Bloomberg Tech at CES | Bloomberg Tech 1/6/2026

Key Concepts

• AI Compute Demand: Exponentially increasing demand for AI processing power, projected to grow from 1 billion to over 5 billion users in the next five years.
• MI455 & MI440X: AMD’s new generation of AI accelerators, designed for cloud data centers (MI455) and enterprise applications (MI440X) respectively.
• Helio: AMD’s first rack-scale system solution, a powerful chip designed to meet the growing demand for AI compute.
• Open Ecosystem: AMD’s commitment to an open and collaborative approach to AI development, partnering with companies like OpenAI and Oracle.
• Compute Deficit: A significant shortage of available computing power to meet the rapidly increasing demands of AI development and deployment.
• Multimodal AI Wearables: The shift towards integrating AI capabilities into everyday wearable devices like headphones, rather than solely focusing on smart glasses.
• Supply Chain Resilience: Challenges and strategies related to securing and localizing supply chains, particularly in light of geopolitical factors and tariffs.

AMD Accelerates AI with New Chips & Partnerships – A Bloomberg Tech Deep Dive

This transcript details a Bloomberg Tech broadcast from CES in Las Vegas, focusing heavily on AMD’s advancements in AI hardware and strategy, alongside broader industry trends. The conversation, led by Caroline Hyde and Ed Ludlow, features an in-depth interview with AMD CEO Lisa Su.

1. AMD’s New AI Hardware: Helio, MI455, and MI440X

AMD unveiled its Helio chip, described as a “massive system” and the world’s first rack-scale system solution. This chip is built using 2 and 3 nanometer technology, boasting 300 billion transistors, and is designed to address the escalating demand for AI compute. Su emphasized that the demand for AI compute is increasing, with active users projected to grow from 1 billion today to over 5 billion in the next five years.

Alongside Helio, AMD announced the MI455 chip, targeted at cloud data centers, and the MI440X, designed for enterprise applications. The MI440X utilizes the same building blocks as the MI455 but is optimized for deployment in existing data centers, allowing enterprises to upgrade their AI capabilities without building new infrastructure. Su highlighted that enterprises are increasingly adopting AI across various sectors, including financial services and healthcare, often preferring to maintain control over their data centers.

2. Addressing the Compute Deficit

A central theme of the interview was the growing “compute deficit” – a significant shortage of processing power needed to support the rapid expansion of AI. Su quantified this, stating that compute needs to increase by 100 times over the next four to five years, equating to 10 to the 24th power in FLOPS (floating-point operations per second). She illustrated this by noting that current models are already powerful, but have room for improvement, particularly in areas like software development.

3. Partnerships & Open Ecosystem

AMD is prioritizing partnerships to address the compute deficit and foster innovation. Su highlighted collaborations with OpenAI, Oracle, and others. She emphasized AMD’s commitment to an “open ecosystem,” believing that collaboration and industry standards are crucial for accelerating AI development. This contrasts with NVIDIA’s more vertically integrated approach.

4. Deployment Timeline & Target Users

The Helio chip is slated for deployment in 2026, with strong partnerships already in place. The MI440X is designed for immediate enterprise adoption, allowing companies to integrate new AI capabilities into existing data centers. Su noted that OpenAI is exploring use cases for Helio, and partnerships with Oracle are also underway.

5. Competitive Landscape & NVIDIA

When asked about AMD’s competitive advantage over NVIDIA, Su emphasized AMD’s focus on “outstanding performance at an advantageous total cost of ownership” and its commitment to an open ecosystem. She positioned AMD as a provider of solutions for a wide range of AI applications, from large-scale cloud models to enterprise use cases.

6. Supply Chain & Geopolitical Considerations

The conversation touched upon supply chain challenges, particularly regarding access to China. Su stated that China is an important market for AMD and that they are working with the U.S. government to obtain licenses for exporting newer generation chips. She acknowledged the need to localize supply chains to mitigate the impact of tariffs and geopolitical tensions, citing plans to bring battery production to the U.S.

7. Future of AI & Innovation

Su expressed optimism about the future of AI, stating, “You ain’t seen nothing yet.” She predicted continued rapid innovation in AI technology and applications, with AI becoming increasingly integrated into everyday life. She highlighted the potential of AI to enhance productivity and improve business processes across various industries. She also discussed AMD’s foray into physical AI with its partnership with Bionics and the development of humanoid robots.

8. Broader CES Trends (Post-Interview Discussion)

Following the interview with Lisa Su, the broadcast discussed broader trends at CES, including:

• NVIDIA’s Rally: NVIDIA’s stock saw a slight increase following positive news regarding potential license approvals for exporting to China.
• Memory Chip Demand: The continued high demand for memory chips, driven by AI, benefiting companies like Sandisk and Micron.
• Robotaxi Development: The collaboration between Lucid and Uber on autonomous vehicles, and the integration of NVIDIA’s technology.
• AI Wearables: A shift towards integrating AI capabilities into diverse wearable form factors beyond smart glasses, such as headphones and jewelry.
• Consumer Sentiment: A general effort by companies to make AI more approachable and less intimidating for consumers, often through the use of “cuteness” and user-friendly interfaces.

Notable Quotes:

• Lisa Su: “You ain’t seen nothing yet.” – Expressing optimism about the future of AI innovation.
• Lisa Su: “We see AI now in every part of compute.” – Highlighting the pervasive nature of AI across various applications.
• Lisa Su: “Our claim to fame is outstanding performance at advantageous total cost of ownership.” – Positioning AMD’s competitive advantage.

Synthesis/Conclusion:

AMD is aggressively positioning itself as a key player in the rapidly evolving AI landscape. Through the development of powerful new chips like Helio, MI455, and MI440X, coupled with strategic partnerships and a commitment to an open ecosystem, AMD aims to address the growing compute deficit and unlock the full potential of AI. The company’s focus on both cloud and enterprise applications, alongside its exploration of physical AI, demonstrates a comprehensive strategy for capitalizing on the transformative power of artificial intelligence. The broader CES coverage highlighted the increasing integration of AI into various consumer products and the ongoing challenges related to supply chain resilience and geopolitical factors.