STM · STMicroelectronics N.V. · Stock research

Current Thesis

The June re-rating happened, and it has now failed its first real test. STM was repriced through Q2 2026 from a low-margin auto/industrial cyclical into an AI-data-center and space-infrastructure supplier — the 2026-06-02 decision to roughly double the 2026 data-center ambition to ~$1B (from "nicely above $500M") and lift 2027 to ~$2B was the catalyst, and shares printed a 25-year-high close of $79.82 on 2026-06-22 on the back of it plus the reported $85B SpaceX IPO headline.

Since then the tape has done nothing but give it back. Price closed $62.06 on 2026-07-17 — down 22% from the June high, below the $71.42 close of 2026-06-26, and under the early-June breakout base that the whole momentum leg was built on. A $1.5B convertible offering announced 2026-06-16 put a hedging overhang on the shares. On 2026-07-13 STM fell 4.2% in the US and 7.1% in Milan to €52.08, the worst performer on that exchange, on a July expiry session and broad Asian semi weakness. Most telling: BofA's upgrade to Buy with a $100 target — a full-fat, +60% target — could not hold the tape.

The narrative is intact on paper and broken on the chart. Everything now compresses into the 2026-07-23 Q2 print, which lands before the European open, three trading days out. This is not a momentum setup any more; it is an earnings coin-flip on a name that just lost its base.

Bullish and bearish views on STMicroelectronics N.V.

The model's bull view on STMicroelectronics N.V. (STM), in brief: Data-center revenue ambition roughly doubled on 2026-06-02 to ~$1B for 2026 and ~$2B for 2027, up from ">$500M" and ">$1B" reaffirmed at the Q1 call — the act of raising a target quarter-on-quarter is the signal. The bear view: The breakout base is gone. Two pushes into the high-$70s (2026-06-03 at $79.71, 2026-06-22 at $79.82) both reversed, and the second unwind has run 22% without finding a bid. An upgrade to a $100 target arriving into a 7% down day is exhaustion behavior. When the most bullish… Both cases follow in full.

Bull Case

• Data-center revenue ambition roughly doubled on 2026-06-02 to ~$1B for 2026 and ~$2B for 2027, up from ">$500M" and ">$1B" reaffirmed at the Q1 call — the act of raising a target quarter-on-quarter is the signal.
• Q1 2026 (2026-04-23): revenue $3.10B, +23% YoY, ahead of consensus; Q2 guided to $3.45B ±350bps, +11.6% sequential and +24.9% YoY. Book-to-bill was described as well above 1.0 across all end markets and regions, with distribution inventory normalized.
• Sell-side is still upgrading into weakness: BofA to Buy, PT $100 from $83; TD Cowen to $70 from $50 on 2026-07-08; Morgan Stanley to €78 from €74 on 2026-06-30. Consensus sits at $75.25 average with 12 Buys, 6 Holds, 0 Sells — that is ~21% above the 2026-07-17 close.
• Space franchise disclosed at >$3B cumulative over three years (2026-05-04); LEO-program revenue ran $175M (2021) to $600M (2025), +243%, with ~7.5B Starlink RF chips delivered and run-rate set to double by 2027.
• AWS multi-year engagement (2026-02-09) reportedly carries warrants toward a ~3% Amazon stake — hyperscaler equity in a component supplier is the strongest demand-durability signal available.
• ST54M single-die post-quantum-crypto secure element (2026-06-24) extends the high-margin payments/identity socket.
• The convertible is cheap money: 0.00–0.50% on the 2031 tranche, 0.625–1.125% on 2033, at a 47.5–55% conversion premium, and it retires the $750M 2027 converts struck at $45.10. Structurally this is a balance-sheet upgrade, not a raise.

Bear Case

• The breakout base is gone. Two pushes into the high-$70s (2026-06-03 at $79.71, 2026-06-22 at $79.82) both reversed, and the second unwind has run 22% without finding a bid.
• An upgrade to a $100 target arriving into a 7% down day is exhaustion behavior. When the most bullish possible sell-side datapoint cannot produce a green close, the marginal buyer is already in.
• The convertible creates mechanical selling: delta-hedging on a $1.5B dual-tranche issue that settled ~2026-06-23 sits directly on top of the June top. That is not sentiment; it is flow.
• Company buybacks are underwater. 96,195 shares repurchased 2026-07-06 to 07-10 at a weighted average €60.27, against a €52.08 Milan print on 2026-07-13.
• Earnings quality is thin. TTM revenue $12.38B but TTM net income only $147M, down 86.6%, EPS $0.16. Q1 GAAP net income was ~$37M on $3.10B of revenue. Forward P/E of 36 requires the 2027 ramp to arrive on schedule.
• Margins are recovering from a low base, not strong: Q1 GM 33.8%, Q2 guided ~34.8% including ~100bps of unused-capacity charges. China SiC wafer ASP pressure and automotive MCU competition (Infineon AURIX, onsemi Treo) still cap the blend.
• AI-DC plus space remain a single-digit slice of a ~$13–14B base. The re-rating is a bet on 2027 mix, and any AWS capex digestion removes the reason the multiple expanded.
• The SpaceX IPO halo is a sentiment proxy carrying zero STM revenue. It unwound faster than it built, which is what halos do.

Setup & Price Structure

• $62.06 close 2026-07-17, -1.13% on the day; market cap $54.72B; forward P/E 36.07, trailing 372 on depressed earnings.
• 52-week range $21.11–$81.42. The name is still roughly +2x off the April lows, so there is a large unrealized-gain overhang above.
• Sequence: $79.71 close 2026-06-03 → $79.82 close 2026-06-22 (rejected -4.6% to $71.42 by 2026-06-26) → $68.47 after a -4.2% session 2026-07-13 → $62.06 on 2026-07-17. Lower highs, lower lows, no reclaim attempt.
• The June breakout base has broken on a weekly-close basis. The next structural reference below is the pre-breakout May consolidation shelf in the high-$50s.
• Price sits ~17% below the $75.25 consensus target and 38% below BofA's $100 — the sell-side is now above the tape after having chased it up in June, which inverts the usual confirmation read.
• Volume on the decline (15.89M share sessions) has not thinned out. Distribution, not drift.

Catalyst Calendar (next 30 days)

• Late July / early August (est.) — Infineon and onsemi prints frame the analog and SiC margin narrative. A China SiC ASP reset disclosed on either call reads directly through to STM's blended margin.

Elapsed catalysts

• 2026-07-23 — Q2 2026 results, released before the European market open; conference call 09:30 CET / 03:30 ET. The binary. Watch three lines: revenue against the $3.45B ±350bps guide, gross margin against ~34.8% including ~100bps unused-capacity charges, and whether the ~$1B 2026 data-center target is reaffirmed with a quantified H2 run-rate or softened to a range. (passed 17d ago)
• 2026-07-23 — Q3 2026 guidance issued on the same call. Given the ramp math implied by ~$1B for 2026, H2 has to step up materially; a flat-to-down Q3 guide breaks the data-center story regardless of the Q2 beat. (passed 17d ago)
• Weekly, ongoing — share-repurchase status disclosures under the existing program (most recent 2026-07-13 covering 2026-07-06 to 07-10). Pace and average price relative to spot are a readable signal on management's own conviction. (passed 27d ago)
• No other confirmed company-specific dated catalyst inside 30 days. Between now and 2026-07-23 the tape is pre-positioning and semiconductor beta. (passed 17d ago)

What Would Change Our Mind

• Bearish confirmation: a weekly close below $58 takes out the pre-breakout May shelf and settles the question — the June re-rating was a failed breakout, and the correct read becomes a broken momentum name in a distribution phase, not a discount.
• Bearish confirmation, fundamental: Q2 revenue below the $3.45B guide midpoint, gross margin below ~33%, book-to-bill under 1.0, or any softening of the ~$1B 2026 data-center language on the 2026-07-23 call.
• Bullish reset: a post-print weekly close back above $70 on 3x average volume, with book-to-bill disclosed above 1.05 and a quantified H2 data-center run-rate consistent with ~$1B. That rebuilds the base and restores a tradeable trend structure.
• Bullish, second-order: a named hyperscaler or optical-module design win with disclosed dollar content on the PIC100 silicon-photonics platform. The AWS warrant structure set the template; a second one converts a single-customer story into a franchise.
• Neutral-to-watch: convertible hedging pressure fading — visible as declining volume on down days through August — would remove the mechanical bid-side vacuum without requiring any fundamental news.

Correlation Notes

• Analog/power semis: trades with Infineon, onsemi, NXP and Texas Instruments on the auto/industrial cycle and SiC pricing. This cohort has been the drag, not the driver, and STM's June outperformance came from decoupling into the AI-DC theme. That decoupling has now reversed.
• AI data-center infrastructure: the second-derivative read is optical/photonics and power-delivery names. STM only participates here through PIC100 and DC power content; it is a late-cycle, low-share participant, so it gets the beta on the way down more reliably than the alpha on the way up.
• Space/satellite connectivity: loosely correlated to SpaceX private-market marks and Starlink terminal volume. This is a headline-driven sleeve with no near-term reporting cadence, so it adds volatility without adding gradeable datapoints.
• Broad semis: high-beta to SOX and to Asian semiconductor tape — the 2026-07-13 decline was explicitly attributed to Asian semi weakness rather than anything company-specific, which is how this name behaves when it has no fresh number of its own.
• FX: dual-listed with a euro-denominated primary line. The €52.08/$62.06 spread means dollar-reported results carry a translation component; a strengthening dollar compresses reported revenue against a guide set in dollars.