HP has put prices and an expected arrival date on its two RTX Spark-powered OmniBooks, turning September’s introduction into a purchasing timetable. The Windows PCs are expected at HP.com on October 16, 2026, starting at $2,999.99 for the OmniBook X 14 and $3,199.99 for the OmniBook Ultra 16.
Two prices, different retail routes
The October 7 announcement follows HP’s introduction of the portfolio at IFA in September. The new information is commercial: when HP expects the machines to become available, what each starts at, and which retailers will carry them. It is not the first introduction of the OmniBook lineup.
The X 14 has the lower entry price; the Ultra 16 starts $200 higher. Both figures are starting prices, not a claim that every configuration costs that amount. HP also cautions that pricing and availability can change without notice, so October 16 remains an expected availability date rather than a delivery guarantee.
OmniBook starting prices
01$2,999.99OmniBook X 14
HP expects availability at HP.com on October 16, 2026. Additional expected sellers include Costco.com, Micro Center, Walmart and other U.S. retailers.
02$3,199.99OmniBook Ultra 16
HP expects availability at HP.com on October 16, 2026. The company also says Best Buy stores and BestBuy.com will carry the machine.
A shared local-AI pitch, not a measured comparison
HP describes RTX Spark as combining NVIDIA AI, RTX graphics and unified memory in a Windows PC. Its stated goal is to run larger AI models locally without depending on the cloud. That puts execution on the computer itself, rather than requiring a remote service for those local workloads.
The company positions the portfolio for AI creation, development, intelligent agents and entertainment, including creative work and gaming. This is a shared capability pitch for the lineup, not a published head-to-head performance comparison between the X 14 and Ultra 16.
HP’s announcement does not identify a specific AI model tested on either machine or provide a measured running speed. Buyers now have a price and retail route for each device, but the claim about running larger models remains a design promise, not a workload result they can compare.
Reader comments
Newest comments first. Replies stay oldest first.