C3 machine series A3 Confidential VMs with H100 GPUs

Confidential computing protects user prompts and AI/ML models during inference and facilitates secure model building collaboration

[{"selector":"#anim-65b4ddae-dc5c-4c33-9ed9-8291881f6c13 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-b71c9b97-f647-44d7-aadf-a624d8c0380f","keyframes":{"transform":["rotate(-540deg) scale(0.1)","none"],"opacity":[0,1]},"delay":0,"duration":1000,"fill":"both","iterations":1}]

At Google Cloud Next, Google introduced two Confidential Computing technologies to protect AI/ML workloads

[{"selector":"#anim-e55d8aac-db7d-47a9-b04b-6a6bab6498f1 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-87c06ec1-6e7a-495d-8e86-2d2b4d4085eb","keyframes":{"opacity":[1,1]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"both"}] [{"selector":"#anim-272fef4a-a0ae-429a-bd0a-53805591d3ff","keyframes":{"transform":["scale(3)","scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}]

HGX protected PCIe-enabled confidential VMs for Intel AMX and NVIDIA H100 Tensor Core GPUs

[{"selector":"#anim-031eb180-9115-458a-9877-94267b4455bd","keyframes":{"transform":["translate3d(118.56677%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-76d73e81-eec3-40e7-9133-44bb35d8a7e9","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-f823daa0-fe2b-4994-ac0d-be98eee33b42","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}] [{"selector":"#anim-bd841df8-f835-4fd3-b591-f38606a23620 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate3d(-34.18992708853111%, 0, 0) translate(-25%, 0%) scale(1.5)","translate3d(0%, 0, 0) translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"fill":"forwards"}]

NVIDIA with Google Cloud allow GPU Confidential Computing. The H100 is NVIDIA's first GPU with a Confidential Computing mode

[{"selector":"#anim-a1ab4fb6-fe9b-4e1d-8adb-15ee932f08a3 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate3d(34.18530341841403%, 0, 0)","translate3d(0%, 0, 0)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"both"}] [{"selector":"#anim-5925bf5d-461e-4791-bbbf-ad270b954a25","keyframes":{"transform":["translate3d(-122.14765%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-b62450d1-8faf-462f-8571-1d7a717218f4","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-6f8486bb-a93b-4cc1-b6d2-c9b670ab3f32","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}]

Powerful hardware-based security restricts Trusted Execution Environment (TEE) data and code execution to approved users

[{"selector":"#anim-e284cf49-83ed-404c-8d6e-baa8abd4327d [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-c3851c7f-5733-4101-a8a2-54e806a22b7f","keyframes":{"transform":["translate3d(-120.5298%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-5efc028c-f5d7-4574-a6e3-08898d5fc9be","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-49fad625-1b7a-4a63-b4d6-9c0f45ec40c0","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}]

computing hardware with Confidential VMs with NVIDIA H100 Tensor Core GPUs with HGX protected PCIe

[{"selector":"#anim-810cb559-6b2a-45ca-880b-f6d7f7dece3f [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-61ecff48-693d-4890-a8ad-733133c60d4c","keyframes":{"transform":["translate3d(123.40426%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-a11a66d0-c911-467c-b8ae-b498aa098b29","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-62d364cf-fbb6-4512-b4ae-6c6c91dfa92d","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}]

A3 Confidential VMs with NVIDIA H100 GPUs may mark private data and trade secrets in a TEE

[{"selector":"#anim-ad608235-a874-4193-8c0d-e1f7d7d3bf9f [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-757e7059-0f86-4959-b4e8-18523d68d647","keyframes":{"transform":["translate3d(-117.64707%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-18aebcff-dfb5-4146-af2c-cdb312496f5c","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-4d30f005-4333-4e00-b772-6a41aea38ed4","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}]

This ensures that any fine-tuning or additional AI/ML training is hidden from view outside the TEE

[{"selector":"#anim-370fe69b-b874-4718-b12b-5c54adff3b84","keyframes":{"transform":["rotate(-540deg) scale(0.1)","none"],"opacity":[0,1]},"delay":0,"duration":1000,"fill":"both","iterations":1}] [{"selector":"#anim-88f60cf8-3bed-421e-91bb-badcc5fb3091 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate3d(-37.075242640019674%, 0, 0) translate(-25%, 0%) scale(1.5)","translate3d(0%, 0, 0) translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"fill":"forwards"}]

All Confidential VMs on the C3 machine series have Intel AMX instruction sets by default to secure AI/ML workloads For more Details Govindhtech.com

[{"selector":"#anim-6e072cfb-5b04-47c8-bc27-ab5baf797754 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-831ffd35-e75d-43bc-8b19-0a43ef793762","keyframes":{"transform":["translate3d(-125.08591%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-77c5dbcf-9a8c-419e-a6b6-08c8f4592b4c","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"both"}] [{"selector":"#anim-026a8b36-7963-4e71-9ebf-7e171de150b3","keyframes":{"transform":["scale(0.15)","scale(1)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.4, 0.4, 0.0, 1)","fill":"forwards"}]