongside its 40,000 staff. This move highlights a growing trend where AI competence is becoming a key requirement for many jobs.

UAE joins US AI supply chain program

The United Arab Emirates joined the U.S.-led Pax Silica initiative on January 14, strengthening economic ties with the United States. This program aims to secure AI and semiconductor supply chains and reduce dependence on rival nations. Other members include Australia, Britain, Israel, Japan, Qatar, Singapore, and South Korea. U.S. Undersecretary of State Jacob Helberg noted the UAE's important contributions in logistics, industrial capacity, capital, and energy. The UAE also has a multibillion-dollar deal to build data center hubs in Abu Dhabi using U.S. technology. While Saudi Arabia is not in Pax Silica, it has a significant bilateral AI deal with the U.S.

AI data centers should pay their own electricity costs

Opinion contributors Mark MacCarthy and David M. Klaus argue that AI data centers should pay their own electricity costs, separate from average consumers. They state that claims from AI developers, like Nvidia's Levi Patterson, that data centers will lower overall rates are incorrect. The electric grid lacks spare capacity, meaning utilities must build new generation and transmission facilities to meet the massive power demands of AI data centers. These new investments, often for expensive power sources like natural gas and nuclear, should not be paid for by regular ratepayers. Data centers also require costly infrastructure upgrades for uninterrupted 24/7 power. The authors suggest a "separate checks" model where hyper-scalers pay for their dedicated infrastructure.

Wall Street stays bullish on AI despite bubble fears

Wall Street remains optimistic about AI, even with concerns about a potential bubble. Alphabet's market cap reached $4 trillion recently after announcing two major partnerships, showing real-world AI applications. The company's stock, which faced pressure in early 2025 due to antitrust issues and competition, has become the top-performing Big Tech stock. Investors now favor companies demonstrating responsible spending and measurable returns from AI, including cost-cutting initiatives like Walmart's AI-powered drone delivery. While some experts like Bob Elliott question the long-term economic impact of AI replacing workers, investors are keen to see earnings expand throughout 2026. Jennifer Bender of Jane Street noted that investors are holding onto AI stocks to capitalize on potential earnings growth.

Sources

• OpenAI partners with Cerebras
• OpenAI Forges $10 Billion Deal With Cerebras for AI Computing
• OpenAI Signs Another Deal With a Computer Chip Maker
• Oracle sued by bondholders over losses tied to AI buildout
• Oracle tumbles as bondholders sue over AI-related losses
• AI bias meter lets Fintech TV viewers see news slant on screen
• TN bills would allow some pregnant drivers to park in accessible spots, create felony for dangerous AI training
• How AI Companies Got Caught Up in US Military Efforts
• McKinsey asks graduates to use AI chatbot in recruitment process
• UAE bolsters economic ties with US by joining AI supply chain program
• AI data centers and average electricity users need separate checks
• Why Wall Street isn't hopping off the AI bubble just yet