{"id":1478,"date":"2026-04-05T12:59:08","date_gmt":"2026-04-05T12:59:08","guid":{"rendered":"https:\/\/pckumar.in\/datacenters\/product\/nvidia-hgx-h100-8gpu8x-nvidia-h100-sxm\/"},"modified":"2026-05-14T07:25:38","modified_gmt":"2026-05-14T07:25:38","slug":"nvidia-hgx-h100-8gpu8x-nvidia-h100-sxm","status":"publish","type":"product","link":"https:\/\/pckumar.in\/datacenters\/product\/nvidia-hgx-h100-8gpu8x-nvidia-h100-sxm\/","title":{"rendered":"NVIDIA HGX H100- 8GPU(8x NVIDIA H100 SXM) India"},"content":{"rendered":"<p>[et_pb_section fb_built=&#8221;1&#8243; _builder_version=&#8221;4.16&#8243; custom_padding=&#8221;0px||||false|false&#8221; global_colors_info=&#8221;{}&#8221;][et_pb_row _builder_version=&#8221;4.16&#8243; background_size=&#8221;initial&#8221; background_position=&#8221;top_left&#8221; background_repeat=&#8221;repeat&#8221; width=&#8221;100%&#8221; custom_padding=&#8221;0px||0px||false|false&#8221; global_colors_info=&#8221;{}&#8221;][et_pb_column type=&#8221;4_4&#8243; _builder_version=&#8221;4.16&#8243; custom_padding=&#8221;|||&#8221; global_colors_info=&#8221;{}&#8221; custom_padding__hover=&#8221;|||&#8221;][et_pb_wc_breadcrumb _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_breadcrumb][et_pb_wc_cart_notice _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_cart_notice][\/et_pb_column][\/et_pb_row][et_pb_row column_structure=&#8221;1_2,1_2&#8243; _builder_version=&#8221;4.16&#8243; background_size=&#8221;initial&#8221; background_position=&#8221;top_left&#8221; background_repeat=&#8221;repeat&#8221; width=&#8221;100%&#8221; custom_padding=&#8221;0px||||false|false&#8221; global_colors_info=&#8221;{}&#8221;][et_pb_column type=&#8221;1_2&#8243; _builder_version=&#8221;4.16&#8243; custom_padding=&#8221;|||&#8221; global_colors_info=&#8221;{}&#8221; custom_padding__hover=&#8221;|||&#8221;][et_pb_wc_images _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_images][\/et_pb_column][et_pb_column type=&#8221;1_2&#8243; _builder_version=&#8221;4.16&#8243; custom_padding=&#8221;|||&#8221; global_colors_info=&#8221;{}&#8221; custom_padding__hover=&#8221;|||&#8221;][et_pb_wc_title _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_title][et_pb_wc_rating _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_rating][et_pb_wc_price _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_price][et_pb_wc_description _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_description][et_pb_wc_add_to_cart _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_add_to_cart][et_pb_wc_meta _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_meta][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=&#8221;4.16&#8243; background_size=&#8221;initial&#8221; background_position=&#8221;top_left&#8221; background_repeat=&#8221;repeat&#8221; width=&#8221;100%&#8221; global_colors_info=&#8221;{}&#8221;][et_pb_column type=&#8221;4_4&#8243; _builder_version=&#8221;4.16&#8243; custom_padding=&#8221;|||&#8221; global_colors_info=&#8221;{}&#8221; custom_padding__hover=&#8221;|||&#8221;][et_pb_wc_tabs _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;]\t\t\t\t\t[\/et_pb_wc_tabs][et_pb_wc_upsells _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_upsells][et_pb_wc_related_products _builder_version=&#8221;4.16&#8243; global_colors_info=&#8221;{}&#8221;][\/et_pb_wc_related_products][\/et_pb_column][\/et_pb_row][\/et_pb_section]<\/p>\n","protected":false},"excerpt":{"rendered":"<p>&nbsp;<\/p>\n<p>The NVIDIA HGX\u2122 H100 8-GPU baseboard is the gold standard for AI supercomputing clusters. It links eight H100 SXM GPUs using NVSwitch technology, allowing every GPU to talk to every other GPU at full NVLink speed (900 GB\/s).<\/p>\n<p>With a total of 640GB HBM3 memory, it is architected for massively parallel workloads like training Large Language Models (LLMs) and mixture-of-experts (MoE) models. This platform serves as the fundamental node for the world&#8217;s largest AI clouds and supercomputers, delivering the bisection bandwidth required to scale workloads across thousands of GPUs.<\/p>\n","protected":false},"featured_media":104,"comment_status":"open","ping_status":"closed","template":"","meta":{"_et_pb_use_builder":"on","_et_pb_old_content":"&nbsp;\r\n\r\n&nbsp;\r\n<table border=\"1\" cellspacing=\"0\" cellpadding=\"8\">\r\n<tbody>\r\n<tr>\r\n<th>Product Specifications<\/th>\r\n<th>H100 SXM<\/th>\r\n<th>H100 NVL<\/th>\r\n<\/tr>\r\n<tr>\r\n<td>FP64<\/td>\r\n<td>34 teraFLOPS<\/td>\r\n<td>30 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>FP64 Tensor Core<\/td>\r\n<td>67 teraFLOPS<\/td>\r\n<td>60 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>FP32<\/td>\r\n<td>67 teraFLOPS<\/td>\r\n<td>60 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>TF32 Tensor Core<\/td>\r\n<td>989 teraFLOPS<\/td>\r\n<td>835 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>BFLOAT16 Tensor Core<\/td>\r\n<td>1,979 teraFLOPS<\/td>\r\n<td>1,671 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>FP16 Tensor Core<\/td>\r\n<td>1,979 teraFLOPS<\/td>\r\n<td>1,671 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>FP8 Tensor Core<\/td>\r\n<td>3,958 teraFLOPS<\/td>\r\n<td>3,341 teraFLOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>INT8 Tensor Core<\/td>\r\n<td>3,958 TOPS<\/td>\r\n<td>3,341 TOPS<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>GPU Memory<\/td>\r\n<td>80 GB<\/td>\r\n<td>94 GB<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>GPU Memory Bandwidth<\/td>\r\n<td>3.35 TB\/s<\/td>\r\n<td>3.9 TB\/s<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Decoders<\/td>\r\n<td>7 NVDEC + 7 JPEG<\/td>\r\n<td>7 NVDEC + 7 JPEG<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Max TDP<\/td>\r\n<td>Up to 700 W (configurable)<\/td>\r\n<td>350\u2013400 W (configurable)<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Multi-Instance GPU<\/td>\r\n<td>Up to 7 MIGs @ 10 GB each<\/td>\r\n<td>Up to 7 MIGs @ 12 GB each<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Form Factor<\/td>\r\n<td>SXM<\/td>\r\n<td>PCIe, dual-slot air-cooled<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Interconnect<\/td>\r\n<td>NVIDIA NVLink: 900 GB\/s\r\nPCIe Gen5: 128 GB\/s<\/td>\r\n<td>NVIDIA NVLink: 600 GB\/s\r\nPCIe Gen5: 128 GB\/s<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>Server Options<\/td>\r\n<td>NVIDIA HGX H100 (4\/8 GPUs)\r\nNVIDIA DGX H100 (8 GPUs)<\/td>\r\n<td>Partner &amp; NVIDIA-Certified Systems (1\u20138 GPUs)<\/td>\r\n<\/tr>\r\n<tr>\r\n<td>NVIDIA AI Enterprise<\/td>\r\n<td>Add-on<\/td>\r\n<td>Included<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>","_et_gb_content_width":""},"product_brand":[50],"product_cat":[37],"product_tag":[],"class_list":["post-1478","product","type-product","status-publish","has-post-thumbnail","product_brand-nvidia","product_cat-nvidia","first","instock","sale","sold-individually","shipping-taxable","purchasable","product-type-simple"],"_links":{"self":[{"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/product\/1478","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/product"}],"about":[{"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/types\/product"}],"replies":[{"embeddable":true,"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/comments?post=1478"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/media\/104"}],"wp:attachment":[{"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/media?parent=1478"}],"wp:term":[{"taxonomy":"product_brand","embeddable":true,"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/product_brand?post=1478"},{"taxonomy":"product_cat","embeddable":true,"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/product_cat?post=1478"},{"taxonomy":"product_tag","embeddable":true,"href":"https:\/\/pckumar.in\/datacenters\/wp-json\/wp\/v2\/product_tag?post=1478"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}