STMicroelectronics N.V. -- Valuation, Catalysts and Risks
STM trades at $66.86 (May 22, 2026) with ~$54B market cap after a +137% 12-month rerate
driven by AI/data-center optimism, the AWS multi-billion engagement, and the $1B/2027 AI
revenue target. The street average price target ($55) is materially BELOW spot -- meaning
sentiment, not earnings, is doing the work. At 19.2x EV/EBITDA on FY26E numbers, STM is
modestly rich vs analog/MCU peers (NXPI 14.2x NTM, IFX 15.7x TTM, RNECY 17.1x). The
fundamental tape is still mid-cycle recovery: FY2025 revenue $11.80B (-11% YoY); FY2025 GAAP
operating income just $175M (1.5% margin); Q1 2026 GAAP op income only $70M. Recovery is
real but priced. Not "priced for perfection on numbers," BUT priced for the AI/LEO narrative
being correct.
Dimension weight: 15%
Dimension Score
4/10
Concerns & Risks
EV / EBITDA (FY26E)
~19.2x
vs peer avg ~17.5x | premium
Enterprise Value
$53.85B
Mkt cap $54B - $0.15B net cash
China Exposure
~15-20%
Of revenue; Asia Pac 63% of FY25
Next Catalyst
Late Jul
Q2 2026 Earnings (~$3.45B guide)
Valuation Summary -- FY26E and FY27E
Inputs: market cap ~$54B, Q1 2026 cash $1,889M
+ marketable securities $832M
= $2,721M liquidity; total long-term debt $2,569M.
Net cash position ~$0.15B → EV ≈ $53.85B. FY26E revenue ~$14.1B and FY27E EPS
~$2.10 from consensus aggregate. FY26E EBITDA built from consensus op income ~$0.9-1.0B + D&A
~$1.9B (run-rating Q1 2026 D&A of $454M
× 4) → ~$2.8B. FY27E EBITDA ~$3.6-3.8B at ~38% GM mid-recovery.
| Metric | FY26E Value | STM Multiple | Peer Avg | Verdict |
|---|---|---|---|---|
| EV / Revenue | $14.1B | 3.8x | ~6.4x (IFX 6.5x, NXPI 6.8x, RNECY 6.1x) | Optically cheap; normalized for STM's lower EBITDA margin |
| EV / EBITDA (primary) | $2.8B | ~19.2x | ~17.5x (NXPI 14.2x NTM, IFX 15.7x, RNECY 17.1x, MCHP 29.3x) | In-line to slightly rich vs peers; cheap vs MCHP, premium to NXPI |
| EV / EBITDA (FY27E) | $3.7B | ~14.6x | ~15x | Reasonable IF recovery comes through |
| P / E (FY26E) | $1.05 EPS | ~63.7x | n.m. (post-trough EPS distort) | Optically broken; trough-EPS denominator |
| P / E (FY27E -- MS) | $2.10 EPS | ~31.8x | ~22-28x | Expensive on FY27 even if EPS doubles |
Consensus EPS from company_context aggregate (S&P Global / ChartMill / Mizuho / Morgan Stanley).
Peer avg multiples from Alpha Spread (IFX), GuruFocus (NXPI), TIKR (NXP NTM), Power Electronics News.
EV Bridge (Q1 2026)
Market Cap
~$54.0B
+ Total Debt
$2.57B
- Cash + Securities
$2.72B
Net Cash
+$0.15B
Enterprise Value
$53.85B
Primary verdict: at 19x EV/EBITDA on a number that
already assumes a ~2.5x rebound from the FY25 trough operating profit of
$175M,
STM is priced for everything to go right. EV/EBITDA is modestly above the
analog/MCU peer median once you strip out Microchip's outlier. EV/Sales of 3.8x looks cheap
only because STM's structural EBITDA margin (~20% peak, ~16% midcycle) is lower than IFX
(~25%) or NXPI (~35%).
Peer Comparison
| Company | Mkt Cap | EV / EBITDA | EV / Revenue | Peak EBITDA Margin |
|---|---|---|---|---|
| STMicroelectronics (STM) | $54B | 19.2x FY26E | 3.8x FY26E | ~20% |
| Infineon (IFX) | $55B | 15.7x TTM | 6.5x | ~25% |
| NXP Semiconductors (NXPI) | $57B | 14.2x NTM | 6.8x | ~35% |
| Renesas (RNECY) | $30B | 17.1x | 6.1x | ~30% |
| Microchip (MCHP) -- outlier | $30B | 29.3x | 7.2x | ~38% |
| Peer Average (ex-STM, ex-MCHP outlier) | — | ~17.5x | ~6.4x | ~30% |
Commentary: STM trades at a modest premium on EV/EBITDA (19.2x vs 17.5x peer avg)
despite a structurally lower EBITDA margin profile (~20% peak vs Infineon ~25%, NXPI ~35%). The
EV/Sales discount (3.8x vs 6.4x peer avg) reflects this margin reality, not a value opportunity.
Multiple expansion thesis depends entirely on AI/LEO narrative being correct and margin recovery
pushing peak EBITDA margins above the historical ~20% ceiling.
Stress Test -- Is STM Priced for Perfection?
Implied normalized earnings at current price:
EV = $53.85B (mkt cap $54B - small net cash position). A peer-median EV/EBITDA of ~17x (mid-cycle analog/MCU) implies the market is pricing ~$3.17B of normalized EBITDA. Backing out FY25 D&A of $1,854M (rising to ~$2.0B with NXP MEMS and capex ramp), that means normalized operating income ~$1.15-1.20B and normalized op margin ~7-8% on $15B revenue, OR ~$1.0B op income on $14B revenue. FY24 was $1,676M operating income on $13.27B revenue (12.6% margin) -- the prior peak.
EV = $53.85B (mkt cap $54B - small net cash position). A peer-median EV/EBITDA of ~17x (mid-cycle analog/MCU) implies the market is pricing ~$3.17B of normalized EBITDA. Backing out FY25 D&A of $1,854M (rising to ~$2.0B with NXP MEMS and capex ramp), that means normalized operating income ~$1.15-1.20B and normalized op margin ~7-8% on $15B revenue, OR ~$1.0B op income on $14B revenue. FY24 was $1,676M operating income on $13.27B revenue (12.6% margin) -- the prior peak.
Implied Normalized EBITDA
~$3.17B
At 17x peer-median EV/EBITDA; ~75-80% of 2024 peak op income
Implied Normalized Op Income
~$1.15-1.20B
~7-8% op margin on $15B revenue; vs FY24 peak 12.6% on $13.27B
Double-Hit Risk
HIGH
AI miss collapses both implied earnings AND multiple simultaneously
Verdict: Not quite "perfection" -- the implied
normalized earnings are NOT at peak-cycle 2024 levels, they're at a reasonable midcycle
~$1.0-1.2B operating income. BUT the multiple ALSO assumes the AI/LEO/silicon photonics
narrative materializes (i.e., that the rerate is not just a cycle bounce but a structural
premium). If AI revenue underdelivers, you simultaneously lose (a) the implied earnings
uplift AND (b) the multiple expansion -- that's the double-hit risk.
Not priced for perfection on numbers, but priced for the AI narrative being broadly
correct. A ~20% margin of safety has been eliminated by the rerate.
China Exposure Analysis
Geographic Revenue Mix (FY2025)
| Asia Pacific | $7,455M (63%) |
| EMEA | $2,449M (21%) |
| Americas | $1,896M (16%) |
| China (within Asia Pac) | ~15-20% of total revenue |
| Q1 2026 Asia Pac | $1,896M (61% of $3.10B) |
Key Dynamics
Export control risk: STM's mix skews to mature-node analog/power/MCU/MEMS -- products NOT on the leading-edge logic export control lists. The China-for-China supply chain (STM32 MCU production at Huahong, SiC JV with Sanan in Chongqing) is an active mitigation.
Retaliation risk: Any escalation of EU-China or US-China trade actions could trigger reciprocal Chinese measures targeting European semis (precedent: Chinese countervailing duties on European auto chips floated 2024-2025).
EU/Italy reshoring offset: Catania €5B Silicon Carbide Campus (€2B Italian state aid under EU Chips Act) anchors a vertically integrated SiC fab targeted to start production 2026 and ramp to full capacity by 2033, up to 15,000 wafers/week at full build. Back-end loaded.
Risk rating: Mid-tier compared to a Qualcomm or Lam Research. The reshoring story is multi-year and dilutive in the near term.
Key Catalysts (Next 6-12 Months)
| Catalyst | Timing | Why it matters |
|---|---|---|
| Q2 2026 print | Late July 2026 | Guidance midpoint $3.45B revenue (+11.6% q/q, +24.9% y/y); GM 34.8% +/- 200bps. Read on whether AI ramp is materializing and whether 35.2% non-GAAP GM is the springboard to 40%. |
| LEO satellite investor call | May 4, 2026 (rearview) | Mgmt announced $3B+ cumulative LEO revenue ambition for 2026-2028. Quantifies the Starlink/SpaceX/LEO thesis underpinning the AI rerate. |
| Q3 2026 print | Late October 2026 | First quarter where AI revenue should be visibly accelerating; H2 seasonality + ramp = test of $500M+ data center target for FY26. |
| SiC ramp at Catania + Sanan JV | Sanan: end 2026; Catania: 2027-2028 | Bull-case unlock if EV demand resumes; bear case if SiC pricing collapse continues. |
| Silicon photonics PIC100 customer wins | Throughout H2 2026 | Mgmt announced HVM start for PIC100 in Q1 2026; needs second/third named hyperscaler beyond AWS to validate platform durability. |
| AI revenue progression vs target | Per-quarter checkpoint | "Nicely above $500M" for 2026, "well above $1B" for 2027 -- these are now consensus expectations, so the bar is on EXCEEDING. |
| EU AGM | May 27, 2026 | Resolutions and any commentary on FR/IT governance, governance friction risk. |
| Capital Markets Day | Not yet scheduled | Last formal CMD was 2022; a 2026/2027 CMD with refreshed long-term targets is overdue and would be a likely positive catalyst. |
Risk Register (Severity x Probability)
| # | Risk | Sev (1-5) | Prob (1-5) | Combined | Notes |
|---|---|---|---|---|---|
| 1 | Auto cycle delay / EV demand soft | 4 | 4 | 16 | Auto = 39% of FY2025 revenue. Q1'26 auto declined 10% sequentially even as YoY grew 15% off trough. Street/S&P Global sees auto recovery delayed to 2027. Tesla/Western OEM exposure at risk if Chinese EV makers (BYD etc.) keep gaining share with domestic chips. |
| 2 | SiC oversupply / pricing pressure | 4 | 4 | 16 | Wolfspeed emerged from Chapter 11 with debt cut 70% and is back competing. ONSemi 5x capacity expansion by 2026, Infineon aggressive on 200mm, Chinese substrate players (TanKeBlue, SICC) at ~40% market share with lower prices. P&D segment ran at -21.5% operating margin in Q1 2026 -- already in pain. |
| 3 | $1B AI target miss / derate | 5 | 3 | 15 | Stock has rerated 137% on this narrative. A "merely $700-800M" 2027 AI revenue print would crater the multiple. Capacity constraints (mgmt admits demand > supply for photonics) raise miss risk. |
| 4 | Capacity reservation fee runoff | 3 | 5 | 15 | Mgmt flagged $140M YoY headwind in 2026 from declining customer capacity reservation fees. Already in numbers but worth watching as Q2-Q4 unfold. |
| 5 | China export controls / retaliation | 4 | 3 | 12 | Less direct than for leading-edge logic, but escalation risk is binary and unmodelable. China-for-China supply chain helps but caps margin. |
| 6 | Margin recovery slower than modeled | 4 | 3 | 12 | Path to 40% GM requires $4B+ quarterly revenue. Currently $3.1B. Reshaping (200→300mm, 150→200mm SiC) means transitional inefficiency through 2026 and into 2027. Consensus FY26 EPS of $1.05 may be too high if GM trails the 35-38% bridge. |
| 7 | Currency (EUR cost vs USD reporting) | 3 | 4 | 12 | Italian/French fabs (EUR cost base) but reports USD. Mgmt called out EUR/USD at 1.15-1.16 in FY26 OpEx guide adds low double-digit OpEx growth. Weaker dollar (EUR strength) directly compresses GM. |
| 8 | Govt-share governance friction (FR/IT) | 3 | 3 | 9 | Combined ~27.5% French/Italian government ownership constrains M&A, capital allocation, and operational restructuring. Past public friction over CEO succession (2024-2025). Real but slow-burn. |
| 9 | Inventory rebuild risk | 2 | 3 | 6 | Inventory $3,170M at Q1 2026 vs $3,140M Q4 2025; DSI at 140 days vs 130. Distribution channel is normalized per mgmt but balance-sheet inventory still elevated. |
Honest Bull Case vs. Honest Bear Case
Bull Case
If AI revenue tracks above the $500M/$1B 2026/2027 path -- which mgmt explicitly said is
constrained by capacity, not demand -- STM compounds at >20% on the highest-margin
segment of its portfolio (silicon photonics, BiCMOS, high-perf MCU for optical interconnect)
and earns a true AI infrastructure multiple.
Concurrent recovery in automotive (especially the NXP MEMS acquisition driving above-market sensor growth) and industrial restocking gets the company back to $4B+ quarterly revenue and 40% GM by 2027, justifying $2.10-2.50 normalized EPS.
The LEO satellite business adds an idiosyncratic $3B+ cumulative tailwind over 2026-2028, and the Catania SiC capacity becomes the European-strategic asset that Western OEMs preferentially source from to de-risk their own China exposure.
Net: with $2.50 normalized EPS, P/E of 27x looks like a fair growth-at- reasonable-price for a unique European AI-infrastructure-meets-power-semis story.
Concurrent recovery in automotive (especially the NXP MEMS acquisition driving above-market sensor growth) and industrial restocking gets the company back to $4B+ quarterly revenue and 40% GM by 2027, justifying $2.10-2.50 normalized EPS.
The LEO satellite business adds an idiosyncratic $3B+ cumulative tailwind over 2026-2028, and the Catania SiC capacity becomes the European-strategic asset that Western OEMs preferentially source from to de-risk their own China exposure.
Net: with $2.50 normalized EPS, P/E of 27x looks like a fair growth-at- reasonable-price for a unique European AI-infrastructure-meets-power-semis story.
Bear Case
The 137% YTM rerate has front-loaded several years of "what if AI works" optionality into
a $54B EV, while the company is still printing GAAP operating income of $70M on $3.1B
of quarterly revenue -- a 2.3% operating margin -- and the P&D segment loses money at
-21.5%.
AI revenue of "$500M" in 2026 is just 3-4% of total revenue, so the AI multiple is being applied to a business that is still 95%+ legacy automotive/industrial/consumer where China local competition (GigaDevice MCUs, BYD SiC self-supply, Chinese photonics entrants) is structurally accelerating, and where SiC pricing is in freefall post-Wolfspeed-emergence.
If FY26 EPS lands at $0.80 (not $1.05) and FY27 EPS at $1.60 (not $2.10) -- both plausible if GM expansion lags by 200bps -- the stock at $66.86 trades at 42x/22x and the AI-narrative premium collapses to a market multiple, implying ~$40-45 fair value (-35-40%).
Add China retaliation tail-risk and the FR/IT government-share governance overhang and the asymmetry from here is unfavorable.
AI revenue of "$500M" in 2026 is just 3-4% of total revenue, so the AI multiple is being applied to a business that is still 95%+ legacy automotive/industrial/consumer where China local competition (GigaDevice MCUs, BYD SiC self-supply, Chinese photonics entrants) is structurally accelerating, and where SiC pricing is in freefall post-Wolfspeed-emergence.
If FY26 EPS lands at $0.80 (not $1.05) and FY27 EPS at $1.60 (not $2.10) -- both plausible if GM expansion lags by 200bps -- the stock at $66.86 trades at 42x/22x and the AI-narrative premium collapses to a market multiple, implying ~$40-45 fair value (-35-40%).
Add China retaliation tail-risk and the FR/IT government-share governance overhang and the asymmetry from here is unfavorable.
Score Rationale: 4/10
STM scores 4/10 on Dimension 5, reflecting the intersection of a fully-priced multiple with material known risks and limited near-term de-risked catalysts.
Why 4 and not lower: STM is a real beneficiary of multiple secular tailwinds (AI infrastructure, LEO satellites, EU sovereignty), has a credible technology stack (silicon photonics on 12-inch is genuinely scarce), modest net cash balance sheet, and the China exposure (~15-20%) is meaningful but not existential. The 4 reflects that an investor isn't getting blown up at this level -- earnings will likely recover, and the LEO/AI numbers are likely directionally correct.
Why 4 and not higher: The full scoring rubric (10 = no China + cheap vs peers + catalyst near + no regulatory risk) is failed on multiple axes:
- ~15-20% China exposure (-1)
- 137% YTM rerate -- valuation is full at 19x EV/EBITDA on a number that already prices recovery (-2)
- Trading well above sell-side average PT of $55 implies asymmetric downside on any guide-down (-2)
- FR/IT government ownership and CHIPS Act dependency = regulatory/governance overhang (-1)
- Auto cycle still soft, P&D segment loss-making, capacity reservation fees rolling off (-1)
- Offsets: real AI catalysts in next 6 months and credible mgmt execution on cost program (+1)
Score = base 5 (neutral) − 3 (valuation + China + governance) + 2 (catalyst proximity + balance sheet) = 4 / 10.
Net assessment: Real beneficiary of secular AI/LEO tailwinds, but valuation is full and any AI/LEO miss carries asymmetric downside given the absent PT cushion. WATCH -- await Q2/Q3 2026 prints before upgrading conviction. Strategic ownership is NOT defensible until three conditions clear: (a) two consecutive quarters of positive FCF, (b) the $1B AI/2027 target is reaffirmed with a named hyperscaler #2 beyond AWS, and (c) gross margin clears 38%+ (currently 33.8%).
Sources: STMicroelectronics Q1 2026 earnings transcript (Apr 2026); Q4 2025 / Q3 2025 / Q2 2025 / Q1 2025 / Q4 2024 earnings transcripts; SEC 20-F FY2024 and FY2025; consensus aggregate from S&P Global / ChartMill / MarketBeat / WallStreetZen / Mizuho / Morgan Stanley (May 2026). Daloopa series IDs embedded throughout for Q1 2026 balance sheet, D&A, and revenue figures. Peer multiples: Alpha Spread (IFX), GuruFocus (NXPI), TIKR (NXP NTM), Power Electronics News (IFX/STM/onsemi/Renesas comparison). Catania SiC Campus: Datacenter Dynamics, ST press release, Evertiq (EU Chips Act). Wolfspeed Chapter 11 emergence: Financial Content / MarketMinute. SiC oversupply: Fortunebusinessinsights. Auto semi recovery: S&P Global Mobility, Manufacturing Dive. Analysis as of May 22, 2026.