AMD Unveils Helios AI Platform Alongside Nvidia's Vera Rubin System

CES 2026 showcased a dynamic landscape for artificial intelligence, with AMD and Nvidia presenting distinct yet powerful visions for the future. AMD CEO Lisa Su unveiled the Helios AI platform, a massive liquid-cooled system featuring advanced Instinct GPUs like the MI455X, boasting 320 billion transistors and 432GB of HBM4 memory. AMD also highlighted its open software ROCm and confirmed a partnership with OpenAI, whose President Greg Brockman stated they use AMD systems for AI agents. The company also launched new Ryzen AI processors for PCs and partnered with Luma AI for video generation. Nvidia CEO Jensen Huang, in contrast, focused on "physical AI," emphasizing systems interacting with the real world through robotics and autonomous vehicles. Nvidia's new Vera Rubin AI chip system, featuring Vera CPU and Rubin GPU, is already in full production and promises significant performance gains, with meaningful revenue expected in the second half of 2026. Nvidia is expanding its ecosystem through partnerships with companies like Cadence, Synopsys, Palantir, Siemens, and PepsiCo, which will use Siemens Digital Twin Composer powered by NVIDIA Omniverse for factory optimization. Market analysts like Morgan Stanley's Joseph Moore reported "skyrocketing" AI demand post-CES, noting Nvidia's Rubin platform as a highlight and a strong ramp-up for AMD's MI455. Wedbush Securities' Joel Kulina affirmed Nvidia's continued leadership, even as the broader AI market expands. Beyond hardware, AI "godfather" Geoffrey Hinton predicts AI will master mathematics within a decade, while former OpenAI expert Daniel Kokotajlo now forecasts autonomous coding in the early 2030s and superintelligence by 2034. OpenAI also reported ChatGPT's significant role in U.S. healthcare, handling nearly 2 million messages weekly related to medical pricing and insurance, with a large portion from rural areas. Marktechpost launched AI2025Dev, a free analytics platform tracking 100 AI releases from 39 companies, noting 69% are open source. Central also introduced CTRL, an AI agent for automating back-office tasks, already serving over 400 companies.

Key Takeaways

• AMD unveiled the Helios AI platform at CES 2026, a liquid-cooled data center system using Instinct MI455X GPUs with 320 billion transistors and 432GB HBM4 memory.
• OpenAI President Greg Brockman confirmed that OpenAI uses AMD systems for AI agents.
• Nvidia introduced the Vera Rubin AI chip system, featuring Vera CPU and Rubin GPU, entering full production in the second half of 2026 with expected revenue.
• Nvidia is expanding its "physical AI" focus, partnering with companies like Cadence, Synopsys, Palantir, Siemens, and PepsiCo for robotics and digital twin applications.
• Morgan Stanley and Wedbush analysts noted skyrocketing AI demand, with strong performance expected from both Nvidia's Rubin and AMD's MI455 products.
• Geoffrey Hinton predicts AI will surpass human capabilities in mathematics within the next 10 years.
• Daniel Kokotajlo, a former OpenAI expert, updated his timeline, forecasting autonomous coding by the early 2030s and superintelligence by 2034.
• OpenAI's ChatGPT is a key tool in U.S. healthcare, handling almost 2 million messages weekly on medical pricing and insurance, with 600,000 from rural areas.
• Marktechpost launched AI2025Dev, a free analytics platform tracking 100 AI releases from 39 companies in 2025, noting 69% are open source.
• Central introduced CTRL, an AI agent runtime designed to automate mission-critical back-office tasks like payroll, already used by over 400 companies.

AMD unveils Helios AI platform at CES 2026

At CES 2026, AMD announced Helios, a powerful new AI platform for data centers. This massive, liquid-cooled system uses advanced Instinct GPUs like the MI455X, which has 320 billion transistors and 432GB of HBM4 memory. AMD also highlighted its open software ROCm and a partnership with OpenAI. The company showed progress with Luma for AI video and world generation and plans for the MI500 chip in 2027. AMD also brings new AI chips to PCs this quarter, including a partnership with Zoom.

AMD CEO Lisa Su unveils new AI chips at CES 2026

At CES 2026, AMD CEO Lisa Su introduced new AI chips and large computing systems, including the Helios rack-scale system for data centers. Helios uses AMD's Instinct MI455 AI chip and EPYC processors to meet the huge demand for AI. OpenAI President Greg Brockman confirmed they use AMD systems for AI agents. AMD also partnered with Luma AI for video generation and launched new Ryzen AI processors for PCs. This shows AMD's focus on integrated AI infrastructure and open tools.

Nvidia and AMD show different AI futures at CES

At CES 2026, Nvidia CEO Jensen Huang and AMD CEO Lisa Su presented different visions for AI's future. Huang focused on "physical AI," where systems interact with the real world through robotics and autonomous vehicles. He sees Nvidia building "AI factories" using digital twins for training. Su, however, emphasized the need for massive computing power, discussing "yottaflop" scale for future AI systems. She positioned AMD's chips as flexible building blocks for data centers, PCs, and edge devices, also highlighting energy efficiency.

Morgan Stanley sees huge AI demand after CES 2026

Morgan Stanley analyst Joseph Moore reported "skyrocketing" AI demand after CES 2026, with Nvidia's new Rubin platform being a highlight. Nvidia's CEO Jensen Huang showcased Rubin, which is already in full production and promises significant performance gains. The company also announced new BlueField DPUs and an inference memory platform. Nvidia expects meaningful revenue from Rubin in the second half of 2026 and continues to see demand from China. For AMD, Moore noted a growing ecosystem around its MI455 product, with OpenAI as a key customer, expecting a strong ramp up later this year.

Nvidia keeps AI lead with Vera Rubin and new partners

Joel Kulina from Wedbush Securities states that Nvidia's leadership in AI remains strong, driven by its new Vera Rubin AI chip system. The Vera Rubin platform, featuring Vera CPU and Rubin GPU, is entering full production in the second half of 2026. Nvidia is also expanding into "physical AI" like robotics and autonomous systems. The company is building a strong ecosystem through partnerships with companies like Cadence, Synopsys, Palantir, and Siemens. While Nvidia's stock has traded flat recently, this is due to the AI market growing broader, not a weakening of Nvidia's core demand.

AI will master math says AI godfather Geoffrey Hinton

Geoffrey Hinton, a "godfather of AI," predicts that artificial intelligence will soon become much better at mathematics than humans. He explains that math is a "closed system" where AI can test ideas and learn, similar to how AI mastered games like Go. Hinton believes this could happen within the next 10 years. Current AI systems like DeepMind's AlphaGeometry already solve complex geometry problems, showing progress in this field.

AI expert delays timeline for superintelligence

Daniel Kokotajlo, a former OpenAI employee and leading AI expert, has updated his timeline for artificial general intelligence or AGI. He now believes that AI will achieve autonomous coding in the early 2030s, later than his previous 2027 prediction. The new forecast sets 2034 as the horizon for "superintelligence," without predicting when AI might destroy humanity. Kokotajlo notes that AI progress is "somewhat slower" and the real world has "enormous inertia" that delays societal change.

Marktechpost launches AI2025Dev for AI data analysis

Marktechpost, an AI news platform, released AI2025Dev, a free analytics platform for AI developers and researchers. This tool turns 2025 AI activity into a searchable dataset covering model releases, training scale, and performance benchmarks. The 2025 release includes data on top research papers, AI startups, and investors. It tracks 100 AI releases from 39 companies, noting that 69% are open source or open weights. AI2025Dev also shows trends like the rise of agentic systems and model efficiency, helping users compare LLM training data and benchmark scores.

OpenAI says ChatGPT helps US healthcare

A report from OpenAI claims that ChatGPT is becoming a key tool in U.S. healthcare. The report shows that almost 2 million messages weekly involve people asking about medical pricing, insurance, and billing. Additionally, 600,000 healthcare messages come from rural areas each week, and most queries happen when clinics are closed. OpenAI's survey found that three in five U.S. adults have used AI tools for healthcare recently. The report highlights how AI can help people in underserved areas and assist doctors with tasks like drafting visit notes.

Dow hits record high as oil and AI stocks rally

The Dow Jones Industrial Average surged over 600 points today, reaching a new record high. The S&P 500 and Nasdaq also saw big gains. This market rally was partly driven by a jump in oil stocks following a US military strike and the capture of President Maduro in Venezuela. At the same time, strong demand for artificial intelligence technologies boosted the tech sector. Investors are watching both geopolitical events and tech advancements.

Central launches CTRL AI agent for back office tasks

Central, a company building AI-native back-office platforms, announced CTRL, an AI agent runtime for mission-critical operations. CTRL helps automate tasks like payroll by translating natural language requests into multi-step workflows. CEO Josh Wymer emphasized that CTRL keeps users in control, allowing them to approve or change each step. The system is designed for security and compliance, offering audit trails and reproducible workflows. CTRL is now available for early access customers and has been used in Central's Slack bot for over two years, serving more than 400 companies.

PepsiCo partners with Siemens and Nvidia for AI digital twins

At CES 2026, PepsiCo announced a major partnership with Siemens and NVIDIA to use AI and digital twins for its plant and supply chain operations. PepsiCo aims to optimize its facilities and meet growing demand by using a digital-first planning strategy. They will use Siemens Digital Twin Composer, powered by NVIDIA Omniverse, to create realistic 3D digital twins of their factories and warehouses. This technology allows them to simulate and optimize layouts before physical construction. This collaboration helps PepsiCo boost capacity and gain a real-time view of its operations.

Sources

• AMD at CES 2026: Helios Signals a Full-Stack AI Push
• AMD Showcases Growing AI Hardware Arsenal at CES2026
• Nvidia and AMD Reveal Dueling Paths for AI’s Future
• CES recap: Morgan Stanley flags ’skyrocketing’ AI demand By Investing.com
• Nvidia’s AI Dominance Is Intact, Driven by Vera Rubin and Ecosystem Expansion
• AI Will Excel At Math Because It’s A Closed System That Can Be Played: Geoffrey Hinton
• Leading AI expert delays timeline for its possible destruction of humanity
• Marktechpost Releases ‘AI2025Dev’: A Structured Intelligence Layer for AI Models, Benchmarks, and Ecosystem Signals
• Report from OpenAI Claims ChatGPT Is Becoming an Important Complement to U.S. Healthcare
• Stock market today: Dow jumps 600 points to record, S&P 500 and Nasdaq rally as oil stocks surge after Venezuela strike
• Central announces CTRL: The AI Agent Runtime for Mission-Critical Products
• PepsiCo Announces Industry-First AI and Digital Twin Collaboration with Siemens and NVIDIA