Gaming, adtech & consumer internet

Claim

One of three July survivors, and the one with the biggest caveat.

NTES — conviction 7.0. It cleared on valuation for the same reason DOCS and ZBRA did: no AI-narrative premium to unwind. But NTES sits in the 🌏 Structural-Risk sleeve of Watchlist.md, and that placement is doing real work.

The distinction the sweep was not built to make

The July screens filtered on valuation risk — is this above its own band? NTES passes. It does not address structural risk — VIE ownership structure, regulatory intervention, delisting exposure, capital-controls risk. Those don't show up in a multiple, and no amount of multiple discount compensates for an outcome where the claim on the asset fails.

A cheap multiple is not compensation for a structural risk; it's a different axis. The sleeve system already encodes this — the note exists so a future screen doesn't quietly rank NTES against non-China names on price alone.

Also covered in this space with own reports: TCEHY, BABA (both July 15), TTD.

What would falsify this

Regulatory clarity that durably de-risks the VIE structure — or, in the other direction, any intervention that demonstrates the structural risk is live.

Related

[[health-tech]] and [[robotics-machine-vision]] — the other two July survivors, both without this caveat.

Gap found 2026-08-05 — APP was never in this sweep and it should have been

A portfolio-specific passage was removed from the public build.

Position and state as of 2026-08-05:

• $417.80, market cap $140.4B, −44.0% off the $745.61 high, −22% over 30 days
• 19.35x forward vs 36.33x trailing · consensus strong_buy, mean target $656.20 (+57.1%)
• FY25: revenue $5.48B, operating margin 78%, gross margin 88%, FCF $3.94B
• 3yr FCF CAGR 112.3%; diluted shares 371.6M → 342.0M (−2.7%/yr)
• 876 employees → ~$6.25M of revenue per employee
• Beta 2.53; P/S 22.8x; P/B 59.5x; debt/assets 48.8%
• ⚠️ Graham IV computes $42.64 vs a $417.80 price — void here, same buyback-hollowed-book artifact documented on GDDY (BVPS $7.03). Do not report it as a valuation signal.
• 2023 revenue of $1.84B is the games divestiture, not a decline — do not read the 2022→2023 line as shrinkage.

The structural lesson, which is not about APP: a name in Held — No Action is invisible to field sweeps, so the sections designed to reduce noise also reduce coverage. Any name parked there for longer than a couple of quarters should be pulled back through a sweep at least once. See [[pitfall-stale-entry-zone-suppresses-a-name]].

The print landed and the reaction was disproportionate — record it, don't resolve it. APP reported Q2 after the close on 2026-08-05 and fell −16% to ~$351, low $324.93 (−22%), below its $359.00 52-week low, on:

	Reported	Consensus	Gap
Revenue	$1.924B, +52.8% YoY	~$1.935B	−0.6%
Adj EPS	$3.76	$3.67	beat
Q3 revenue guide	$2.055-2.085B	~$2.068B	−0.6% at midpoint

A 0.6% revenue shortfall on 52.8% growth, alongside an EPS beat, removed ~$40B of equity value at the lows. That ratio — reaction to miss — is the fact worth keeping. It is the signature of a name priced for perfection rather than a name whose earnings broke, and it is the same shape as GDDY on 7/31 (−21.1% on a guide the market misread, subsequently analyzed and put in zone at [6.5]).

Forward multiple compressed 19.4x → ~16.3x in one session; ~30x FCF/share of $11.53. ⚠️ The $656.20 mean target predates the print and will be cut — not a usable anchor.

Still open: whether 52.8% growth is decelerating in a way the guide is conceding, or whether this is an expectations reset on an intact compounder. A 0.6% miss cannot distinguish the two — that needs the segment detail and the Axon / e-commerce split, which belongs in the /analyze file, not here.

History

• 2026-07-29 — swept in round 2.
• 2026-08-05 — APP added after the movers screen found it held, uncovered, and unscored. Sweep scope corrected: the July pass missed it because Held — No Action is not read.
• 2026-08-05 (post-close) — APP Q2 printed: −16% (low −22%) on a 0.6% revenue miss with an EPS beat. Recorded as a reaction-to-miss ratio, not as a verdict.