Nutanix has acquired Ryax Technologies, a French compute-orchestration specialist, and plans to fold its GPU scheduling into Nutanix Kubernetes Platform and Nutanix Enterprise AI. The Ryax acquisition is the company’s first since it bought D2iQ’s assets in December 2023. Both prices stayed undisclosed; both deals share a habit of buying small and quiet.
The announcement’s language is big — “agentic AI anywhere”, “transformative technology” — while the target is not. Ryax is a Lyon startup of three to five people that raised roughly €2.3 million across its entire life. The gap between those two facts is the story.
Why the Ryax Acquisition News Arrived Now
Three threads explain the timing.
A month after a strong year
Nutanix reported fiscal 2026 results on 26 August: revenue of $2.85 billion, up 12%, and ARR of $2.55 billion, up 16%. Free cash flow reached $840.7 million, and the balance sheet held $2.36 billion in cash. A Rule-of-40 score of 42 marked the third straight year above 40.
That performance buys patience, and cash for tuck-ins. It also leaves a gap in the narrative. Nothing in the reported numbers says “AI platform company”, the self-description Nutanix now leads with. Four weeks after earnings, this deal supplies the missing beat.
The agentic AI land grab
April 2026 brought Nutanix’s “complete platform for the agentic AI era” — NKP Metal for bare-metal Kubernetes, new neocloud partnerships, expanded AI services. June brought NVIDIA certification for its storage in AI factories. September needed a follow-on, and the release itself names the pressure: GPU shortages are pushing enterprises to split compute across private data centres, hyperscalers and neoclouds.
Scheduling across that fragmented footprint is a real problem. Nutanix just bought itself an answer — small, French and cheap.
A rare Nutanix acquisition
Nutanix rarely buys companies. Its previous purchase was D2iQ’s Kubernetes assets in December 2023, lifted from a business liquidating through an insolvency process. That code became Nutanix Kubernetes Platform, the very product line Ryax will now enhance.
The pattern is consistent: distressed or tiny targets, no disclosed prices, technology folded into the platform. Nutanix competes against Broadcom’s vastly larger balance sheet, so it buys components, not companies.
The Competitive Picture Behind the Ryax Acquisition
The category this deal enters already has a giant in it.
The Run:ai comparison
NVIDIA bought Run:ai, the Israeli GPU-orchestration leader, for a reported $700 million, closing in December 2024 after an EU antitrust review. It then open-sourced the core scheduler — KAI Scheduler — under Apache 2.0 in April 2025. Those two moves set the category’s reference price and, simultaneously, made a free version public.
Ryax sits in the same category with a fraction of one percent of Run:ai’s funding. Nutanix is not trying to outspend NVIDIA. It wants scheduling that spans NVIDIA GPUs and everything else, tied to a hybrid-cloud platform rather than one vendor’s stack.
What the others offer
IBM pairs watsonx with Red Hat OpenShift for cross-environment AI workloads. Broadcom’s VMware, remarkably, has no equivalent play in this space. Any Kubernetes team can already pull free schedulers — KAI Scheduler, Volcano, Kueue — off the shelf.
So the competitive question the release never addresses: why pay Nutanix for what the ecosystem gives away? The honest answer is integration and support, not raw capability.
What the Data Shows About the Ryax Acquisition
The numbers on both sides of the deal deserve attention.
The company Nutanix bought
Ryax Technologies started in Lyon in November 2017, founded by three computer-science PhDs whose prior work managed resources for supercomputers. Its first product, launched alongside a €1.34 million seed round in 2019, aimed to industrialise data-science pipelines. The company then pivoted to an open-source “hybrid IT workflow orchestrator”, and later to AI compute orchestration with a hosted cloud that rents H100s by the second.
French registry filings show three to five employees in 2022 and 2021 revenue under €700,000. Lifetime funding comes to roughly €2.2–2.4 million, including a final €400,000 round in 2023. Blocks & Files estimates the price at two to four times that funding — call it under $15 million. Nutanix earns that back in about an hour of revenue.
Nutanix by the numbers
The buyer enters this deal healthy. ARR grew 16% to $2.55 billion in fiscal 2026, net retention held at 106%, and the company added more than 3,000 customers. It claims over half the Global 2000 as customers. Server supply constraints, though, are stretching delivery timelines and pushing license start dates into the future — a headwind management says will persist through fiscal 2027, when guidance calls for $3.18–3.23 billion in revenue.
One footnote matters for headline readers: fiscal 2026 GAAP net income of $1.51 billion includes a one-time $1.21 billion tax benefit.
The deal math
“Financially not material” is the release’s phrase, and the arithmetic backs it. If three to five engineers and an open-source engine command even a generous multiple, the Ryax acquisition is an acqui-hire wearing strategy clothing. That is a rational move — Nutanix needs the capability, and the price cannot hurt. But the milestone framing should be read accordingly.
What’s New vs. What’s Repackaged
New: cross-platform agentic AI orchestration — private data centres, hyperscalers, neoclouds, even Slurm HPC clusters — embedded in a hybrid-cloud platform. Also new, and genuinely useful for Nutanix: a European engineering beachhead with HPC DNA.
Repackaged: GPU scheduling itself. Kubernetes-native schedulers are abundant, and NVIDIA’s is free. Ryax’s own marketing has recycled the same Gartner line — 85% of data-science projects never reach production — since 2019, updating only the buzzword, from “smart data” to “agentic AI”.
The Question the Press Release Doesn’t Answer
What happens to Ryax’s existing customers? The company ran a hosted cloud, listed on AWS Marketplace, and maintained an open-source engine. The release says nothing about any of them. Nor does it give an integration timeline beyond “future releases”, or state how many engineers actually join. And Nutanix’s Global 2000 claim arrives with no disclosure of AI revenue — so the acquisition’s success will be invisible in the numbers the company does report.
What the Ryax Acquisition Means for You
Different readers should take different lessons from this deal.
If you run Nutanix today, nothing changes this quarter. Ask your account team when Ryax-derived scheduling lands in NKP and NAI, and request GPU-utilisation benchmarks against your current setup. That is the only test that matters.
If you are choosing GPU orchestration, note that NVIDIA’s KAI Scheduler is free under Apache 2.0, and Run:ai ships inside the NVIDIA stack. Ryax’s exit shows where independent options end up: absorbed. Design your stack so the scheduler can be swapped without rewriting applications.
If you follow the hybrid-cloud market, watch the D2iQ playbook repeat. Asset purchases, quiet terms, platform integration — that is how Nutanix fights Broadcom’s scale. The question for fiscal 2027 is whether agentic AI moves ARR, or remains a press-release adjective.
Editor’s Note
This article draws on Nutanix‘s 22 September 2026 press release and its 22 September Indian agency distribution; Nutanix’s fiscal 2026 fourth-quarter results and earnings commentary of 26 August 2026; French registry data and funding records for Ryax Technologies from 2017 to 2023; Blocks & Files’ analysis of the acquisition; CRN’s interview with Nutanix’s chief AI officer; reporting on NVIDIA’s Run:ai acquisition by Reuters and TechCrunch; and Nutanix’s D2iQ announcements. The purchase price remains undisclosed, and estimates cited here are third-party calculations. Nothing here is investment advice.