FLEX · Flex Ltd. · Stock research

Reference close: $110.45 on 2026-08-21 (split/dividend-adjusted daily bars).

FLEX — Flex Ltd.

Current Thesis

The 8 August note put the trend marker at the lost ~$142 weekly shelf and described the leg as a de-rating: operating numbers compounding while the multiple contracts. Two weeks on, the compression has continued. The last completed close is $110.45 (2026-08-21), 31.9% below the $162.07 52-week high, with a three-month price change of -16.6% and RSI(14) at 42.8 — falling, but nowhere near a washout reading.

Nothing company-specific is dated in the window that produced the latest leg. No SEC filings were reported between the 2026-07-29 8-K and 2026-08-21; the only Flex-tagged item in the news stream was a 2026-08-14 screen of AI-buildout names trading more than 20% below their highs. A Form 4 shows the Chief Commercial Officer selling 2,755 ordinary shares on 2026-08-18 at weighted-average prices of $119.7946–$123.248, footnoted as tax withholding on RSU vesting — which also dates the intraweek level from which the shares fell to Friday's $110.45.

What an investor buys here is the bet that multiple compression stops before the two scheduled re-anchoring events: the 2026-11-10 Investor Day, where standalone SpinCo financial targets would first be presentable, and the tax-free separation of Cloud & Power Infrastructure targeted for Q1 CY2027. The underlying business is not the question — Q1 FY2027 net sales were $7.928B (+21% YoY) against $7.557B consensus, CPI revenue $2.2B (+35% YoY), and the FY2027 guide was raised to $33.7–35.2B and $4.42–4.74 adjusted EPS on 2026-07-29. The question is what the market pays for it.

The narrative is saturated. Three dated events that should have produced a bid did not. S&P 500 inclusion effective 2026-06-22 gave a +6.0% pop and round-tripped -9.04% to $146.70 by 06-26. The beat-and-raise on 2026-07-29 closed -9.06%. And the post-print sell-side response was to cut targets while keeping ratings. The mainstream coverage cluster is behind the name, not ahead of it: Josh Brown's CNBC segment 2026-05-06, Cathie Wood's "blasts from the past" 2026-05-11, Cramer's +50% call 2026-05-27, IBD SwingTrader 2026-06-02 — all before the June top. The August coverage shape is a bargain screen. A flip to dead would be dated by a weekly close below $100 alongside the 2026-11-10 Investor Day passing without SpinCo numbers.

Bullish and bearish views on Flex Ltd.

The model's bull view on Flex Ltd. (FLEX), in brief: Q1 FY2027, reported 2026-07-29: net sales $7.928B, +21% YoY, above the $7.557B consensus; adjusted EPS $1.00 versus $0.91 estimate; GAAP EPS $0.76; adjusted operating margin 6.7%, +70bps YoY; adjusted gross margin 9.6%. The bear view: Every post-print target action was a cut. Both cases follow in full.

Bull Case

• Q1 FY2027, reported 2026-07-29: net sales $7.928B, +21% YoY, above the $7.557B consensus; adjusted EPS $1.00 versus $0.91 estimate; GAAP EPS $0.76; adjusted operating margin 6.7%, +70bps YoY; adjusted gross margin 9.6%.
• Cloud & Power Infrastructure revenue $2.2B in Q1 FY2027, +35% YoY, accelerating from +31% YoY to $1.8B in Q4 FY2026 (reported 2026-05-06). The segment growth rate went up, not down.
• FY2027 guidance raised on 2026-07-29 to $33.7–35.2B sales from $32.3–33.8B, adjusted EPS to $4.42–4.74 from $4.21–4.51 against a $4.45 consensus, adjusted operating margin 7.0–7.2%. The near quarter was also guided above the street: Q2 FY2027 sales $7.95–8.25B versus $7.835B consensus, adjusted EPS $1.00–1.07 versus $0.99.
• Product cadence attaches revenue to named rack roadmaps: 110kW power shelf for NVIDIA Vera Rubin NVL72, 30kW capacitive energy storage and the BMR317 converter shown at COMPUTEX 2026-06-02; Cerebras CS-3 wafer-scale production ramp at Flex's California facilities announced 2026-07-09.
• Sell-side consensus sits far above spot: 11 analysts, average 12-month target $160.50, "Strong Buy" aggregate, against the 2026-08-21 close of $110.45 (stockanalysis.com consensus screen, 2026-08-21).
• Separation machinery is moving: leadership teams for both companies named in the 2026-07-29 8-K, with CEO Revathi Advaithi slated to run SpinCo, and Investor Day booked for 2026-11-10.

Bear Case

• Every post-print target action was a cut. Barclays went to $144 from $203 (Overweight maintained) on 2026-07-30 — a 29% reduction taken after a beat and a raise. Goldman cut to $154 from $177 (Buy). JPMorgan cut to $160 from $175 (Overweight). Baird cut to $142 from $165 (Outperform). When targets fall on good numbers, the model input that moved is the multiple.
• The gap between the $160.50 average target and the $110.45 close is wide enough that it closes one of two ways, and the documented direction of travel since 2026-06-04 has been targets falling toward price.
• Cash conversion in Q1 FY2027: operating cash flow $276M, free cash flow $41M, cash $2.84B, and no share repurchases in the quarter ended 2026-06-26. The company itself was not a bidder for the stock through the drawdown.
• Adjusted gross margin of 9.6% and an adjusted operating margin guided to 7.0–7.2% define the ceiling on how much of the AI revenue converts to earnings. Growth at contract-manufacturing economics does not automatically earn a data-center multiple.
• The 2026-08-05 AGM approved Board authority to issue ordinary shares up to 20% of issued share capital — a standing overhang independent of the spin.
• Order-of-operations risk in the separation: no SpinCo Form 10 was on file as of 2026-08-21, and the SEC declaring it effective, plus Singapore High Court approval, are conditions to a Q1 CY2027 close.
• The 2026-08-18 Form 4 sale, whatever its tax purpose, is the only insider transaction disclosed in the window, and it is a sale.

Setup & Price Structure

• $110.45 close (2026-08-21); $162.07 52-week high; 31.9% below it; three-month price change -16.6%; RSI(14) 42.8.
• The prior structural marker at ~$142 is gone, and the drawdown proceeded through the previously cited ~$130 area without a visible shelf. No higher low has been confirmed since the 2026-06-26 close of $146.70.
• Crowding evidence points at absence rather than excess: RSI in the low 40s is not an extended-above-a-rising-average condition, and the retail-attention cluster (May–early June 2026) is more than two months stale. The observable is a thin new bid, with two positive catalysts sold and a further leg lower into no news.
• Positioning items on the record: no buyback in the June quarter; a 20%-of-share-capital issuance authority approved 2026-08-05; a CCO Form 4 sale on 2026-08-18; passive index flow now two-way since S&P 500 membership took effect 2026-06-22.
• No base has formed. The next scheduled company-specific test is roughly nine weeks out.

Catalyst Calendar (next 30 days)

• 2026-08-26 — NVIDIA Q2 FY2027 report, after the close. Not a Flex event. NVIDIA guided to approximately $91.0B revenue; consensus per 40 analysts is $91.85B and $2.08 EPS. Rack-scale and capex commentary sets the multiple for the whole AI-EMS/data-center-power cohort, and Flex's CPI growth is levered to the Vera Rubin NVL72 build.
• No Flex-specific dated event falls inside the 30-day window. The next company catalysts sit beyond it: ~2026-10-28 (est.) Q2 FY2027 results, and 2026-11-10 Investor Day.
• Undated but live: the SpinCo Form 10 filing, which is the first hard document on separated economics and could land any time before the Q1 CY2027 target.

What Would Change Our Mind

The structure that would have to hold is a higher low above the 2026-08-21 close, formed without the help of a company catalyst, because there is no company catalyst until late October. Absent that, the de-rating is the whole price action. A weekly close below $100 would confirm the compression running on with no base formed and would put the shares below every reference level cited in this coverage since first write-up. Secondary conditions, each independently gradeable: CPI segment growth printing below 30% YoY at the ~2026-10-28 Q2 FY2027 report after +35% in Q1; the 2026-11-10 Investor Day passing without standalone SpinCo financial targets; FY2027 adjusted operating margin guided below the 7.0–7.2% band; or a second consecutive quarter with free cash flow far below adjusted net income and continued zero repurchases.

What would rebuild the case in the other direction: a Form 10 on file with SpinCo segment economics that carry a gross margin above the 9.6% consolidated Q1 figure, resumption of buyback disclosure at the Q2 print, or any sell-side action that raises rather than trims a target — the first upward revision since Keybanc's $180 on 2026-05-07.

Correlation Notes

• Flex trades inside the AI-EMS and data-center-power complex; NVIDIA's 2026-08-26 print is the cohort's shared factor, and the CPI segment's growth is attached to accelerator rack build-outs by disclosed product wins (110kW NVL72 power shelf, 2026-06-02).
• Cerebras is a second read-across since the 2026-07-09 CS-3 production ramp at Flex's California facilities; a Cerebras earnings-cycle event was being previewed in the 2026-08-12 press flow.
• The RemainCo half is consumer, automotive, industrial and health-solutions EMS, which tracks the industrial cycle rather than AI capex. That mismatch inside one share class is the argument the spin is meant to settle, and until the Form 10 exists there is no filed document letting the market price the halves separately.
• S&P 500 membership since 2026-06-22 makes index flow a two-way factor; the +6.0% inclusion pop had fully round-tripped by 2026-06-26.