Supermicro raises quarterly revenue estimate by $4 billion on AI demand

Super Micro Computer, the US maker of servers used in artificial intelligence data centres, said on Tuesday, October 20, 2026 that it now expects revenue of $18.5 billion to $19.5 billion for the first quarter of its fiscal year 2027, far above its earlier guidance of $14.5 billion to $15.5 billion. The company, known as Supermicro, said the increase came mainly from demand for its AI platforms. Its shares rose by about 10% to 11% in after-hours trading on the news, according to market reports.

Supermicro builds the physical machines that run AI: racks of servers packed with graphics processing units (GPUs), the chips that do most of the work in training and running AI models, along with cooling and networking. When technology companies and cloud providers spend more on AI data centres, companies like Supermicro are among the first to see the orders. A jump of this size in a single quarter's estimate is a direct sign that spending on AI hardware was still accelerating in the July to September period.

For businesses, this is the industrial side of the AI boom. Every AI service a company buys, from chatbots to document processing, runs on hardware like this somewhere. Strong orders mean more computing capacity is being built, which over time can make AI services more widely available.

The gross margin figure is worth reading closely. Server makers like Supermicro earn relatively thin margins compared with chip designers, because much of the price of a server goes to the components inside it. A margin of 12.0% to 12.5% means that out of every $100 of sales, about $12 is left after the direct cost of making the product. The company attributed the improvement from the earlier range of 10.4% to 10.8% mainly to the mix of customers and products it sold.

The same evening, another AI infrastructure business, SpaceX's AI computing unit, also reported revenue well above analysts' expectations. Together, the two updates point to continued heavy investment in AI computing at a time when some investors have questioned whether that spending will pay off.

For Indian companies and IT service firms that sell AI projects to global clients, the numbers are a sign that the hardware base for those projects is still growing quickly.

Short-term: The estimate is likely to lift Supermicro shares when regular trading opens on October 21, though the after-hours move can change. Investors will look for confirmation in the full results on November 3.

Long-term: If demand stays at this level, server makers, chip companies and power and cooling suppliers will keep expanding capacity. The thin margins of the server business mean that execution and supply of parts remain key risks.

Who is affected: Supermicro shareholders, cloud providers and companies building AI data centres, chip suppliers whose products go into the servers, and businesses that depend on growing AI computing capacity.

Questions and Answers