The annual CES extravaganza kicked off 2025 with a trio of groundbreaking keynotes from NVIDIA, Intel, and AMD, setting an electrifying tone for a week brimming with innovation. Here’s a detailed look at the standout announcements from Day 1, along with our thoughts on what’s next.

NVIDIA’s Big Reveal
As expected, NVIDIA’s keynote unveiled their much-anticipated 50-series GPUs. While the headline was predictable, the details showcased why NVIDIA remains a leader in the industry.

NVIDIA cemented its leadership in graphics and AI technology with the unveiling of the GeForce RTX 50 Series, powered by the new Blackwell architecture. The keynote delivered by Jensen Huang revealed an astonishing leap forward in AI-driven rendering, ray tracing, and gaming technology. NVIDIA’s Blackwell GPUs, including the flagship RTX 5090, feature dual flow through fans (yes dual), a staggering 92 billion transistors and deliver up to 3,352 trillion AI operations per second. These advancements promise unprecedented performance gains, with the RTX 5090 outperforming its predecessor, the RTX 4090, by up to 2x in key benchmarks.

Complementing this hardware is DLSS 4. It introduces Multi Frame Generation, leveraging AI to generate up to three frames for every rendered frame, boosting performance by up to 8x. DLSS 4 also employs transformer-based models for ray reconstruction and super resolution, enhancing image quality with reduced ghosting and improved anti-aliasing. The technology will debut in over 75 games.
NVIDIA didn’t stop there. The RTX 50 Series GPUs also enable new levels of realism with RTX Neural Shaders, bringing AI capabilities directly into programmable shaders. This unlocks real-time rendering of film-quality materials and digital humans, using AI to produce lifelike characters and environments. RTX Neural Faces, for example, transforms simple face data into high-quality digital avatars in real time, a breakthrough for gaming and virtual production.
Beyond gaming, NVIDIA showcased innovations for creators and developers with its AI Foundation Models and NIM (NVIDIA Inference Microservices). These technologies enable RTX AI PCs to run advanced generative AI models locally, unlocking capabilities like image generation, speech recognition, and real-time virtual assistants without reliance on data centers.
The GeForce RTX 50-series GPUs, starting at $549, promise next-gen gaming performance and improved efficiency. The 50-series GPUs will also be available in laptops, setting a high bar for portable gaming, with pricing details hinting at a wide range of configurations for various budgets.
RTX 5090: Features 21,760 CUDA cores, a 2.01 GHz base clock, 2.41 GHz boost clock, 32 GB of GDDR7 memory on a 512-bit bus, delivering 1,792 GB/s bandwidth. It has a total graphics power (TGP) of 575 watts, requiring a 1,000-watt PSU. - $1999 USD
RTX 5080: Equipped with 10,752 CUDA cores, 2.3 GHz base clock, 2.62 GHz boost clock, 16 GB GDDR7 memory on a 256-bit bus, offering 960 GB/s bandwidth. It has a 360-watt TGP, with an 850-watt PSU recommended. - $999 USD
RTX 5070 Ti: Comes with 8,960 CUDA cores, 2.3 GHz base clock, 2.45 GHz boost clock, 16 GB GDDR7 memory on a 256-bit bus, providing 896 GB/s bandwidth. It has a 300-watt TGP, requiring a 750-watt PSU. - $749 USD
RTX 5070: Includes 6,144 CUDA cores, 2.16 GHz base clock, 2.51 GHz boost clock, 12 GB GDDR7 memory on a 192-bit bus, delivering 672 GB/s bandwidth. It has a 250-watt TGP, with a 650-watt PSU recommended. - $549 USD
1 of 4
Our limited hands-on experience with the RTX 5090 hinted at a major leap forward for NVIDIA’s architecture. While specifics remain under embargo, it’s clear the 50-series will shake up the competitive landscape. We can’t wait to get our hands on it when we return to the Labs for benchmarking.
Intel’s Early Morning Surprises
Intel’s keynote brightened the early hours, launching at 8:30 AM with a focus on AI advancements. Fresh off last month’s Battlemage GPU launch, Intel had more surprises in store, solidifying its commitment to redefining hardware and software integration.


Intel revealed their new Core Ultra H and HX-series processors, emphasizing high performance and improved efficiency. These chips are expected to set benchmarks for gaming and productivity laptops. The company also showcased updates to the XMX cores in their mobile GPUs, promising significant improvements in AI inference capabilities. However, their integrated GPUs remain based on the Alchemist architecture. Attendees at an invite-only session experienced Intel’s latest innovations in AI and 3D spatial audio. Notable demos included a prototype Audioscenic AMPHI Hi-D laptop with multichannel audio featuring 3D-printed components, as well as a Tencent handheld device showcasing an 11” 3D hardware display.

Despite the lack of direct NVIDIA comparisons, Intel’s hardware has demonstrated its ability to compete in certain areas, as evidenced by our testing of the B580 versus the RTX 4060. NVIDIA continues to dominate at the high end and in datacenter applications, but Intel’s OpenVINO AI framework, combined with XMX cores on both discrete and integrated graphics, offers an accessible alternative for consumers exploring AI workloads.
The upcoming H and HX lineup, paired with high-TDP configurations, might pave the way for some of the most powerful gaming laptops yet.
AMD’s Power-Packed Innovations
Running parallel to Intel’s announcements, AMD’s keynote introduced a slate of hardware aimed at gamers, creators, and AI enthusiasts.
AMD’s unveiling of the Radeon RX 9070 XT, part of the RDNA 4 GPU lineup, occurred not during their keynote but through a separate press briefing—a decision that underscored how little focus was given to this next-generation architecture. While the RX 9070 XT boasts enhanced AI compute, advanced ray tracing, and improved media encoding, these details felt overshadowed by other announcements. Similarly, FSR 4, the latest iteration of FidelityFX Super Resolution promising breathtaking visuals, was briefly mentioned without much fanfare. Despite the potential of these technologies, AMD’s presentation offered minimal depth, leaving enthusiasts eager for more information. A robust lineup of partners, including ASUS, Sapphire, and Gigabyte, is set to support these advancements, but the lack of emphasis during the keynote was noticeable.

The Ryzen 9 9950 X3D took center stage among AMD’s Ryzen 9000 series. Offering 16 Zen5 cores, a 5.7 GHz max boost, and 144MB of cache, it aims to deliver exceptional performance for gamers and creators alike. AMD claims a generational uplift of 8% in gaming and 13% in creative tasks, and their Ryzen 9 9000 HX mobile processors are set to provide desktop-grade performance in portable devices. The presentation also highlighted the Ryzen AI 300 and AI MAX series, which focus on efficient, high-performance machine learning capabilities in laptops.

Despite these advancements, AMD’s keynote was light on additional GPU details, leaving some attendees (and us) eager for more information. Most of these products are slated for release in Q1 and Q2 2025, indicating a strong year ahead for AMD.
What’s Next?
Day 1 of CES 2025 was just the beginning. The week promises more exclusives, surprises, and hands-on demos.
Stay tuned for our coverage as we dive deeper into the groundbreaking products and innovations shaping the future of tech.