Earnings Preview | ON Semiconductor (ON.US) Q3 Results to be Announced Soon: AI Data Center Business, Gross Margin, and Q4 Guidance Are the Key Focus

According to Zhitong Finance APP, ON Semiconductor (ON.US) will announce its third quarter 2026 results after the US market close on Monday, November 2. The market currently expects ON Semiconductor’s Q3 revenue to increase by 10% year-on-year to $1.7 billion, with adjusted earnings per share growing by more than 39% year-on-year to $0.88.

In addition to the third-quarter results, the market is also focused on ON Semiconductor’s guidance for the fourth quarter. According to current analyst expectations, ON Semiconductor's Q4 revenue is projected to be about $1.74 billion, with adjusted earnings per share around $0.96.

As the current market begins to reassess the return on capital expenditure for artificial intelligence (AI), and doubts whether the massive investments by technology giants can be converted into actual income and cash flow, ON Semiconductor’s Q3 results will serve as an important window to observe whether demand for AI infrastructure can continue to flow upstream to semiconductor suppliers.

The key to this financial report is not only whether revenue and earnings per share can exceed expectations, but whether the AI data center power business can demonstrate real demand, sustained growth, and conversion into improved gross margin and cash flow. For investors, Q3 results provide demand verification, while Q4 guidance will determine whether such growth can continue.

1. Can the AI data center business maintain rapid growth?

Data show that ON Semiconductor’s second-quarter revenue grew about 9% year-on-year to $1.6 billion, slightly exceeding the analyst consensus of $1.59 billion; adjusted earnings per share increased about 40% year-on-year to $0.74, higher than the expected $0.71.

Segment data further confirm that growth is increasingly focused on the AI data center power semiconductor business. Power Solutions Group (PSG) revenue reached $829 million, up 18.7% year-on-year and 13% quarter-on-quarter, making up 51.7% of total revenue versus 47.5% a year ago, becoming the company's main source of income; Analog & Mixed Signal Group (AMG) revenue was about $545.7 million, down 1.8% year-on-year; Intelligent Sensing Group (ISG) revenue was $228.8 million, up 6.6% year-on-year.

ON Semiconductor’s robust Q2 results indicate that AI chip demand is spilling over from GPU, ASIC, and HBM to power semiconductors, analog control, and monitoring signal chains. This is not simply analog chip companies “piggybacking on the AI concept,” but rather a real physical need arising from the increase in compute density.

In the analog/power chip and power semiconductor/discrete device fields, the core underlying logic for strong demand fueled by the unprecedented AI wave is that AI racks are jumping from tens of kilowatts for traditional servers to over 100 kilowatts, moving toward 600 kilowatts or even 1 megawatt. Next-gen GPU/TPU/ASICs may each consume 2-4 kilowatts, traditional 48V/54V architectures will generate thousands of amps, and copper losses, heat, and busbar size will rise steeply. This is driving global AI data centers to shift to the ±400V or 800V DC architecture jointly promoted by ON Semiconductor and NVIDIA.

Before power enters the GPU from the grid, it goes through AC/DC rectification, PSU, BBU, high voltage distribution, 800V to 50V bus conversion, and then multiphase VRMs bring voltage down to below 1V for the GPU core. Each stage requires Si, SiC, or GaN power switches, gate drivers, digital controllers, power management ICs, and electronic fuses. ON Semiconductor, Texas Instruments, and Infineon all sell these “must-pass” components, so the higher the rack power and the more conversion stages, the greater the per-cabinet semiconductor value.

ON Semiconductor is able to provide EliteSiC MOSFETs/JFETs, high-and-low voltage silicon MOSFETs, GaN devices, hot swap smart fuses, multiphase controllers, power stages, and PoL regulators in this “power tree” from grid to GPU. This means each AI rack not only needs more chips, but also higher-priced, higher-voltage, and higher-efficiency power devices. The NVIDIA MGX platform is viewed as a core strategic entry for ON Semiconductor’s data center business. ON Semiconductor already supplies power FETs, multiphase power, SiC JFET and GaN solutions for existing MGX systems, serving NVIDIA as well as PSU, BBU, and future 800V power distribution board suppliers in the MGX ecosystem; standard server and rack designs on MGX also make it easier for certified power devices to be adopted by multiple OEMs and ODMs.

ON Semiconductor has defined AI data center business as its fastest growing segment, and expects related revenue to more than double by 2026. If management confirms continued strong growth in the AI data center business, along with improvements in orders, customer onboarding, and product shipments, investors may see this as evidence that AI capital expenditure is spreading throughout the wider semiconductor industry chain, not just focusing on GPU, memory, and advanced process chips.

It is noteworthy that order visibility for the AI business is especially important. If management confirms that customer demand is not just concentrated in the coming quarters, but has already formed sustainable platform design engagement and subsequent mass production opportunities, the market may upgrade its valuation logic for ON Semiconductor from short-term cyclical rebound to medium- and long-term growth.

Additionally, the recovery of traditional businesses may also be a point of interest for investors. Can the automotive segment continue to improve? When will the industrial business return to growth? The answers to these questions may test the quality of ON Semiconductor's earnings growth.

2. Gross margin and cash flow are also key

ON Semiconductor’s Q2 GAAP gross margin rose from 37.6% to 38.4%, and non-GAAP gross margin increased to 39.3%. The company previously stated that Q2 capacity utilization had increased from 77% in Q1 to 83%, and management estimates that each percentage point increase in utilization can contribute about 25 to 30 basis points to gross margin.

As of July 3, ON Semiconductor’s operating cash flow grew 149.4% year-on-year to $459.7 million, and free cash flow jumped from $106.1 million to $425.4 million, about four times the figure a year earlier, with the free cash flow rate rising from about 7% to 27%. In other words, while Q2 revenue only grew about 9%, profit, cash flow, and shareholder return all expanded several-fold, indicating that ON Semiconductor's Fab Right capacity optimization, cost discipline, and product mix upgrades are having a real operational leverage effect.

If Q3 gross margin reaches or exceeds the upper guidance limit of 42%, and free cash flow continues to improve, it would suggest that the growth of the AI data center business, improved capacity utilization, and product mix optimization are driving profitability. Conversely, if revenue exceeds expectations but gross margin remains stagnant—or is dragged down by higher raw material, substrate, and external manufacturing costs—the market may question whether AI business growth truly brings high-quality earnings.

Moreover, ON Semiconductor is advancing its acquisition of Synaptics. After the deal is complete, the company will expand its presence in connected computing, edge AI, and related technologies. However, investors will also focus on deal financing, integration costs, synergies, and gross margin impact. Synaptics itself has higher gross margins and theoretically could improve ON Semiconductor's product mix, but whether the acquisition can actually enhance long-term profitability still depends on post-merger business integration and revenue synergy, rather than just Synaptics’ historical gross margin.

3. Will AI orders squeeze out traditional business supply capacity?

ON Semiconductor previously gave priority to AI customer orders, causing delivery schedules for automotive and industrial clients to be somewhat impacted. Additionally, product delivery cycles reportedly extended from about 27 weeks to about 32 weeks.

This indicates that AI power semiconductor demand is increasing, but also creates allocation issues in supply. Investors should pay attention to whether management expects further delivery extension and whether the company can maintain normal delivery to automotive and industrial clients while meeting AI demand. If AI growth comes at the expense of other businesses, the overall growth quality will be discounted.

4. Q4 guidance will determine whether investors are willing to keep paying a premium for AI growth

Compared to the Q3 results themselves, Q4 guidance may have a more direct impact on ON Semiconductor's share price. According to current analyst forecasts, ON Semiconductor’s Q4 revenue will be about $1.74 billion, with adjusted EPS around $0.96, which means the market has already priced in continued growth for the AI data center business, improved capacity utilization, and margin expansion to some extent.

Therefore, even if Q3 revenue and EPS beat expectations, if management’s Q4 guidance only meets consensus, or if their commentary on AI business outlook becomes more cautious, the stock price may not see a significant boost.

In summary, the impact of this earnings report on market sentiment may not be limited to ON Semiconductor’s own stock price, but will also provide important insight for investors judging the prosperity of the AI infrastructure supply chain. The most important metric to watch is whether AI data center business growth can simultaneously drive revenue, gross margin, and free cash flow improvements.

If all three rise in tandem, ON Semiconductor is likely to become a beneficiary of the diffusion of AI investments from computing chips to power infrastructure, and may provide a positive validation of demand for the AI supply chain as the market questions the sustainability of AI capital expenditures. But if the AI business grows rapidly while overall gross margin and cash flow do not significantly improve, the market may further distinguish true profit beneficiaries from mere demand growers within the AI supply chain.