Tickers: $NXP | Related: $TXN, $ON, $STM, $MCHP
If you'd asked any of us a year ago which sleepy auto-chip company was going to print the best single trading day in its history off the back of a data center line item, nobody on the desk would've picked NXP. This is the company most people slot mentally next to "Texas Instruments, but for cars." Boring. Cyclical. Useful but not exciting.
Then yesterday happened. The stock closed up 18%. That's the largest one-day move in NXP's history as a public company. And here's the thing your group chat probably got wrong: it wasn't really about the auto recovery, even though that's the headline most outlets ran with.
The real story is buried four lines down in the earnings deck. NXP did $200 million in data center revenue last year. CEO Rafael Sotomayor just told the Street that number is going to $500 million-plus in 2026 — a 150% jump in a segment most analysts weren't even modeling. That's the line that broke the stock open.
So the question isn't "did NXP have a good quarter." It did. The real question is whether a company nobody calls an AI stock just quietly turned itself into one — and whether the multiple you can still buy it at today is going to look obvious-in-hindsight by Christmas.
Let's get into the numbers.
What Actually Happened in Q1 2026
NXP reported EPS of $3.05 against a $2.95 consensus, on revenue of $3.18 billion — up 12% year over year and the best top-line growth this company has put up since the post-COVID auto restock cycle in 2022. Gross margin held at 58.4%. Operating margin expanded 130 basis points sequentially.
The segment breakdown is where it gets interesting:
- Automotive: $1.78 billion (+14% YoY). The auto cycle isn't just back, it's accelerating, and ADAS / electrification content per vehicle is doing the heavy lifting.
- Industrial & IoT: $620 million (+18% YoY). This is where the data center contribution lives, and it's why the segment outpaced auto.
- Mobile: $410 million (+5% YoY). Steady.
- Communications & Other: $370 million (+3% YoY). Lagging, as expected.
Free cash flow came in at $890 million for the quarter, and management raised the buyback authorization by $2 billion. The dividend is up 8%. This is what a healthy semi cycle looks like when capacity finally lines up with demand.
But none of those numbers, on their own, justify an 18% move. To get to 18%, you need a re-rating. And to get a re-rating, you need a story.
The Data Center Line Nobody Was Modeling
Here's what most of the wire copy missed. On the call, Sotomayor spent more time on data center than on auto. That's never happened before at NXP.
The pitch is straightforward: hyperscalers building out AI training and inference clusters need a lot of non-compute silicon — power management ICs, isolated gate drivers, high-speed interface chips, secure microcontrollers for board management. NXP has quietly been winning sockets in this space for two years, mostly through its analog and mixed-signal portfolio inherited from the Freescale combination.
The numbers Sotomayor put on the slide:
- 2024 data center revenue: ~$120 million
- 2025 data center revenue: $200 million (+67% YoY)
- 2026 guide: $500 million-plus (+150% YoY)
- Multi-year target: $1B+ by 2028
If they hit the 2026 number, data center goes from a rounding error to roughly 4% of total revenue — and more importantly, to the fastest-growing line in the company by a country mile. That's the kind of mix shift that re-rates a multiple.
For context, when analog peer Monolithic Power (MPWR) crossed the $500M data center threshold in 2023, its forward P/E expanded from ~28x to ~45x in twelve months. NXP currently trades at ~17x forward — a multiple that prices it as a pure auto cyclical with zero AI optionality.
Why the Sell-Side Just Sprinted to Update Models
The PT revisions came out fast and steep:
- TD Cowen: $250 → $310 (Buy)
- Morgan Stanley: $299 → $335 (Overweight)
- Goldman: $280 → $320 (Buy)
- Citi: $265 → $305 (Buy)
The new Street median sits around $315, with a high of $345. Our own work lands at $330, which I'll walk through in the scenario section below.
What's notable is why the upgrades came in. Read the TD Cowen note carefully and the language is telling: "Data center exposure now warrants partial re-rating toward analog peers with comparable hyperscaler revenue mix." Translation: NXP is no longer being valued as a pure auto play.
That's the multiple expansion thesis in one sentence.
The Auto Story Isn't Bad — It's Just Not the Story
To be clear: the auto business is genuinely strong. Content per vehicle is up to ~$140 in 2026 from ~$95 in 2021, driven by ADAS sensor fusion, battery management ICs for EVs, and the slow-but-real shift to zonal architectures. NXP, NXP's S32 processor family is winning design wins at most major OEMs.
But auto is a known story. It's already in models. What's not in models is data center — and that's the variance, the optionality, the thing that actually moves the stock.
If you're an investor sizing this position today, your decision tree looks like this:
1. Do I believe Sotomayor's $500M+ data center guide? (I do — channel checks support it.)
2. Do I believe the multi-year $1B+ trajectory? (Probability-weighted yes.)
3. If the answer to both is yes, what multiple am I willing to pay?
That last one is the whole game.
Competitive Context: Where NXP Fits in the Analog Landscape
The cleanest peer set is TXN, ON Semi, STM, and Microchip. Here's how they stack up on forward P/E:
- TXN: ~24x (premium; pure-play analog leader, no auto cyclicality concern)
- ON Semi: ~18x (auto-heavy, similar profile to NXP)
- STM: ~14x (cheaper but Europe / weak end-markets)
- Microchip: ~22x (recovering from inventory cycle)
- NXP: ~17x (current)
If NXP simply re-rates to ON Semi's multiple — same profile, but with the data center kicker priced in — you get to $290 mechanically. To get to our $330 PT, you need NXP to trade somewhere between ON and Microchip, which is exactly what the data center pivot earns it.
Three-Scenario Price Target
This is where I break it out for sizing. All targets are 12-month.
Bull Case — $385 (probability: 25%)
Data center hits $550M+ in 2026, beats the guide. 2027 visibility extends to $800M run-rate. Auto stays strong. Multiple expands to 22x forward EPS of ~$17.50. Re-rating completes within 12 months.
Base Case — $330 (probability: 50%)
Data center hits the guide ($500M). Auto grows mid-teens. EPS lands at $16.80. Multiple expands modestly to 19.5x as Street partially re-rates. This is our central case and the rating anchor.
Bear Case — $245 (probability: 25%)
Data center disappoints — hits $400M, not $500M. Hyperscaler capex digestion in 2H 2026 hits the segment. Auto cycle fades faster than expected. EPS at $15.50, multiple compresses back to 15.8x.
Probability-weighted PT: (0.25 × $385) + (0.50 × $330) + (0.25 × $245) = $321, which rounds to a $330 base-case PT and supports a Buy rating from $255.
What to Watch Next
Three near-term checkpoints:
- Q2 2026 earnings (late July): Does data center revenue hit ~$110M for the quarter? That's the run-rate consistent with the $500M annual guide.
- Hyperscaler capex commentary: Watch META, MSFT, GOOG, and AMZN July earnings for any AI infra capex guidance changes. NXP's data center business is concentrated in the top 5 buyers.
- Analyst Day (September): Sotomayor has hinted at a multi-year data center roadmap reveal. If the 2028 number lands above $1B with credible margin assumptions, the multiple expansion goes another leg.
Bottom Line
NXP isn't an AI stock the way NVIDIA is an AI stock. Nobody's pitching it on a CES keynote slide. But it's becoming an AI infrastructure stock — one that ships the unsexy silicon hyperscalers can't run racks without — and the market just woke up to that.
At $255, you're paying ~17x forward earnings for a business with a refreshed auto cycle, structurally higher content per vehicle, and a data center segment growing 150% with multi-year visibility. That's a setup that doesn't usually stay this cheap.
We're Buy-rated with a $330 12-month PT. To pressure-test against the Street and see how our scenarios stack up against consensus, check Edgen's live consensus and analyst targets page.
For the broader AI hardware infrastructure cluster, also worth reading: Seagate's HDD shortage thesis and Mozaic margins, our take on LRCX and the AI semi-equipment cycle, and the 2026 semiconductor sector overview.
FAQ
Q1: Why did NXP stock jump 18% if the EPS beat was only $0.10?
The beat itself was modest. The move came from the data center revenue guide ($200M to $500M+), which forces analysts to model a new growth vector and re-rate the multiple. That's a thesis change, not a number change.
Q2: Is NXP an AI stock now?
Sort of. It doesn't sell AI accelerators or GPUs. But it sells the analog, power, and interface silicon that surrounds those chips on hyperscaler boards — and that exposure is finally large enough to move the P&L.
Q3: What's the biggest risk to the bull case?
Hyperscaler capex digestion. If META, MSFT, or GOOG cut 2H 2026 AI infra spend, NXP's data center ramp slows and the re-rating thesis stalls. Auto cyclicality is a secondary risk but more priced-in.
Q4: How does NXP compare to TXN as an analog play?
TXN is more diversified, higher margin, and trades at a premium multiple (~24x). NXP is more auto-concentrated but with faster near-term growth and a cheaper entry. If you want quality at a price, TXN. If you want growth-at-reasonable-price with a re-rating catalyst, NXP.
Q5: What's the right position size?
That depends on portfolio context, but for a thesis with 50% base-case probability and ~30% upside to base case, a 2–4% position is consistent with how most institutional analog books are sized. Bull case implies ~50% upside; bear case implies ~4% downside from $255 — that's a favorable asymmetry.
