Helping companies and data centers become more cost and energy-efficient.
Bamboo's mission is to become the world's first true net-server compute platform combining AI and hardware. 14,000 companies have committed to net-zero carbon emissions by 2030. The CIOs of these companies are struggling to meet this goal.
A material piece of the problem is the $80 billion on-premise server market. According to IDC, Arm servers are predicted to overtake x86 and become a $100 billion market by 2030 due to the superior cost/power characteristics.
Arm-designed CPUs are driving a paradigm shift in data centers delivering significantly cheaper, faster, and efficient capabilities. According to IDC, Arm is beginning to outpace x86 moving forward. We have already seen several major players make the transition:
December 2019 -- AWS Garaviton2 enables the best price-performance in EC2
June 2020 -- ARM-based Fugaku supercomputer is not the fastest in the world
July 2020 -- Apple announces Mac transition to all Apple Arm-based silicon
September 2020 -- NVIDIA to acquire Arm for $40 billion creating the world's premier computing company for AI
October 2020 -- VMware announces ESXi bare-metal hypervisor on Arm
December 2020 -- Microsoft commits to powering 50% of Azure Cloud with Arm servers
Arm-designed CPUs are beneficial to large parallel workloads from Kubernetes, TensorFlow, MySQL, MariaDB, Ceph, Docker, MongoDB, Elasticsearch, Apache, and AWS. It's also supporting Linux distros Ubuntu 20.04, Centos 8.4, and Red Hat 8.4.
Why do enterprise customers care?
When you consider the rack space, power consumption, CapEx, OpEx, heat output, power consumption, and CO2 output of a 16-server x86 deployment versus a Bamboo box, enterprise data centers realize a 75 to 80% savings. This is the type of orders of magnitude reduction in emissions attracting CIOs and COOs charged with meeting their company's carbon-neutrality goals.
The Arm ecosystem is growing as more applications are being built for Arm with app modernization, edge computing, IoT devices, AI/ML, and platform-as-a-service. Arm is prolific at the edge. The number of container images has gone from 5,000 to 125,000 in under five years. Developers are now Arm-first due to the lower cost. They're also able to recompile legacy apps on Arm in 90 minutes or less.
Arm servers also have linear scale-out performance with sustained, predictable core frequencies and predictable performance at scale delivering excellent performance at a lower cost per watt -- a key metric for data centers today.
There are two keys driving the adoption of Arm:
Desire to be net-zero by 2030.
The inability to get more power into data centers to drive and cool servers.
As companies throw more compute to solve business problems, they need a way to power compute more efficiently.