Samsung officially unveiled the Gauss2 GenAI model at SDC24 Korea. The company’s DX Division is already using it to improve internal productivity. Samsung could utilize Gauss2 GenAI to unlock faster and smarter experiences on Galaxy AI.

Last year, Samsung debuted the first-gen Gauss2 AI model. The second-gen offers improved performance and efficiency. It can also simultaneously handle various data types as a multimodal model integrating language, code, and images.

It is available in three distinct models catered to different purposes: Compact, Balanced, and Supreme. The Korean tech giant plans to integrate it into products to deliver higher levels of convenience and personalization.

Samsung Gauss2 Models – Compact, Balanced & Supreme

Compact Model

• It delivers optimized performance for on-device usage by maximizing the utilization of the device’s computing resources.

Balanced Model

• The Balanced model focuses on achieving a balance between performance, speed, and efficiency, providing consistent performance across diverse tasks.

Supreme Model

• It aims for top-notch performance by leveraging Mixture of Experts technology atop the Balanced model.

The Balanced and Supreme models bring equal or superior performance in key metrics such as generating responses in various languages including English and Korean, and coding capabilities compared to leading open-source GenAI models.

Adding to this, the processing speed (per hour) of Balanced and Supreme models is 1.5 to 3 times faster. Faster and stable processing speed offers quicker AI response generation, reduced user wait times, and more efficient task handling.

Samsung says Gauss2 supports 9 to 14 languages as well as various programming languages. The company has integrated its stabilization techniques for training LLMs and designed a custom tokenizer to ensure maximum efficiency.

The Korean tech giant also revealed that it will continue to expand the reach of its AI-based services across all product lines. By integrating knowledge graph technology with AI, Samsung expects to provide even more enhanced personalization services.

Samsung has invested in the US-based AI (artificial intelligence) semiconductor startup, Enfabrica. The company reportedly plans to introduce the 3.2Tbps ACF SuperNIC chip and pilot systems in the first quarter of next year.

According to the information (via Reuters), Samsung Catalyst Fund (SCF) participated in the Series C funding round of $115 million recently completed by Enfabrica.

SCF joined this investment round, which was led by Spark Capital, as a new investor. Other new investors include Arm, Cisco Investments, Maverick Silicon, and VentureTech Alliance.

Following the investment, Enfabrica shared plans to strengthen R&D and move forward with mass production of its “Accelerated Computing Fabric Super Network Interface Card (NIC)” chips.

About Enfabrica

Former Broadcom and Alphabet employees founded Enfabrica in 2019. The firm develops semiconductors and software, which support the performance and scalability of AI and accelerated computing workloads.

Samsung’s investment in AI startups

Samsung Catalyst Fund has previously invested in US AI semiconductor startup DreamBig Semiconductor and Dutch AI solution startup AxelleraAI.

The consecutive investments in promising overseas AI companies are seen as efforts to secure potential clients and gain leadership in new technologies.

Enfabrica co-founder and Chief Executive Rochan Sankar told Reuters that current technologies can string together about 100,000 AI computing chips before the network starts to bog down.

AMD could launch Ryzen AI Mobile SoC, potentially entering the smartphone application processor market. Currently, the company is known for designing and developing computer processors and graphics technologies.

SmartphoneMagazine reports that AMD could enter the Mobile SoC market with Ryzen AI. If it happens, the company will rival Apple (A Series), Samsung (Exynos), Qualcomm (Snapdragon), MediaTek (Dimensity) and Google (Tensor).

The report says that AMD is already in talks with integrators to have the forthcoming “Ryzen AI” mobile SoCs used in smartphones. It could be a significant development in the tech industry as AMD currently focuses on PC chips and drivers.

It is rumored that AMD plans to introduce “Ryzen AI” chips, similar to what is found in handheld devices. That said, the company will likely have an implementation like Phoenix, Hawk Point, and Strix Point APUs.

Starting with Exynos 2200, AMD has been integrating RDNA-based Xclipse GPU in Samsung’s flagship chipset. The Exynos 2400 also features Xclipse GPU and we’ve seen the expansion to Galaxy A55’s Exynos 1480 chipset earlier this year.

The US chip firm is allegedly setting its sights on the mobile industry and will allegedly introduce “APU-like” Ryzen AI SoCs. The company is considering mobile market entry to answer Qualcomm and MediaTek as they eye the PC market.

Apple itself designs its mobile SoCs and gets produced through TSMC. Qualcomm and MediaTek chips are widely used in Android devices. Samsung also has its own Exynos processor and Google has its Tensor for Pixel devices.

AMD’s entry into the smartphone market could mark a radical development. The chip’s performance and pricing factors will decide the heat of competition. If such things are planned, AMD may partner with a brand to ship its Ryzen AI Mobile chips.