📢 𝗝𝗨𝗦𝗧 𝗜𝗡: Google Cloud and NVIDIA Launch Vera Rubin A5X AI Instances, Target 960,000 GPUs - $Alphabet(GOOGL.US) $NVIDIA(NVDA.US)

👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:

➤ 𝗚𝗼𝗼𝗴𝗹𝗲 𝗖𝗹𝗼𝘂𝗱 and 𝗡𝗩𝗜𝗗𝗜𝗔 launch generally available 𝗔𝟱𝗫 𝗔𝗜 𝗶𝗻𝘀𝘁𝗮𝗻𝗰𝗲𝘀.

➤ A5X instances are powered by 𝗡𝗩𝗜𝗗𝗜𝗔 𝗩𝗲𝗿𝗮 𝗥𝘂𝗯𝗶𝗻 𝗡𝗩𝗟𝟳𝟮 systems.

➤ Google Cloud plans infrastructure expansion to 𝟵𝟲𝟬,𝟬𝟬𝟬 𝗥𝘂𝗯𝗶𝗻 𝗚𝗣𝗨𝘀.

➤ European availability is planned for 𝗡𝗼𝘃𝗲𝗺𝗯𝗲𝗿 𝟮𝟬𝟮𝟲.

➤ A5X integrates 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗫-𝟵 𝗦𝘂𝗽𝗲𝗿𝗡𝗜𝗖𝘀 and Google's 𝗩𝗶𝗿𝗴𝗼 𝗡𝗲𝘁𝘄𝗼𝗿𝗸.

➤ 𝗜𝗻𝗲𝗳𝗳𝗮𝗯𝗹𝗲 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 becomes the first A5X customer.

➤ 𝗪𝗣𝗣 demonstrates AI workloads running on a single A5X instance.

➤ Google Cloud expands 𝗡𝗩𝗜𝗗𝗜𝗔 𝗖𝗼𝗻𝗳𝗶𝗱𝗲𝗻𝘁𝗶𝗮𝗹 𝗖𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴 across GPU infrastructure.

➤ 𝗚𝟰 𝘃𝗶𝗿𝘁𝘂𝗮𝗹 𝗺𝗮𝗰𝗵𝗶𝗻𝗲𝘀 gain generally available Confidential Computing capabilities.

➤ Companies explore 𝗗𝗦𝗫 𝗠𝗮𝘅𝗟𝗣𝗦 to optimize AI performance per watt.

👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:

➤ Strengthens 𝗡𝗩𝗜𝗗𝗜𝗔's position in next-generation hyperscale AI infrastructure.

➤ Planned 𝟵𝟲𝟬,𝟬𝟬𝟬-𝗚𝗣𝗨 deployment signals substantial long-term AI infrastructure ambitions.

➤ Expands 𝗚𝗼𝗼𝗴𝗹𝗲 𝗖𝗹𝗼𝘂𝗱's capacity for advanced AI training and inference.

➤ Single-instance workload consolidation could reduce 𝗔𝗜 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 complexity and costs.

➤ Confidential Computing could accelerate 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗜 adoption for sensitive workloads.

👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁:

𝗗𝗮𝘃𝗶𝗱 𝗦𝗶𝗹𝘃𝗲𝗿, Founder and CEO of Ineffable Intelligence:

"We're building superlearners that continually improve through their own experience," said David Silver, founder and CEO of Ineffable Intelligence. "Vera Rubin's combination of compute performance and high-bandwidth connectivity allow us to push the scale and ambition of our reinforcement learning research further."