Postcard from @CES 2025: Key Takeaways on AI, Connectivity, and Robotics

Postcard from @CES 2025: Key Takeaways on AI, Connectivity, and Robotics

CES 2025 lived up to its reputation as the global stage for innovation, showcasing major advancements across artificial intelligence (AI), connectivity, robotics, and more. From NVIDIA’s thought-provoking keynote to engaging panel discussions, the event was a rich confluence of ideas, technology, and people shaping the future. As a speaker and participant, I had the opportunity to contribute and absorb insights that reflect the industry’s rapid evolution.

英伟达 Keynote: Setting the Tone for CES 2025

NVIDIA kicked off CES 2025 on Day 0 with a keynote full of insights and bold predictions. Jensen Huang ’s presentation highlighted:

  • AI Evolution: A framework for the next phases of AI, moving from: Perception AI (e.g., computer vision) to Generative AI (e.g., ChatGPT and image synthesis), to Agentic AI (autonomous agents making decisions), and ultimately to Physical AI (embodied AI in robotics).

AI Evolution

  • Scaling Laws of AI: Huang outlined three critical scaling laws:

Pre-training Scaling: Growth driven by compute, model size, and data.

Post-training Scaling: Reinforcement learning boosts capabilities.

Test-Time Scaling: Enhancements from "thinking" at inference.

OpenAI’s O1/O3 models and Google’s Gemini Pro leverage chain-of-thought reasoning, aligning with these scaling laws.

AI Scaling Laws

  • Robotics: Huang’s declaration that the “ChatGPT moment for general robotics is just around the corner” resonated deeply. He categorized robots into three types:

  1. AI Agents (digital robots).
  2. Autonomous Vehicles.
  3. Humanoids.

He predicted that future IT departments will serve as HR for digital agents—a provocative vision for workforce transformation.

Panel Spotlight: Smart Connectivity in the Age of Operational AI

It was exciting to share my perspective on smart connectivity and operational AI as a participant in PwC’s panel, “Smart Connectivity in the Age of Operational AI,” alongside exceptional industry leaders:


Our discussion examined how enterprises can harness IoT data to enable operational AI by leveraging a variety of networks, including Bluetooth, LoRa, Wi-Fi, cellular, and satellite. Key requirements for enterprise connectivity include coverage, data bandwidth, throughput, latency, cost, security, flexibility, and reliability. Balancing these priorities is essential for determining the optimal connectivity solution for each use case.

In practical terms, most enterprise networks will be hybrids—a combination of technologies tailored to specific needs. For example, in a sports arena, low-cost sensors connected via Wi-Fi might detect issues like pipe leakage, while high-speed 5G could support webcams for crowd control, facial recognition, or live video feeds. This hybrid approach ensures each technology is utilized where it performs best, delivering efficient and effective outcomes.

When it comes to AI, the role of connectivity is critical. As we often say, 'data is the new oil,' and as Sam Altman puts it, 'compute is the currency of the future.' Networks are what bring data to compute in the right place, at the right time, and at the right cost.

?The video of our panel discussion should be available soon, and I will inform everyone when it’s live.

Emerging Tech Trends for 2025

CES highlighted several transformative trends shaping the tech landscape:

  1. AI at the Edge: Edge AI allows AI inferencing on or close to devices, reducing latency and cost, while enhancing security and privacy. Edge AI can be on the device or on intermediate points close to the device, like gateways and routers. AI inferencing can also take place on telco edge at different layers of the telecom network. 高通 showcased Snapdragon 8 Gen 3 for AI on smartphones and announced Snapdragon X Series for AI on PC’s and laptops.
  2. Software-Defined Vehicles (SDVs): The automotive buzzword this year, which automakers have gone all-in on with AI integrations, autonomous and advanced driving, in-vehicle displays, passenger comfort features, smart charging and much more.
  3. Spatial Computing: Beyond AR/VR/XR that blend the virtual and physical worlds, spatial computing is about enabling enterprise applications like digital twins, simulating real-world scenarios for better decision-making.
  4. Quantum Computing: While NVIDIA predicts mainstream adoption is 20 years away, some argue quantum is already delivering early value. Whenever quantum computing arrives, the cybersecurity implications will be significant. Some even call it Y2Q (reference to Y2K). Some hackers may stockpile encrypted data now, hoping to decrypt it with quantum tools in the future.
  5. AI is the New UI: As interactions with AI become more intuitive, they are redefining how users engage with technology.
  6. AI agents disrupting SaaS: As AI agents become increasingly capable of executing business logic autonomously, the SaaS model is poised for disruption.


SDV - big buzzword at CES

Lessons from AI Deployment in Enterprises

Panel discussions and keynotes reinforced critical lessons for leveraging AI in the enterprise:

  • Data Management is Paramount: Effective AI starts with clean, well-governed data.
  • Streamline Processes First: Introducing generative AI is futile without optimizing existing workflows.
  • Human-Centric AI: Despite automation, humans remain central to critical decision-making.
  • Responsible AI: includes is a combination of compliance with government regulations plus adherence to the company’s own ethical frameworks.

Networking and Reflections

CES 2025 was more than just an exhibition of technology; it was a gathering of brilliant minds, fostering meaningful connections. From reconnecting with colleagues to meeting new collaborators, the event reaffirmed the value of in-person engagement in driving innovation.

?

Patrick Parodi, Azita Arvani, and John Hoffman





Julia Sweet, Accenture CEO at fireside chat with CNBC's Julia Boorstin
Durga Malladi, SVP and GM at Qualcomm and Azita Arvani


Conclusion

As CES 2025 demonstrated, the convergence of AI, connectivity, and robotics is shaping a future that’s both exciting and complex. From NVIDIA’s visionary keynote to dynamic panels like the one I participated in, the event underscored the industry’s relentless drive toward an autonomous future.

I’d like to close with an inspiring quote that Julie Sweet , CEO of 埃森哲 , says she lives by: "If your dreams don’t scare you, they are not big enough."

?Look forward to collaborating with you on big deployments of AI agents, robots, and enabling networks in 2025!

Here is to my new robot friends at CES:

Artly barista robot making me a chai latte



Robot playing a keyboard


MICHAEL GENNACO

PRINCIPAL: OIR GROUP

1 个月

Thanks for the recap. Very helpful.

Monica Paolini

Principal at Senza Fili

1 个月

Excellent - I now feel almost like I was there :)

要查看或添加评论,请登录

Azita Arvani, Stanford Sloan Fellow的更多文章

社区洞察

其他会员也浏览了