onsemi Forecasts Upbeat Revenue as AI Data Center Chip Demand Surges

onsemi issued an upbeat revenue forecast after AI data center sales accelerated, giving the company a prominent place in Google News despite uneven demand elsewhere. The result matters because onsemi does not make the graphics processors that train large AI models. It supplies power semiconductors that help deliver electricity efficiently across increasingly dense computing systems.

That distinction creates the central tension. AI spending is expanding beyond Nvidia and other accelerator vendors, pulling less visible component suppliers into the growth cycle. Yet onsemi still depends heavily on automotive and industrial customers, whose recoveries remain less certain than hyperscaler infrastructure budgets.

The company’s outlook therefore tests a broader claim about the AI market. If power demand becomes a durable revenue engine, onsemi can reduce its dependence on traditional cycles. If the lift remains small beside its core businesses, the latest forecast will look more like a welcome offset than a strategic transformation.

The forecast shows AI demand spreading through the server rack

onsemi’s latest outlook suggests that the AI infrastructure boom is reaching components far beyond processors and memory.

The company forecast quarterly revenue above market expectations after reporting stronger demand for chips used in AI data centers. The upbeat forecast, reported by Reuters, followed the company’s August 3 earnings call.

The quarter ended July 3, according to onsemi’s earnings schedule. That timing makes the outlook a fresh reading on infrastructure spending during the middle of 2026.

The key development is not simply that another semiconductor company offered favorable guidance. The important change is where onsemi says the momentum originates. AI data centers are becoming a meaningful source of incremental demand for its power-management portfolio.

A modern AI server requires much more than an accelerator. Electricity moves through several conversion stages before reaching processors, memory, networking equipment, storage devices, and cooling systems. Each stage creates opportunities for power-management chips, controllers, and switching components.

This chain is often called the power tree. It describes the hardware that converts, distributes, monitors, and protects electricity from the facility connection to individual computing devices. Greater computing density usually increases the importance of efficient conversion at every level.

onsemi had already reported evidence of this shift. In its first-quarter results, the company said AI data center revenue more than doubled from the previous year. It attributed that increase to broader adoption across the power tree.

The company also said its AI data center business grew more than 30 percent sequentially during that earlier quarter. Those figures establish momentum before the new forecast, rather than relying on a single period’s guidance.

First-quarter companywide revenue reached $1.513 billion. That result exceeded the midpoint of onsemi’s guidance and rose 5 percent from the comparable period. The Power Solutions Group produced $736.6 million, up 14 percent year over year.

Those numbers help explain why investors care about the latest report. Power products were already outperforming several other parts of the portfolio. AI infrastructure can strengthen that position if deployments continue at the current pace.

The signal also matches developments across the wider chip market. Texas Instruments has connected stronger analog-chip expectations to AI data center demand. GlobalFoundries has cited accelerating deployments, while Tower Semiconductor has reported large commitments involving data-center connectivity chips.

These companies do not compete with Nvidia in the most direct sense. They sell supporting components or manufacturing capacity needed around high-performance processors. Their results show how infrastructure spending can travel through several semiconductor categories.

Google News coverage can flatten those distinctions by placing every supplier under the broad label of “AI chips.” That shorthand is convenient, but it can mislead readers about what each company provides.

onsemi’s opportunity sits in electrical conversion and control, not model training performance. Its products help an expensive accelerator receive stable power with less energy lost as heat. That role becomes more valuable as racks consume more electricity.

The latest forecast consequently changes the onsemi story in one specific way. AI data centers are no longer only a distant market attached to a presentation. They are influencing reported growth and near-term management expectations.

Still, the development does not erase the company’s traditional exposure. Automotive and industrial customers remain central to overall revenue. The next question is whether AI power chips can become large enough to change that balance.

Google News attention hides the real contest for onsemi

The main contest is not onsemi against Nvidia, but AI infrastructure growth against onsemi’s dependence on cyclical automotive and industrial demand.

That framing matters because the company’s favorable forecast arrived during a recovery that remains uneven. AI-related demand is rising quickly, while several established semiconductor markets continue working through inventory and spending cycles.

onsemi has historically generated a large share of its business from automotive applications. Its chips support electric-vehicle power systems, driver assistance, image sensing, charging, and vehicle networking. Industrial equipment provides another major source of demand.

Both markets can produce attractive long-term growth. They can also move slowly when customers reduce inventories, delay capital purchases, or respond to weaker economic conditions. A strong AI data center quarter does not automatically remove those pressures.

First-quarter segment results showed the imbalance. Power Solutions Group revenue rose from the previous quarter and year. Advanced Solutions Group revenue declined 3 percent sequentially and 5 percent year over year.

Intelligent Sensing Group revenue fell 5 percent sequentially, although it remained slightly above the prior-year period. Companywide revenue decreased 1 percent from the fourth quarter. AI momentum existed alongside that mixed performance.

This contrast is why the latest guidance deserves scrutiny. A rising forecast says demand improved enough to lift the whole company’s near-term range. It does not show that every end market recovered at the same rate.

Management described the first quarter as a move beyond the cyclical trough. That is a company assessment, not an independently established turning point. The next several quarters must show whether automotive and industrial orders support that view.

The AI contribution can still change investor expectations before it becomes the largest business. Markets often respond to the direction and durability of incremental revenue. A fast-growing category can affect valuation even while mature categories provide most sales.

However, readers should separate three claims that often merge in headlines. The first is that AI data center revenue is growing quickly. The second is that this growth materially lifts companywide revenue. The third is that it permanently changes onsemi’s business mix.

Available company disclosures strongly support the first claim. The new forecast provides some support for the second. The third remains an unproven strategic outcome.

That distinction also applies to the label “surging demand.” Growth rates can appear large when measured from a modest base. onsemi disclosed percentage increases, but its first-quarter release did not provide a separate dollar total for AI data center revenue.

The company’s 2025 annual report offered useful context. It said AI data center revenue exceeded $250 million for that year. That amount established the business as real, although still limited beside annual companywide revenue.

If that base continues doubling, AI power sales can become more consequential quickly. If growth normalizes after early design adoption, the business will remain a valuable supplement. The difference will not be visible from one optimistic forecast.

The pressure therefore falls on onsemi’s core businesses as much as its competitors. Strong AI growth raises the standard for automotive and industrial execution. Investors will expect those markets to recover without consuming the gains produced by data centers.

Texas Instruments creates a relevant comparison because it also sells analog and power-management components. Its broad customer base gives it exposure across industrial, automotive, consumer, and communications markets.

A Reuters report on analog chip demand described AI data centers as a driver of Texas Instruments’ favorable outlook. That parallel shows onsemi does not own the supporting-chip opportunity.

GlobalFoundries and Tower Semiconductor offer another comparison. Their manufacturing and specialty-chip businesses serve different parts of the infrastructure chain. Both have also pointed to data-center deployments when explaining stronger expectations.

This wider participation is positive for the infrastructure thesis. It is less comfortable for any company claiming a uniquely protected position. Customers can source related power and analog products from multiple vendors, depending on performance requirements and design qualifications.

onsemi must therefore convert early demand into durable design wins. A design win places a supplier’s component inside a customer platform, often creating revenue across that platform’s production life. Repeated wins matter more than temporary component shortages.

The latest Google News cycle captures the forecast, but not that longer contest. The headline establishes momentum. Product adoption, customer breadth, and companywide mix will decide whether momentum becomes structural growth.

Power efficiency is becoming part of AI performance

AI data center power chips matter because electrical efficiency now constrains how much computing equipment operators can deploy.

AI performance discussions usually focus on processor speed, memory bandwidth, or networking capacity. Those metrics remain essential. Yet every accelerator must operate inside a facility with finite electrical and cooling capacity.

Power conversion always loses some energy. The lost energy becomes heat, which cooling systems must remove. Small efficiency gains can matter when multiplied across thousands of processors operating continuously.

This relationship gives power-semiconductor suppliers a clear technical route into AI infrastructure. They can provide devices that switch electricity efficiently, tolerate higher voltages, occupy less space, and respond quickly to changing computing loads.

The market includes traditional silicon components and newer materials such as silicon carbide and gallium nitride. These wide-bandgap materials can operate under demanding voltage, temperature, and switching conditions. They can also introduce higher costs or manufacturing challenges.

onsemi has invested heavily in silicon carbide, commonly shortened to SiC. The material has been associated closely with electric vehicles because efficient inverters can improve range and charging performance. Similar efficiency priorities apply to data-center power systems.

However, the products used in vehicles and servers are not interchangeable simply because they share a material. Voltage levels, reliability requirements, switching frequencies, packaging, and customer qualification processes differ. Each application requires its own engineering work.

onsemi’s data-center pitch spans more than one component. The company says it is gaining adoption across the power tree with multiple chip vendors and leading hyperscalers. Hyperscalers are companies that operate computing infrastructure at enormous scale.

That customer language sounds encouraging, but the company has not publicly identified every buyer behind the reported growth. It also has not supplied enough detail to calculate concentration or compare revenue by platform.

The mechanism nevertheless fits a visible industry problem. AI accelerators have become more demanding, and rack-level power density keeps increasing. Operators must deliver more electricity while controlling heat, space, reliability, and operating costs.

A supplier does not need to build the accelerator to benefit from that transition. It needs products qualified for the electrical architecture surrounding the accelerator. That expands the potential group of AI beneficiaries.

Omdia has forecast that annual power-chip revenue will rise from roughly $80 billion in 2026 to $100 billion by 2029. The forecast, discussed in an analysis of power-chip demand, covers far more than data centers.

The broad market creates both opportunity and caution. AI infrastructure can accelerate growth within power semiconductors, but vehicles, industrial equipment, electronics, and communications remain important demand sources. Suppliers must manage capacity across those uses.

Data-center operators also have several ways to improve energy efficiency. They can change facility voltages, redesign power shelves, use liquid cooling, optimize workloads, or select more efficient processors. Power chips represent one part of a system-level response.

That system view favors vendors able to work with several equipment makers. A component validated across different server and power architectures can reach more deployments. It also reduces reliance on a single accelerator generation.

onsemi says its growth reflects adoption with multiple chip vendors and hyperscalers. If accurate, that breadth would make the opportunity more durable than one short product cycle. Future disclosures need to show whether the breadth continues.

The company’s manufacturing strategy also matters. onsemi has been adjusting its production footprint while focusing resources on differentiated products. It announced agreements in July to divest two manufacturing facilities under its “fab right” strategy.

Such changes can improve utilization and capital efficiency. They can also introduce execution risk when demand rises faster than expected. Capacity, outsourced production, inventories, and qualification schedules must remain aligned.

A proposed European investment offers another piece of the supply strategy. In 2024, onsemi discussed a multiyear investment of up to $2 billion in the Czech Republic. The planned facility would support advanced power semiconductors.

That project was presented as serving electric vehicles, renewable energy, and AI data centers. Its breadth illustrates the company’s central manufacturing challenge. The same strategic capacity must address several markets with different cycles.

Power efficiency also changes how readers should interpret the phrase “AI chip demand.” The category is broader than processors with a model-training benchmark. It includes components that make dense computing systems practical and economical.

For enterprise buyers, the implication reaches beyond semiconductor stocks. AI deployment costs depend on infrastructure availability, electricity, cooling, and utilization. Faster accelerators cannot eliminate those constraints by themselves.

Developers rarely select rack-level power components directly. They still feel the consequences through cloud availability, workload costs, and capacity limits. Hardware efficiency can influence how quickly providers add computing resources.

This is the mechanism behind onsemi’s favorable forecast. More AI computing creates more electrical complexity. More complexity increases semiconductor content throughout the power path, giving specialized suppliers an opening beside accelerator leaders.

The numbers still leave important questions unanswered

The upbeat outlook supports onsemi’s AI narrative, but public disclosures do not yet prove that data centers can reshape the entire company.

The largest uncertainty is scale. onsemi has reported fast percentage growth and identified AI data centers as an important opportunity. It has not consistently broken that business into a separate reporting segment.

Without quarterly dollar revenue, readers cannot determine how much of the companywide improvement came directly from AI power products. They also cannot measure margins, customer concentration, or the mix between new designs and expanding existing programs.

This limitation does not make the growth claim false. It makes its strategic importance harder to test. Percentage growth, absolute revenue, and profitability answer different questions.

The next uncertainty concerns durability. Hyperscalers are spending heavily on AI infrastructure, but capital budgets do not rise at the same rate forever. Customers can revise deployment schedules because of electricity access, permits, financing, or changing model economics.

Power components can also experience ordering volatility. Customers may build inventory when supplies appear constrained, then reduce purchases after capacity improves. Reported semiconductor revenue can temporarily run ahead of end demand.

The industry learned that lesson after pandemic-era shortages. Customers accumulated components to protect production, followed by a prolonged inventory correction in several markets. Automotive and industrial suppliers were particularly exposed to that adjustment.

AI data centers differ from vehicles, but they are not immune to supply-chain behavior. Rapid architecture changes can shorten product relevance. Buyers may also redesign power systems as rack voltages and cooling methods evolve.

The competitive landscape creates another risk. Texas Instruments, Infineon, STMicroelectronics, Analog Devices, Monolithic Power Systems, and other suppliers participate in power management. Product overlap varies, but customers have alternatives across many functions.

Competition does not depend only on headline efficiency. Buyers evaluate price, availability, reliability, packaging, software support, reference designs, and manufacturing continuity. A technically attractive device can lose if the broader supply plan looks weak.

onsemi’s acquisition strategy adds another execution question. In June, it agreed to acquire Synaptics in an all-stock transaction with an enterprise value of approximately $7 billion. The companies expect the deal to broaden onsemi’s reach into intelligent systems.

The transaction would combine onsemi’s power and sensing portfolio with Synaptics’ processing and connectivity products. According to the transaction filing, the combined company was projected to have $7.8 billion in 2026 revenue.

The presentation also described $200 million in expected annual synergies within 18 months after closing. The transaction was expected to close in mid-2027, subject to shareholder and regulatory approvals.

Those are forward-looking company estimates, not completed results. Integration can consume management attention while onsemi adjusts factories and pursues data-center designs. The acquisition could strengthen the portfolio, but it also raises the execution burden.

The deal additionally broadens onsemi’s AI story beyond data centers. Synaptics sells technology for connected devices and edge computing, where workloads run closer to users or machines. That expansion can diversify growth but complicate performance comparisons.

Investors should avoid combining every AI-adjacent product into one undifferentiated category. Data-center power, edge processing, vehicle sensing, and wireless connectivity serve different buyers. They carry different margins, product cycles, and competitive pressures.

The company’s financial presentation requires similar care. In the first quarter, GAAP operating margin was negative 3.5 percent, while non-GAAP operating margin reached 19.1 percent. Restructuring and asset-related charges drove much of that difference.

Non-GAAP measures can help show underlying operations, but they exclude selected costs. Readers should examine both sets of figures as onsemi changes its manufacturing footprint and prepares for a major acquisition.

Gross margin also deserves attention. First-quarter GAAP gross margin was 38.5 percent, compared with 40 percent a year earlier. Sequential improvement was encouraging, yet the annual comparison showed that recovery was incomplete.

AI revenue becomes more convincing when it raises companywide growth without weakening margins. Revenue gained through unfavorable pricing or expensive capacity expansion would carry less strategic value.

There is also a verification gap between company language and public customer evidence. onsemi says it works with multiple chip vendors and leading hyperscalers. Named deployments would help establish which architectures use its products and for how long.

Suppliers often cannot identify customers because of confidentiality agreements. Even so, later financial results can reveal whether demand remains broad. Sustained growth across several quarters would reduce the importance of missing names.

The skeptical reading is therefore straightforward. AI data centers are growing quickly for onsemi, but the company has not yet shown that this business outweighs weakness elsewhere. The forecast improves the case without settling it.

That is a more useful conclusion than either extreme. The report is not merely an “AI” label attached to an old semiconductor company. It is also not proof that onsemi has escaped automotive and industrial cycles.

Three signals will show whether the forecast marks a lasting shift

The next test is whether onsemi turns rapid AI growth into visible scale, stronger margins, and repeatable customer adoption.

The first signal is a clearer quarterly AI data center revenue figure. Management has provided percentage growth and an annual reference point. A regular dollar disclosure would let readers measure the business against companywide sales.

If AI revenue keeps doubling from a larger base, the structural-growth argument strengthens. If percentage growth slows before the business becomes substantial, the latest guidance will look more cyclical.

Readers should also watch whether onsemi repeats its claim of broad adoption across the power tree. Additional design wins, named platforms, or expanding content per system would show that demand extends beyond one component.

A decline in customer breadth would weaken the thesis. One major program can produce fast growth, but concentration makes revenue more vulnerable to architecture changes or customer negotiations.

The second signal is gross-margin performance. AI power products should help margins if customers value efficiency, reliability, and specialized design. They should also support operating leverage when factories run at healthier utilization.

Margin improvement would indicate that AI demand contributes more than volume. Flat or declining margins during rapid AI growth would raise questions about pricing, product mix, or manufacturing costs.

This measure must be read alongside restructuring adjustments. Large exclusions can obscure how much the operating transition actually costs. GAAP and non-GAAP results should move toward a more consistent picture over time.

Automotive and industrial demand remains part of this test. A recovery in those markets would let AI growth lift the entire company. Continued weakness could absorb the new business and limit reported expansion.

The third signal is execution across manufacturing changes and the Synaptics transaction. onsemi is trying to optimize factories, expand AI data-center power sales, and prepare for a large acquisition. Each task requires capital and management attention.

Successful factory divestitures should improve utilization without limiting supply for growing products. Stable delivery performance would support the company’s strategy. Shortages or rising costs would weaken it.

The September 16 analyst day provides an early opportunity for greater detail. Investors should look for a defined AI data center revenue path, product categories, addressable content, and capital requirements. General statements about market growth will offer less value.

The company should also explain how Synaptics changes the data-center strategy. The acquisition was presented partly as an expansion from AI data centers into physical AI. Readers need to know where the portfolios genuinely connect.

Physical AI refers to systems that use computing, sensors, and controls to interact with the physical environment. Robots, vehicles, and industrial machines fit that description. It is distinct from power delivery inside a data center.

A credible integration plan should preserve those distinctions. Otherwise, the combined AI narrative risks becoming too broad to measure. Clear reporting will matter as much as a wider product portfolio.

Competitor results provide an external check. Continued data-center strength at Texas Instruments, GlobalFoundries, and specialty foundries would support the broad infrastructure thesis. Diverging results might reveal company-specific wins rather than a universal trend.

Hyperscaler capital spending is another background indicator, although it should not replace onsemi’s own results. Large budgets support demand across processors, networking, memory, cooling, and power. They do not guarantee equal gains for every supplier.

The most informative evidence will come from the intersection of those signals. Rising customer breadth, expanding dollar revenue, and improving margins would establish a durable shift. One measure alone would remain open to alternative explanations.

For readers arriving through Google News, the immediate answer is clear. onsemi’s forecast adds credible evidence that AI spending is spreading beyond accelerator vendors. Power delivery has become part of the infrastructure race.

The harder judgment remains unresolved. onsemi must show that AI growth can change its overall financial profile while its established markets recover. That proof will require several quarters, not one favorable headline.

Watch the next revenue disclosure first, then margins, and finally execution around factories and Synaptics. Together, those measures will show whether the current Google News moment marks a new growth engine or a temporary cushion.