Marvell Acquires Celestial AI for $3.25B and XConn for $540M to Fuel AI Connectivity
Marvell Technology reported 41% year-over-year adjusted revenue growth and expanded its AI connectivity portfolio by acquiring Celestial AI for $3.25 billion and XConn for $540 million. The deals accelerate UALink and CXL product rollouts, potentially boosting data-center revenue while raising balance-sheet leverage by nearly $3.8 billion.
1. Strategic Acquisitions Boost AI Connectivity
Marvell Technology has completed two major acquisitions this quarter, acquiring Celestial AI for $3.25 billion and XConn for $540 million. These deals expand Marvell’s data‐center connectivity portfolio by integrating optical engines and high‐speed cable systems into its existing switch and PHY product lines. The combined technology set is expected to support next‐generation AI workloads by increasing bandwidth density and reducing power per bit transmitted, positioning Marvell as a leading supplier for hyperscale data centers and supercomputing environments.
2. Financial Performance and Growth Metrics
In the most recent fiscal quarter, Marvell reported adjusted revenue growth of 41% year-over-year, driven largely by strength in its data‐center segment. Operating margins expanded by 520 basis points sequentially, reflecting higher volume leverage and disciplined cost control. Adjusted earnings per share more than doubled compared with the prior year, underscoring the company’s ability to translate top‐line gains into bottom‐line profitability even as it invests heavily in R&D and strategic M&A.
3. UALink and CXL Technology Integration
Marvell’s UALink chip portfolio, designed for switchless connectivity within GPU clusters, is now being enhanced with Compute Express Link (CXL) support following the XConn acquisition. This integration allows memory pooling across multiple accelerators, reducing latency by up to 30% and improving memory utilization by as much as 40% in AI training workloads. Early customer trials with two leading cloud providers have demonstrated data throughput increases of 25% in multi‐node configurations.
4. Balance-Sheet Implications and Outlook
While the acquisitions are expected to lift revenue by approximately $700 million in the next four quarters, they also add $3.79 billion of debt to Marvell’s balance sheet. Management projects net leverage to peak at 2.8x EBITDA before debt paydown begins in fiscal 2027. Free cash flow generation of over $1 billion annually should enable rapid deleveraging, but investors will be closely watching leverage metrics and interest coverage over the next 12 to 18 months as AI investments continue to scale.