MPWR · Monolithic Power Systems, Inc.

Current Thesis

The AI data-center power-delivery narrative is still ACCELERATING, but the tape and the governance backdrop both changed in the last week. Q1 2026 (reported 2026-04-30) printed $804.2M revenue (+26.1% YoY, +7.1% QoQ) versus $781.6M consensus, with Enterprise Data nearly doubling YoY to $262.8M the GPU/accelerator power-management content ramp that is the actual narrative leg. Q2 guide of $890–910M (midpoint ~$900M, ~+35% YoY / ~+12% QoQ) implies the ramp is mid-cycle, not topping. What's new: the stock flushed -10.38% on 2026-06-05 to $1,481.05, off the 2026-06-02 high near $1,625 and the 52-week high of $1,714.09, dragging it back onto its 50-day SMA (~$1,455). That is the pullback the early-June setup lacked. The other new development is a Feb-26-2026 non-reliance restatement plus a wall of CEO/insider selling, a sentiment overhang on a 106x trailing P/E. So entry quality improved while a fresh, non-operational risk got layered on. This is now a probe-able pullback to rising MA support, not a chase into a vacuum but the restatement and insider wall keep it short of a fat pitch.

Bull Case

• Enterprise Data ~2x YoY to $262.8M in Q1 2026 (2026-04-30) direct read-through that MPWR is winning power-management content on AI accelerator boards, refuting the late-2024 fear it lost Blackwell vertical-power-delivery sockets to Infineon/Renesas.
• Q2 2026 guide $890–910M (2026-04-30), ~+35% YoY at the midpoint and ~+12% sequentially off a record $804.2M quarter. Sequential acceleration is the cleanest sign the content cycle is not topping.
• Communications $111.5M, +55.5% YoY / +33.1% QoQ (Q1 2026) a second growth vector firing alongside data center.
• Margins held: GAAP GM 55.3%, non-GAAP 55.5% (Q1 2026), roughly flat QoQ through the volume/mix shift; operating profit $241.2M (+42.9% YoY), net income $193.2M (+44.4%). Non-GAAP EPS $5.10 beat the $4.90 estimate.
• Sell-side targets reset higher and now sit above spot: Truist $1,805 (raised from $1,396), KeyBanc $2,000, TD Cowen $1,850, Needham $1,750; median ~$1,662 versus the $1,481.05 close (2026-06-05). For the first time in weeks there is PT headroom rather than a price that has outrun every target.
• Cluster confirmation: SOXX +27% in April 2026 (its best month on record), with a combined $5.45B SOXX+SMH inflow a complex-wide momentum regime, not a single-name pop.

Bear Case

• Non-reliance restatement (8-K, 2026-02-26). The audit committee determined certain prior financials should not be relied upon: a non-cash deferred income-tax error tied to a one-time foreign tax incentive, restating FY2024 audited results and all 2025 interim periods, including a $195M reduction in FY2024 income tax benefit. Revenue and non-GAAP metrics are unchanged, so the operating story is intact but it is a live governance overhang the market is still digesting, and it raises the bar of trust ahead of any amended filings.
• Insider selling is one-directional and heavy. Trailing six months: ~127 insider sales, 0 buys. Additional EVP selling into late May 2026. Clustered, no offsetting buys, at the highs.
• The 6/5 flush was a ~10% single-day air-pocket, with semis dominating the day's S&P 500 losers distribution behavior, not orderly profit-taking. The pullback may not be finished.
• Valuation leaves no cushion. 106x trailing / ~58x forward P/E on a $72.76B cap; any AI-capex-digestion wobble has room to mean-revert toward the 200-day (~$1,182).
• Retail-saturation marker: a 2X leveraged single-stock MPWR ETF (Tradr) launched 2026-04-24. Leveraged retail vehicles arrive at narrative maturity.
• NVIDIA concentration remains the single point of failure: the Enterprise Data ramp is both the bull case and the headline risk if a Rubin-generation socket-share loss or hyperscaler capex pause surfaces.

Setup & Price Structure

Close $1,481.05 (2026-06-05), down 10.38% on the day, off the 2026-06-02 spike near $1,625 and the 52-week high of $1,714.09. The flush parked price directly on the 50-day SMA (~$1,454.93) the first test of rising trend support since the April run. The 200-day SMA sits well below at ~$1,182.46, and a 20-week EMA in the ~$1,250 zone marks the deeper trend line. YTD performance is roughly +86.5%, so this is a high-beta leader giving back a slice of an outsized run, not a broken chart. The structurally important read: the early-June problem was a price extended above every analyst target into a catalyst vacuum; after the pullback, spot now sits below the median PT (~$1,662) with the 50-day as the line in the sand. A daily close that holds the 50-day and reclaims $1,550 re-arms the uptrend; a close that loses ~$1,455 without a quick reclaim turns this into distribution toward the 20-week EMA / 200-day. RSI is not at blow-off extremes after the drop, so this is a support test rather than a reversal signal for now.

Catalyst Calendar (next 30 days)

• No scheduled binary catalyst before ~2026-07-07. The Q1 print already fired (2026-04-30) and Q2 earnings are out at 2026-08-03 (after close, est.), outside the 30-day window.
• Restatement resolution (undated, watch June–July): amended/restated filings for FY2024 and 2025 interims tied to the 2026-02-26 non-reliance 8-K are the next company-specific event; clean resolution removes the governance discount, further delay extends it.
• Form 4 flow (ongoing): any pause in CEO/insider selling or a first insider buy would be an early sentiment tell; continued selling into the pullback is the opposite.
• Sector reads: semiconductor-complex tape (SOXX/SMH), NVIDIA commentary, and any hyperscaler capex updates can move the name 5–10% intraweek regardless of MPWR-specific news.

What Would Change Our Mind

Bullish confirmation: a daily/weekly close that holds the 50-day (~$1,455), reclaims $1,550+, and prints with the semi complex re-accelerating, plus clean restatement filings and a halt to insider selling that combination upgrades the read from probe to size-up on the next higher low. Bearish invalidation: a daily close below the 50-day (~$1,455) that is not reclaimed within a week (distribution, target the 20-week EMA ~$1,250 / 200-day ~$1,182); restatement scope widening beyond the non-cash tax item into revenue or operating metrics; or the Q2 print (2026-08-03) coming in below the $890M guide floor or Enterprise Data posting its first sequential decline that is the AI-power content ramp rolling over. Continued one-way insider selling plus a failed 50-day test is the practical "stand aside" combination even with the narrative intact.

Correlation Notes

Archetype: picks & shovels: a power-management content supplier into NVIDIA accelerator boards, so it co-moves with SOXX/SMH/NVDA and trades on AI-capex sentiment, not on a retail squeeze the 2X leveraged ETF is noise around the float, not the driver. High beta to the semi complex (SOXX +27% in April 2026 was the tailwind; the 6/5 semi-wide selloff was the headwind). Idiosyncratic risk is concentrated in two places: NVIDIA socket-share at each accelerator generation, and the restatement/governance overhang that is independent of the tape. Pairs/peers for confirmation or divergence: monitor whether other AI-power and analog names (and NVDA itself) hold their own MAs on the same days MPWR tests its 50-day cluster strength validates the pullback-buy; MPWR breaking while peers hold flags a name-specific (restatement/insider) problem.

Notes

Q1 2026 reference figures (reported 2026-04-30): revenue $804.2M (+26.1% YoY), Enterprise Data $262.8M, Communications $111.5M, GAAP GM 55.3% / non-GAAP 55.5%, non-GAAP EPS $5.10 vs $4.90 est. Q2 guide: $890–910M revenue, GAAP GM 55.1–55.7%.