NXP Semiconductors: Data Center Inflection and SDV Ramp Make This More Than an Auto Story
TD Cowen 54th Annual Technology Conference, May 27, 2026
NXP Semiconductors walked into TD Cowen's annual technology conference with a message that goes well beyond its familiar automotive identity. Speaking with analyst Josh Buchalter, Jeff Palmer, EVP of Investor Relations, laid out a company that is quietly building a credible data center revenue stream, accelerating its software-defined vehicle bet, and entering what management believes is the early innings of a genuine upcycle. The combination is more interesting than the stock's traditional narrative suggests.
The Cycle Has Turned, and NXP Is Not Being Shy About It
Palmer opened with a tone shift that is worth noting. "Compared to 90, 180 days ago, I think we're more optimistic than we have been in a while," he said, pointing to a set of KPIs that collectively signal tightening conditions: book-to-bill solidly above 1, customer backlog through distribution building, late orders and expedites increasing, and lead times stretching out. NXP has already raised prices in the first quarter and is preparing further increases into the second half, primarily driven by energy, transportation, precious metals, and substrates. Palmer was clear that the intent is margin preservation, not margin expansion — "all we're looking to do is maintain our gross margins" — and that any inflationary pass-throughs are first tested against the company's own operational capacity to absorb them.
Internal factory utilization is guided to run in the low 80% range in the first half and the mid-80s in the second half, with four non-fungible 8-inch facilities running proprietary mixed-signal processes. One clarification Palmer offered should update some investor models: automotive is not over-indexed to internal manufacturing. All bulk CMOS products, including automotive, are built externally through foundry partners. Internal factories are reserved for analog, mixed-signal, and RF products. That distinction matters for how investors should think about the utilization leverage story as automotive demand recovers.
Automotive: The Inventory Overhang Is Finally Behind Them
The nine-quarter Tier 1 inventory digestion that weighed on NXP's Western markets business is over. "That's finally behind us," Palmer said. But the recovery is not a restocking story — it is a return to end-demand buying, and in some cases, Tier 1s are running dangerously lean. "We have a few other large Tier 1s who are running 3 and 4 weeks," Palmer noted, explaining that squeezed margins and limited OEM compensation are driving that behavior. NXP is not expecting a restocking wave and says customers who fail to provide adequate forecasts risk line-down situations, given the company's three-to-six-month fab-to-finished-product cycle time.
On China specifically, Palmer pushed back on the bear case. NXP's China automotive business was actually up in the first quarter, consistent with the seasonal pattern the company had anticipated, and Auto is guided up again in the second quarter with China participating. The broader automotive thesis, he argued, should be understood as a content-per-vehicle story rather than a SAAR story. NXP's auto revenue has grown at a 9% CAGR over three years and 13% over five, against essentially flat global unit production.
The Four Growth Vectors, and Why Software-Defined Vehicles Are the Biggest
Palmer broke NXP's automotive growth into a core franchise growing at low single digits and four accelerated growth drivers that represent the real investment case. The largest is the software-defined vehicle platform built around the S32 MPU family, zonal processors, Automotive Ethernet, and software. That business was $1 billion in 2024, crossed $1 billion again in 2025, and is targeted at roughly $2 billion by the end of 2027. Critically, Palmer emphasized that this is not speculative: "In that horizon, that is not design wins we don't have. We have the design wins. We just have to wait for customers to get ready to go to production."
The 77-gigahertz radar business, currently approaching $900 million annually, is expected to grow at 15% to 20% as the product line transitions toward imaging radar. Electric vehicle content — battery management systems and gate drivers — was approximately $500 million in 2024 and carries a similar 15% to 20% CAGR outlook, recovering from a program setback last year. Connectivity, spanning in-cabin WiFi and Bluetooth as well as ultra-wideband for phone and vehicle integration, rounds out the four-driver framework.
Seven years ago, NXP made a deliberate decision not to invest in another generation of automotive microcontrollers, instead betting on a hierarchical processing architecture for the vehicle of the future. The S32N product that has emerged from that bet is, in Palmer's words, "effectively a virtual ECU" — a device with 16 independent ECUs built into a single die, dynamically reprogrammable by the automaker. Western OEMs are expected to begin rolling out software-defined vehicle platforms in the 2028 model year, arriving in late 2027. The bulk of current SDV revenue is being driven by Chinese and Korean OEMs. Palmer was candid about the risk: "SDV is probably more like the opportunities are bigger... but if you don't win them, if they decide to go in a different direction, it can be big as well, to the negative."
Two Acquisitions Aimed at Owning the Software Stack
To reinforce its SDV position, NXP made two acquisitions Palmer discussed in detail. TTTech Auto, an Austrian firm with 1,200 automotive software engineers specializing in functional safety and security, brought a middleware operating system called MotionWise. The rationale was straightforward: "It was really a make versus buy. We knew we were going to need more software engineers going forward than we currently had." The team is now working to enable software on top of the S32 families in addition to continuing development of MotionWise.
The second deal, Aviva Links, brings a multi-gigabit asymmetric connectivity technology called Certus that is suited for ADAS applications — where sensor data flowing upstream exceeds the data pushed downstream — as well as in-cabin display applications. The acquisition was nudged along by existing customers, and NXP already has design wins, though revenue is not expected until next year at the earliest.
The Data Center Business Is the Part of the Story Investors Have Been Missing
NXP formally called out its data center revenue for the first time on its most recent earnings call: $200 million in 2025, scaling to $500 million in 2026. Palmer walked through the two components in detail, and the explanation is worth understanding for investors who have not been tracking this.
The first is the Layerscape-based product line — a 16-nanometer chip with sixteen 64-bit ARM cores and a large switch fabric — which goes into top-of-rack switches and, in an 8-core variant, into network interface cards. This sits within NXP's digital networking segment. The revenue acceleration is tied to a small number of hyperscalers building proprietary AI rack infrastructure. NXP won these designs years ago; the ramp simply took longer than expected to materialize. "We didn't talk about it until recently because we've been awarded the design win several years ago. They just weren't going anywhere," Palmer said. "It wasn't until late last year that they started to accelerate."
The second component is board management control — power management, cooling, security, and inter-card communication functions across server line cards. This sits in the IoT segment and draws on NXP's i.MX application processors, MCX RoT security microcontrollers, and a portfolio of I2C and I3C connectivity products. The customer base here is broader, spanning hyperscalers, Taiwanese server OEMs, and U.S. reference design partners. Palmer acknowledged the cross-segment placement creates investor confusion and suggested the company is considering how to restructure its reporting segments.
NXP's stated serviceable addressable market for the control plane of data centers is $4 billion, growing at roughly 10% annually. Management targets growth at a multiple of the SAM, with 2x the SAM growth rate as the ideal. At $500 million on a $4 billion SAM, the penetration story has significant runway if execution holds.
Edge AI Is Starting to Show Up in Industrial Orders
On the industrial and IoT side, Palmer described a shift in how customers are specifying processors. Distilled AI models running locally without cloud dependency are moving from concept to procurement criteria. NXP's i.MX families already carry embedded NPUs under the EIQ brand for moderate-performance inference workloads. The higher-performance answer is the Kinara acquisition, which added the Ara NPU family — a chip that can be ganged up to three units per i.MX instance. Palmer framed this as genuinely early-stage but real: "We're really kind of amazed with the ideas customers are coming up with."
The Gross Margin Path to 60%-Plus and What VSMC Adds in 2028
NXP operates within a 57% to 63% long-term gross margin target and is currently mid-range. The company's rule of thumb — 100 basis points of gross margin expansion for every $1 billion of incremental revenue — combined with management's stated confidence in low-double-digit revenue growth in both 2026 and 2027 implies a trajectory toward approximately 60% gross margins as the company approaches $15 billion to $15.5 billion in revenue. Palmer cautioned against applying the rule of thumb on a quarterly basis, noting that mix shifts create variability.
The more significant catalyst arrives in 2028, when the VSMC joint venture in Singapore — a partnership with TSMC and GlobalFoundries that has been in development for three to four years — reaches full load. At that point, NXP expects an incremental 200 basis points of gross margin expansion at the corporate level. If the company exits 2027 near 60%, VSMC's contribution would put gross margins comfortably into the upper half of the target range for the first time.
Capital Return Story Should Re-Emerge as Investment Cycle Peaks
NXP has returned approximately $23 billion to shareholders over the past decade, representing roughly 95% to 96% of all free cash flow generated. The past few years have been an exception, with cash directed toward three acquisitions and the Singapore joint venture. Palmer indicated those demands are now tailing off, with 2026 and 2027 expected to mark a return to 100% free cash flow distribution through dividends and buybacks. For investors who have found NXP's capital return profile underappreciated relative to peers, the company's own message is that the window is reopening.