AAI 2026: AMD Hadirkan Komputasi Full-Stack untuk Era Agentic AI

SAN FRANCISCO – 23 Juli 2026 – AMD (NASDAQ: AMD) hari ini meluncurkan portofolio infrastruktur AI generasi berikutnya dan physical AI di Advancing AI 2026, dipimpin oleh solusi rack-scale AMD Helios yang kini telah memasuki produksi dan akan diterapkan oleh perusahaan AI terkemuka dalam skala gigawatt.

Seiring AI berkembang dari pelatihan menuju inferensi dan beban kerja agentic, permintaan komputasi meningkat dengan cepat. AMD menghadirkan platform AI terbuka dan full-stack yang memberikan fleksibilitas kepada pelanggan untuk menerapkan komputasi yang tepat bagi setiap beban kerja.

“Fase AI berikutnya akan mencakup frontier models, agents, dan physical AI, sehingga menciptakan peluang baru untuk menghadirkan kecerdasan di mana saja,” ujar Dr. Lisa Su, Chair dan CEO AMD. “Untuk mewujudkan potensi tersebut, seluruh industri perlu bekerja sama. AMD bermitra di seluruh ekosistem untuk menghadirkan komputasi terdepan dan platform terbuka yang memberi pelanggan performa, fleksibilitas, dan pilihan untuk menskalakan AI dari data center hingga edge.”

AMD Helios: Solusi AI Rack-Scale dengan Performa Tertinggi

Menghadirkan frontier AI membutuhkan arsitektur rak yang terintegrasi penuh, dengan setiap bagian dari stack mendorong batas performa. Solusi rack-scale AMD Helios dirancang untuk kebutuhan ini, dengan silikon yang dioptimalkan bersama, mencakup 72 GPU AMD Instinct™ MI455X berkinerja tinggi dan 18 CPU AMD EPYC™ “Venice” Generasi Keenam yang bertenaga, terhubung melalui jaringan front-end, scale-up, dan scale-out AMD Pensando™, serta dipercepat oleh software terbuka AMD ROCm™. AMD Helios menggabungkan performa komputasi terdepan, kapasitas memori, dan bandwidth jaringan untuk menghadirkan hingga 30% lebih banyak token per dolar dibandingkan solusi kompetitor terdepan.

Laboratorium AI dan penyedia cloud terkemuka memilih AMD Helios karena performa full-stack dan keterbukaannya. Selain itu juga mencakup OpenAI, Anthropic, Meta, Microsoft, Oracle, HUMAIN, Tensorwave, Vultr, Cirrascale, dan lainnya. Sistem akan tersedia dari OEM terkemuka, termasuk Bull, HPE, Lenovo, dan Supermicro, serta mitra infrastruktur Sanmina dan Wiwynn.

Dalam acara Advancing AI, para mitra AMD menjelaskan bagaimana mereka menerapkan infrastruktur AI AMD dalam skala besar untuk pelatihan frontier dan inferensi:

• Anthropic dan AMD menjabarkan lebih lanjut pengumuman kemitraan strategis pada hari Rabu untuk menerapkan GPU AMD Instinct MI455X hingga 2 gigawatt dalam solusi rack-scale AMD Helios. Kedua perusahaan memulai kolaborasi engineering multiyear untuk menggunakan Claude guna mempercepat pengembangan software AMD. Secara khusus, tim akan menggunakan Claude untuk mengoptimalkan beban kerja pada GPU AMD Instinct dan mempercepat pengembangan software ROCm. AMD juga akan mengadopsi Claude secara luas di seluruh tim engineering dan pengembangan produknya.

• OpenAI dan AMD bermitra untuk mengoptimalkan seluruh AI stack, dari silikon hingga software. Dengan memanfaatkan framework Triton milik OpenAI bersama software AMD ROCm, kedua perusahaan mengoptimalkan beban kerja kelas GPT pada GPU AMD Instinct MI455X dan rak AMD Helios. OpenAI memperkirakan mulai mengoperasikan Helios pada kuartal keempat 2026, dengan penerapan yang meningkat sepanjang 2027.

• Meta dan AMD melakukan co-design untuk penerapan skala gigawatt, dengan mengoptimalkan full AI compute stack AMD bagi beban kerja Meta. Meta kini memvalidasi platform CPU EPYC Generasi Keenam di laboratoriumnya dan telah mulai menguji serta memvalidasi beban kerja pada rak AMD Helios sebagai persiapan untuk penerapan berskala besar.

• Cerebras dan AMD berkolaborasi menghadirkan solusi gabungan komputasi AI ultra-low-latency Cerebras dan infrastruktur rack-scale high-throughput AMD Helios untuk membantu meningkatkan efisiensi, skalabilitas, dan keekonomian inferensi bagi layanan inferensi ultra-low-latency.

Menghadirkan CPU dan GPU Data Center dengan Performa Tertinggi

Prosesor EPYC Generasi Keenam menghadirkan portofolio CPU server terluas untuk agentic AI,2 yang mencakup beban kerja cloud, enterprise, general-purpose, dan high-performance computing (HPC). Dengan performa per core yang terdepan dan kepadatan thread tertinggi,3 prosesor ini memungkinkan jumlah agent terbanyak per watt, per dolar, dan per rak.4,5,6 Untuk AI host node, CPU EPYC Generasi Keenam menghadirkan kecepatan dan bandwidth memori yang dibutuhkan agar accelerator terus memperoleh pasokan data secara penuh. Untuk server general-purpose, prosesor ini menghadirkan performa dan efisiensi energi terdepan guna menjalankan aplikasi business-critical serta tugas pendukung AI.

Dengan GPU AMD Instinct™ MI400 Series, AMD menghadirkan performa tinggi untuk beban kerja cloud, enterprise, dan HPC. GPU AMD Instinct MI455X menghadirkan throughput token 34 kali lebih tinggi dibandingkan GPU MI355X.7 Untuk beban kerja presisi tinggi, accelerator AMD Instinct™ MI430X merupakan solusi paling canggih untuk HPC dan sovereign AI, dengan performa FP64 berbasis hardware hingga 288 TFLOPS untuk komputasi ilmiah. Accelerator Instinct MI430X mendukung gelombang berikutnya dari superkomputer kelas exascale di Amerika Serikat dan Eropa.

AMD juga meluncurkan GPU Instinct MI350P, yang menghadirkan akselerasi AI secara mulus pada infrastruktur yang ada dengan keekonomian token terdepan. GPU MI350P menghadirkan hingga 4,2 kali lebih banyak token per detik per dolar dibandingkan kompetitor.8

Memajukan Ekosistem Open Software

Bagi developer, AMD ROCm adalah platform software terbuka yang menyediakan performa, fleksibilitas, dan dukungan ekosistem yang diperlukan untuk membangun serta menerapkan AI pada hardware AMD. Berlandaskan fondasi tersebut, AMD memperkenalkan ROCm.ai, sebuah platform pengembangan berbasis AI yang membantu developer membangun, mengoptimalkan, dan menerapkan software GPU dengan lebih cepat di seluruh platform AMD. ROCm.ai menghadirkan pemrograman GPU berbantuan AI dengan memungkinkan coding agent populer seperti Claude, Codex, dan Cursor memahami platform AMD dan ROCm secara native. 

ROCm.ai mempercepat enablement software bagi GPU AMD Instinct MI455X sekaligus mengoptimalkan performa. Framework open-source terkemuka, termasuk PyTorch, Hugging Face, vLLM, dan SGLang, telah di-enable pada MI455X dan menunjukkan hasil yang sangat baik.

Mempercepat Infrastruktur AI Generasi Berikutnya

AMD memperpanjang cadence inovasi tahunan untuk CPU, GPU, jaringan, dan rack-scale hingga 2030. Perusahaan membagikan rincian baru mengenai roadmap, termasuk:

• CPU server EPYC generasi berikutnya berbasis arsitektur “Zen 7” akan hadir pada 2028. CPU “Florence,” “Ferrara,” dan “Fidenza” diperkirakan memperluas kepemimpinan AMD dalam kepadatan, performa, performa per dolar sistem, dan performa per watt.

• CPU “Ravenna” berbasis arsitektur “Zen 8” akan hadir pada 2030, dirancang untuk melanjutkan kepemimpinan CPU server AMD.

• GPU AMD Instinct MI500 Series generasi berikutnya akan hadir pada 2027 dengan teknologi komputasi, memori, dan interconnect generasi berikutnya untuk performa terdepan.

• GPU AMD Instinct MI600 Series akan hadir pada 2028.

• Solusi rack-scale AMD Helios 500 generasi berikutnya akan ditenagai GPU AMD Instinct MI500 Series dan CPU AMD EPYC “Verano,” dengan jaringan Pensando “Como” dan “Monza” generasi berikutnya. Solusi rack-scale AMD Helios 600 akan menyusul, ditenagai GPU AMD Instinct MI600 Series, CPU EPYC “Ferrara,” serta jaringan Pensando “Palma” dan “Levanzo.”

Menskalakan AI di Seluruh Enterprise

Enterprise terkemuka menjalankan operasinya pada infrastruktur AMD, mulai dari cloud, hybrid, dan data center on-premises hingga armada PC berkemampuan AI. Teknologi AMD membantu pelanggan menskalakan dengan cepat dan mempercepat transformasi enterprise.

Dalam event Advancing AI, AT&T menunjukkan bagaimana perusahaan menerapkan enterprise AI yang fleksibel menggunakan teknologi AMD di lingkungan cloud, on-premises, dan air-gapped. AT&T juga menggunakan GPU AMD Instinct dan software ROCm untuk mendukung model OTel 2.0, sebuah model open-source yang dilatih secara khusus untuk telekomunikasi.

Dengan platform developer AMD Ryzen™ AI Halo, AMD menghadirkan performa, efisiensi, dan kesederhanaan yang membuat pengembangan AI lokal lebih mudah diakses. Lebih banyak platform AMD Ryzen AI Halo, yang ditenagai prosesor Ryzen™ AI Max PRO 400 Series, akan tersedia pada akhir tahun ini dari AMD dan mitra OEM.

Cisco dan AMD berkolaborasi menggabungkan mesin inferensi berkinerja tinggi AMD, termasuk sistem AMD Ryzen AI Halo, dengan kapabilitas jaringan, observability, dan keamanan Cisco, sehingga enterprise dapat menerapkan, mengelola tata kelola, dan mengelola agentic AI hybrid dan lokal dalam skala besar.

Memajukan Frontier Berikutnya dari Physical AI

Seiring AI meluas di seluruh cloud, enterprise, dan sistem lokal, frontier berikutnya adalah menghadirkan kecerdasan ke dalam mesin yang dapat memahami, bernalar, dan bertindak di dunia fisik. Berdasarkan rekam jejak panjang dalam robotika dengan AMD FPGA dan adaptive SoC, AMD memperkenalkan solusi AMD Kria™ AI, yang memperluas kapabilitas robotika perusahaan dari tubuh robot hingga otak robot. AMD secara unik menyatukan persepsi AI, penalaran, pengambilan keputusan agentic, dan kontrol dalam satu platform untuk menghadirkan performa yang dibutuhkan sistem robotika dunia nyata yang menuntut.

Portofolio ini mencakup system-on-module (SOM) AMD Kria AI baru, yang ditenagai prosesor AMD Ryzen AI Embedded X100 Series, serta AMD Kria AI Robotics Developer Platform, platform terintegrasi turnkey dan terbuka pertama untuk robotika otonom yang menggabungkan komputasi CPU, GPU, NPU, dan FPGA. Bersama ekosistem open software yang diperluas, solusi AMD Kria AI menghilangkan vendor lock-in dan membantu developer serta pelanggan mempercepat perjalanan dari prototipe menuju produksi bagi sistem physical AI generasi berikutnya.

Sumber Pendukung

• Ikuti AMD di Advancing AI 2026 (Press Kit)

• Saksikan tayangan ulang keynote AAI

Tentang AMD

AMD (NASDAQ: AMD) mendorong inovasi dalam komputasi berkinerja tinggi dan AI untuk menyelesaikan tantangan paling penting di dunia. Saat ini, teknologi AMD mendukung miliaran pengalaman di seluruh cloud dan infrastruktur AI, sistem embedded, AI PC, dan gaming. Dengan portofolio luas CPU, GPU, jaringan, dan software yang dioptimalkan untuk AI, AMD menghadirkan solusi AI full-stack yang menyediakan performa dan skalabilitas yang dibutuhkan untuk era baru komputasi cerdas. Pelajari lebih lanjut di www.amd.com.

CAUTIONARY STATEMENT This press release contains forward-looking statements concerning Advanced Micro Devices, Inc. (AMD) such as AI accelerating demand for the full range of AMD silicon, driving AMD’s total addressable market to ~$2 trillion in 2030; compute demand accelerating; the next phase of AI; the features, functionality, performance, availability, timing and expected benefits of AMD products, including, but not limited to, 6th Gen AMD EPYC™ CPUs, AMD Instinct™ MI400 Series GPUs, AMD Helios™ AI rackscale solutions, AMD Ryzen™ AI Embedded X100 processors and the AMD Kria™ AI SOM and Robotics Developer Platform; expected plans, benefits, scalability and deployments with AMD partners including Anthropic, OpenAI, Meta and Cerebras; AMD’s AI infrastructure product roadmaps through 2030; and AMD’s collaboration with AT&T and Cisco, which are made pursuant to the Safe Harbor provisions of the Private Securities Litigation Reform Act of 1995. Forward-looking statements are commonly identified by words such as "would," "may," "expects," "believes," "plans," "intends," "projects" and other terms with similar meaning. Investors are cautioned that the forward-looking statements in this press release are based on current beliefs, assumptions and expectations, speak only as of the date of this press release and involve risks and uncertainties that could cause actual results to differ materially from current expectations. Such statements are subject to certain known and unknown risks and uncertainties, many of which are difficult to predict and are generally beyond AMD's control, that could cause actual results and other future events to differ materially from those expressed in, or implied or projected by, the forward-looking information and statements. Material factors that could cause actual results to differ materially from current expectations include, without limitation, the following: impact of government actions and regulations such as export regulations, import tariffs, trade protection measures, and AMD General AMD General AMD General AMD General licensing requirements; competitive markets in which AMD’s products are sold; the cyclical nature of the semiconductor industry; market conditions of the industries in which AMD products are sold; AMD’s ability to introduce products on a timely basis with expected features and performance levels; loss of a significant customer; economic and market uncertainty; quarterly and seasonal sales patterns; AMD's ability to adequately protect its technology or other intellectual property; unfavorable currency exchange rate fluctuations; ability of third party manufacturers to manufacture AMD's products on a timely basis in sufficient quantities and using competitive technologies; availability of essential equipment, materials, components (such as memory supply), substrates or manufacturing processes; ability to achieve expected manufacturing yields for AMD’s products; AMD's ability to generate revenue from its semi-custom SoC products; potential security vulnerabilities; potential security incidents including IT outages, data loss, data breaches and cyberattacks; uncertainties involving the ordering and shipment of AMD’s products; AMD’s reliance on third-party intellectual property to design and introduce new products; AMD's reliance on third-party companies for design, manufacture and supply of motherboards, software, memory and other computer platform components; AMD's reliance on Microsoft and other software vendors' support to design and develop software to run on AMD’s products; AMD’s reliance on third-party distributors and add-in-board partners; impact of modification or interruption of AMD’s internal business processes and information systems; compatibility of AMD’s products with some or all industry-standard software and hardware; costs related to defective products; failure to maintain an efficient supply chain as customer demand changes; AMD's ability to rely on third party supply-chain logistics functions; AMD’s ability to effectively control sales of its products on the gray market; impact of climate change on AMD’s business; AMD’s ability to realize its deferred tax assets; potential tax liabilities; current and future claims and litigation; impact of environmental laws, conflict minerals related provisions and other laws or regulations; evolving expectations from governments, investors, customers and other stakeholders regarding corporate responsibility matters; issues related to the responsible use of AI; restrictions imposed by agreements governing AMD’s notes, the guarantees of Xilinx’s notes and the revolving credit agreement; AMD’s ability to satisfy financial obligations under guarantees, leases and other commercial commitments; impact of acquisitions, joint ventures and/or investments on AMD’s business and AMD’s ability to integrate acquired businesses; impact of any impairment of the combined company’s assets; political, legal and economic risks and natural disasters; future impairments of technology license purchases; AMD’s ability to attract and retain key employees; and AMD’s stock price volatility. Investors are urged to review in detail the risks and uncertainties in AMD’s Securities and Exchange Commission filings, including but not limited to AMD’s most recent reports on Forms 10-K and 10-Q.

1Based on AMD Performance Labs estimates as of July 2026, tokens-per-dollar performance was calculated using the Kimi K2 Thinking workload (32K input / 8K output) on an AMD Helios rackscale solution compared to an NVIDIA Vera Rubin NVL72 rack. Results reflect estimated aggregate throughput across low, medium, and high-interactivity operating points and hourly pricing projection of system GPUs based on market conditions. System configurations may vary by manufacturer and may produce different results. MI400-025 2EPYC-068 - The AMD EPYC server CPU portfolio spans the industry’s broadest ranges of data center deployments, from general purpose enterprise, cloud, telecom, SMB, and HPC systems to emerging AI environments including sandboxed agentic AI deployments and GPU head node servers. AMD EPYC 6th Generation platforms extend this breadth by uniquely combining high core and thread density of up to 512 threads, advanced memory bandwidth of up to 16 channels of 12.8 GT/s MRDIMM support, next-generation PCIe® Gen 6 connectivity, and select SKUs with boost frequencies up to 5 GHz. 3EPYC-025D: As of July 2026, 6th Gen EPYC 9996 has 256 cores and 512 threads with SMT enabled which is higher than any other publicly disclosed 1P CPU 49xx6-012: Based on estimated performance data for Nvidia, Intel®, and AMD EPYC™ Server Processors for Agentic AI, the AMD EPYC9996 provides the most agents per rack at a 100Kw power envelope per rack. Compared to the cores per rack of Nvidia Vera (88c) powered server racks: - The AMD EPYC 9996 (256C) provides 2.08x the cores (and threads with SMT) per rack - The AMD EPYC 9965 (192C) provides 1.86x the cores (and threads with SMT) per rack - The Intel Xeon 6980P (128C) provides 1.24x the cores (and threads with SMT) per rack Source: https://www.amd.com/content/dam/amd/en/documents/solutions/ai/methodology-description.pdf Agent counts are estimates derived from available CPU thread resources used as a proxy under a consistent theoretical workload. Actual agent capacity and throughput will vary based on workload, model, memory, software, orchestration, and system configuration. AMD General AMD General AMD General AMD General 59xx6-013: Comparison based on published Top-of-stack core counts and 1Ku pricing for estimated highest Agents / CPU $ and threads / CPU $ AMD EPYC™ 9006 (512 threads), AMD EPYC™ 9005 (384 threads), and Intel® Xeon® 6 (256 threads) SKUs as of 7/22/2026. Threads derived as 2 threads per core (SMT). Intel and Xeon are trademarks of Intel Corporation or its subsidiaries. Source: https://www.amd.com/content/dam/amd/en/documents/solutions/ai/methodology-description.pdf Agent counts are estimates derived from available CPU thread resources used as a proxy under a consistent theoretical workload. Actual agent capacity and throughput will vary based on workload, model, memory, software, orchestration, and system configuration. 69xx6-014: Comparison based on published Top-of-stack core counts and CPU W, Default CPU Power for 6th Gen EPYC, and TDPs for 5th Gen EPYC, Intel® Xeon®, and Nvidia Vera for estimated Highest Agents / CPU W across AMD EPYC™ 9956 (400W Default CPU Power), AMD EPYC™ 9965 (500W TDP), Nvidia Vera (450W TDP), ARM AGI (300W TDP), and Intel® Xeon® 6980P (500W TDP) powered servers as of 7/22/2026. 2 threads per core (SMT). 1 thread per core for ARM AGI. Starting with the 6th Gen AMD EPYC™ server processor family, AMD uses Default CPU Power to describe processor power consumption, succeeding AMD's historical TDP reference. Default CPU Power reflects total power consumed across the processor's compute and I/O dies for the stated performance target. Default CPU Power and TDP may both serve as processor power references for product comparison, platform planning, and performance-per-watt analysis. Agent counts are estimates derived from available CPU thread resources used as a proxy under a consistent theoretical workload. Actual agent capacity and throughput will vary based on workload, model, memory, software, orchestration, and system configuration. Intel Xeon TDP from ark.intel.com. Nvidia Vera TDP from https://developer.nvidia.com/blog/nvidia-vera-cpu-sets-a-new-standard-for-agentic-workloads in-ai-factories/. ARM AGI Specifications from https://www.arm.com/products/cloud-datacenter/arm-agi-cpu#Specifications 7MI400-020: Based on measurements and calculations by AMD Performance Labs in July 2026, for the AMD Instinct™ MI455X GPU to determine measured token throughput at high, medium and low interactivity points run on Deepseek V4 Flash with FP4 serving compared to AMD Instinct™ MI355X GPU. System manufacturers may vary configurations, yielding different results. 8MI350P-007: Based on AMD internal testing (July 2026), on a (1x) AMD Instinct MI350P GPU vs (1x) NVIDIA H200 NVL GPU on the Llama 3.3 70B Instruct (FP8) online serving output-throughput per dollar (tok/s/USD) comparison at ISL/OSL 1024/1024 across concurrency levels 1, 4, 8, 16, 32, 64, 128, 256, 512; median of 3 runs per point. MI350P based server internal AMD estimated pricing as $327,238.40 USD. RTX_PRO_6000 based server public list price reported on OEM website as $265,928.24 USD as of 7/16/2026. Stated results are the peak per-concurrency ratios: MI350P served via AIMS silogenai/aim-instinct-meta-llama-llama-3-3 70b-instruct:0.12.0-rc6; H200 NVL via NVIDIA NIM nvcr.io/nim/meta/llama-3.3-70b-instruct:2.0.6; RTX PRO 6000 via NVIDIA NIM nvcr.io/nim/meta/llama-3.3-70b-instruct:2.0.6. Configuration: 8x AMD Instinct MI350P PCIe Card (CDNA4, gfx950, 128 CUs, 144 GB HBM3E, SPX compute / NPS1), vBIOS 113-350P-01-1K1-000A, GPU driver 6.19.13-2353916.24.04, ROCm 7.14.0 (AMD-SMI 26.5.0); host 2P AMD EPYC 9455 (48-core), Dell PowerEdge XE7745, BIOS 1.7.6, microcode 0xb002162, SMT Enabled, Ubuntu 24.04.4 LTS, Linux 6.8.0-124-generic || NVIDIA RTX PRO 6000: 8x NVIDIA RTX PRO 6000 Blackwell Server Edition, vBIOS 98.02.8D.00.01, GPU driver 595.45.04, CUDA 13.2; host 2P AMD EPYC 9455 (48-core), Dell PowerEdge XE7745, BIOS 1.6.4, microcode 0xb00215a, SMT Enabled, Ubuntu 24.04.4 LTS, Linux 6.8.0-124-generic. Sever manufacturers may vary configurations, yielding different results. Results may vary due to factors including system configurations, software versions and BIOS settings.