Microsoft to scale custom AI chips in multi-gigawatt push to cut costs Luke Juricic Mon, October 12, 2026 at 10:33 AM EDT 2 min read MSFT +1.33% Trade MSFT on Coinbase Trading disclosure Trading disclosure The above button links to Coinbase. Yahoo Finance is not a broker-dealer or investment adviser and does not offer securities or cryptocurrencies for sale or facilitate trading. Coinbase pays us for certain activity generated through this link. Prices displayed are informational.
Investing.com -- Microsoft Corporation (NASDAQ:MSFT) plans to deploy custom-designed artificial intelligence processors across its global datacenter footprint at a multi-gigawatt scale starting in 2027, marking a aggressive push to curb reliance on third-party hardware vendors and lower operating costs. The technology giant confirmed Monday that its upcoming Azure Maia 300 chip will serve as the primary inference engine powering internal AI applications, including GitHub Copilot, Microsoft Foundry, and its suite of Copilot enterprise tools.
Before you buy NVIDIA: our AI model's stock picks are beating the market by 120%+. See the list now »
The announcement comes as cloud infrastructure providers face escalating expenses and power constraints tied to hosting resource-intensive AI models. By shifting major production workloads to its proprietary silicon architecture, Microsoft aims to significantly optimize datacenter economics while insulating its operations from ongoing supply shortages for graphics processing units.
The Maia 300 represents the third generation of Microsoft's custom silicon program, following the active rollout of its Maia 200 processor. The Maia 200 is currently handling live production tasks across North America, Europe, and Australia, delivering up to 1.7 times higher output per watt for Microsoft Teams and Copilot workloads compared to standard commercial GPU setups.
Not finding the right stocks? Our AI's list of investor-grade stock picks for October just dropped »
The upcoming Maia 300 chip features a multi-chiplet architecture built on TSMC's N2P process node and incorporates eight HBM4 memory stacks. Microsoft projects the system will deliver a 40% improvement in energy efficiency per megawatt relative to competing hardware, along with a sevenfold increase in rack-scale token throughput over its current-generation system.
The dual-chip roadmap underscores a broader strategic pivot among major hyper-scalers toward vertically integrated hardware designs. Google, Amazon, and Meta Platforms have similarly accelerated proprietary chip programs to handle specialized AI calculations, challenging Nvidia Corp.'s historic dominance over the datacenter accelerator market.
For investors, the multi-gigawatt commitment reflects a clear trajectory toward disciplined capital expenditure in AI infrastructure. By engineering silicon tailored specifically for long-context reasoning and multi-step autonomous agents, Microsoft is positioning itself to expand service capacity while expanding margins across its Azure cloud platform.
Microsoft to scale custom AI chips in multi-gigawatt push to cut costs
Google expands wearables lineup with $180 Fitbit Edge to boost digital health push
C3.ai rallies after Northland upgrades stock on enterprise AI platform launch
