Microsoft’s recently launched Maia 200 AI inference accelerator is now performing at a level “more comparable to its competitors,” according to a report from Goldman Sachs. The assessment comes after previous limited data on Microsoft’s earlier Maia 100 chip suggested the company lagged behind rivals.
Goldman Sachs analyst Gabriela Borges reiterated a “Buy” rating and a $600.00 price target for Microsoft (NASDAQ:MSFT) stock following the Maia 200 update on January 26, 2026. Prior to the Maia 200 announcement, Borges noted that available benchmarks for the Maia 100 were scarce and industry feedback indicated Microsoft was behind competitors in AI chip performance.
The analyst now believes Maia 200’s performance is “more comparable to the competition in terms of raw compute performance,” a development she described as “exceptionally positive for the price/performance of Microsoft’s AI compute services.” This improvement also supports Microsoft’s long-term strategy of achieving gross margins on AI compute in Azure that are comparable to CPU-based Azure workloads, Borges added.
“We believe diversification of Microsoft’s silicon footprint will be key to Microsoft achieving better gross margins and ROI in AI compute,” Borges wrote in a research note. The Maia 200 is built on TSMC’s 3nm process and features native FP8/FP4 tensor cores, a redesigned memory system with 216GB HBM3e at 7 TB/s and 272MB of on-chip SRAM, and data movement engines designed to maximize model utilization.
Microsoft claims the Maia 200 delivers three times the FP4 performance of Amazon’s third-generation Trainium chip and FP8 performance exceeding Google’s seventh-generation TPU. The company also states that Maia 200 is its most efficient inference system to date, offering 30% better performance per dollar than its current generation hardware.
The Maia 200 is currently deployed in Microsoft’s US Central datacenter region near Des Moines, Iowa, with the US West 3 datacenter region near Phoenix, Arizona, scheduled to come online next. Microsoft is also previewing the Maia SDK, which includes tools for building and optimizing models for the Maia 200, including PyTorch integration, a Triton compiler, and optimized kernel library.
Microsoft shares were trading at $388.73 at the time of publication on Monday, February 23, 2026, down 2.14%, according to Benzinga Pro data. The launch of Maia 200 followed Microsoft’s earlier decision to delay the delivery of its latest-generation AI chip from 2025 to 2026, as the company refocused its AI chip plans.
The Microsoft Superintelligence team will utilize Maia 200 for synthetic data generation and reinforcement learning to enhance next-generation in-house models. The company highlighted the chip’s unique design as accelerating the generation and filtering of high-quality, domain-specific data for downstream training.
