ON Semiconductor CEO cites power density as next AI bottleneck
- ON Semiconductor CEO identifies power density as the next major AI bottleneck
- Company targets $11 billion revenue by 2030 with 12-14% CAGR
- AI data center revenue expected to grow >50% CAGR from 2026 to 2030
- J.P. Morgan sees potential for $10-$11 EPS by 2030 vs current $8 estimate
- New Vcore design win secured with revenue expected by end of 2026
*this image is generated using AI for illustrative purposes only.
ON Semiconductor Corporation (NASDAQ: ON) CEO Hassane El-Khoury identified power density as the next major bottleneck for AI infrastructure. He stated that the industry challenge is shifting from compute and memory to delivering more power in less space with greater efficiency and better thermal performance.
El-Khoury dismissed concerns of an impending slowdown in AI infrastructure spending, citing committed data center projects and rising compute requirements. The company reaffirmed its raised long-term revenue target of roughly $11 billion by 2030.
Strategic Growth And Addressable Market
The company outlined a total addressable market (TAM) of $213 billion for its core business. This figure excludes the roughly $30 billion opportunity associated with its pending acquisition of Synaptics Inc. Once the transaction closes, El-Khoury expects the combined addressable market to exceed $243 billion.
He argued that combining Synaptics with ON Semiconductor would create complementary capabilities, stating the companies together could produce value where "1 + 1" becomes "more than 2." Integration planning is progressing well, though the companies must continue operating independently until the deal closes.
Revenue And Margin Targets
ON Semiconductor increased its projected compound annual growth rate to 12% to 14%, up from the previous 10% to 12% range. Management expects AI data centers to drive much of this growth, with AI-related revenue potentially rising from about $500 million in 2026 to more than $2.5 billion by 2030.
J.P. Morgan analyst Harlan Sur highlighted AI data centers as one of the clearest examples of ON Semiconductor’s expansion strategy. The firm noted that management expects AI data center revenue to grow at a compound annual rate of more than 50% from 2026 through 2030. The company is targeting the entire "grid-to-rack-to-chip" power chain as AI systems demand more electricity in smaller spaces.
The company recently secured a new Vcore design win, with revenue expected to begin by the end of 2026.
The company provided detailed projections for its key business units through 2030:
| Segment | Projected Annual Growth | Key Driver |
|---|---|---|
| AI Data Centers | >100% (2026-2027) | Power architecture adoption |
| Automotive | ~9% | Electrification, autonomous driving |
| Industrial | ~10% | General industrial demand |
Financially, ON targets a 53% gross margin and a 38% operating margin under the 2030 model. Free cash flow margins are targeted at 30% to 35%. Capital expenditures are expected to remain at about 5% of revenue. CFO Thad Trent noted that recent price increases should support margins as the company offsets higher input costs.
El-Khoury characterized the 53% gross margin level as a milestone rather than an endpoint. He expects newly introduced technologies across automotive, AI data centers, and industrial applications to represent a larger share of revenue over time, improving the product mix. The company has already delivered step-function margin improvements as it ramps newer technologies and plans to reassess its margin opportunity once it reaches the 53% milestone.
What The Numbers Show
The shift in growth assumptions highlights a significant concentration on high-value power solutions. With semiconductor content per AI rack expected to jump from about $15,000 to more than $115,000, the company’s margin expansion targets rely heavily on this value-add increase rather than volume growth alone. The $8.7 billion invested over the past five years, including $4.3 billion in capital expenditures, underpins this strategic pivot toward higher-density power architectures.
J.P. Morgan said ON Semiconductor’s updated financial targets could support earnings of about $10 to $11 per share by 2030. That compares with the current Wall Street estimate of $8 per share cited in the report. The analyst also pointed to opportunities in automotive, industrial applications and physical AI. ON Semiconductor sees potential semiconductor content of up to about $900 per humanoid robot by 2030.
Technical And Analyst Outlook
Despite the stronger long-term outlook, ON Semiconductor shares rose 2.34% to $68.16 in trading. The stock remains in a technical repair phase, trading 6.8% below its 20-day simple moving average of $72.87 and 14.3% below its 200-day SMA of $79.21. Resistance sits near $74, while support is around $55.
Analysts maintain a Buy consensus with an average price forecast of $106.16. Recent analyst moves include:
- Stifel: Hold (Lowers forecast to $75.00)
- Mizuho: Outperform (Lowers forecast to $95.00)
- Rosenblatt: Neutral (Maintains forecast to $85.00)
Cantor Fitzgerald maintained a Neutral rating with a $90 price target, while Needham kept a Buy rating at $116. Wells Fargo maintained an Overweight rating but lowered its forecast to $95.
How might the integration of Synaptics impact ON Semiconductor's ability to meet its aggressive 53% gross margin target by 2030?
What specific thermal management technologies is ON Semiconductor developing to address the power density bottleneck in AI infrastructure?
Could the projected >100% growth in AI data center revenue from 2026-2027 create supply chain constraints for ON Semiconductor's new Vcore designs?


























