TK Edge Lab

Data-driven investing by TradingKey
Stock scoring • Strategy research • Market signals
Analysis, not advice

[TK Edge Lab @quantX02ge]

$NVDA reports earnings tomorrow.

What happens when NVIDIA enters earnings already under pressure?

Historically, when NVDA was trending lower in the 10 trading days before earnings, weakness was harder to reverse after the report.

In our sample, 60% of those events continued in the same direction after earnings.

That makes NVDA different from stocks where a strong report can quickly reset bearish sentiment.

For a high-growth stock like NVIDIA, falling into earnings may signal something deeper: the market is already lowering its expectations for the future growth path.

And once those expectations start moving down, simply delivering a “good” quarter may not be enough to reverse the trend.

The key question:

Can earnings reset expectations — or will they confirm what the market has already started pricing in?

$NVDA #NVIDIA #Earnings #AI

[TK Edge Lab @quantX02ge]

GS flagged a sharp momentum rotation last week:
- “12M vs 3M winner overlap” fell to multi-year lows
- The GS TMT Momentum Pair dropped ~18%.
- Options positioning showed increasing interest in Energy names such as $XOM and $GEV.

On Aug. 24, semiconductors ($NVDA, $AVGO, $MU ...) again led the decline. Stepping back to the big picture, we score energy the highest in the 5-layer cake, given that grid access and reliable power are becoming key constraints on AI data-centre growth.

Key Investment Theses under Energy:
✅ Execution: Power infrastructure lags data-centre construction, favouring projects with secured grid access and near-term delivery.
✅ Power Quality: Large AI loads increase demand for UPS, BESS and power electronics to ensure stability and load management.
✅ Value Shift: Near term, value favours equipment, storage and onsite power; longer term, firm generation and grid expansion.

What to Prioritise & Monitor within Energy👇

[TK Edge Lab @quantX02ge]

@StragglerLiu KRW40T buyback/cancel + >50% FCF return target is also a loud signal. Q2 net cash was KRW69.4T, while HBM4 mass shipments started in the quarter.

[TK Edge Lab @quantX02ge]

@MikeLongTerm @Meta META grew Q2 revenue 28% to $60.8B, but operating income fell 8%, and free cash flow was just $784M against $31.1B of capex. That cash-conversion gap is the real counterpoint to higher AI-driven earnings expectations.

[TK Edge Lab @quantX02ge]

@TrendSpider @DailyStockPick3 @leadlagreport @Sam_Badawi @willrhind NVDA guided Q2 FY27 revenue to $91B ±2% and GAAP gross margin to 74.9% ±50bp.
Q1 Data Centre revenue already hit $75.2B, up 92% YoY, so margin execution matters.

[TK Edge Lab @quantX02ge]

@zerohedge Reuters says Broadcom could raise more than $60B, potentially reaching $100B, through an SPV. The interesting part is how much AI infrastructure risk gets pushed into project-level financing.

[TK Edge Lab @quantX02ge]

@pequityresearch Morgan Stanley still has AMD at 240k CoWoS wafers for 2027, with MI455/MI450 shipments around 1.5M units. TSMC is also moving 5.5x-reticle CoWoS into volume production in 2026, so the mix shift may partly reflect capacity expansion.

[TK Edge Lab @quantX02ge]

@WOLF_Financial Druckenmiller sold; NVIDIA just authorized another $80B of buybacks. Different smart-money clocks, btw. Wednesday’s guide probably settles the argument better.

[TK Edge Lab @quantX02ge]

@TradexWhisperer MU already locked price and volume for all 2026 HBM supply, and sees HBM TAM reaching ~$100B by 2028. Contracted demand tells more.

[TK Edge Lab @quantX02ge]

@wliang Last quarter NVDA beat its revenue midpoint by $3.6B and still kept non-GAAP GM at 75.0%. If the stock dips, I’d care about the next guidance than the first candle.

[TK Edge Lab @quantX02ge]

@DarioCpx Q1 printed $81.6B vs a $78B guide midpoint; Q2 midpoint is now $91B. Guidance quality is a test.

[TK Edge Lab @quantX02ge]

$NVDA reports earnings next week.

One historical pattern is worth watching:

When NVIDIA was already trending higher in the 10 trading days before earnings, that momentum often continued after the report.

Among the Mega 7, NVDA stands out as a momentum-confirmation stock around earnings.

The reason may be simple: for NVIDIA, earnings aren’t just about whether this quarter beats expectations. The bigger question is whether the results and guidance give investors another reason to raise their assumptions for the quarters ahead.

So heading into next week, the key question isn’t just:

Will NVIDIA beat?

It’s:

Will earnings confirm — or break — the trend the market has already been pricing?

$NVDA #NVIDIA #Earnings #AI

[TK Edge Lab @quantX02ge]

@MicronCEO @jimcramer MU’s strategic agreements are multi-year take-or-pay with enforceable volumes. Most use fixed prices or price bands. That changes the fab math.

[TK Edge Lab @quantX02ge]

@GeorgeGammon May TIC: foreign official flows were -$39.9B, versus +$172.0B from private investors. Foreign flow data is less monolithic.

[TK Edge Lab @quantX02ge]

@epictrades1 MU’s Q3 DRAM sales rose 67% QoQ, driven by low-60% ASP growth and only low-single-digit bit growth. Pricing did most of the lifting.

[TK Edge Lab @quantX02ge]

@StockMKTNewz MU’s FQ3 adjusted FCF hit $18.3B, up from $6.9B in FQ2. Cash generation is the signal I’d keep higher on the dashboard

[TK Edge Lab @quantX02ge]

$MRNA +177% today.
This is the first positive Phase 3 readout for a personalised mRNA cancer therapy, but melanoma alone cannot justify a move like this, and it might not be a good time to take either side now.

What the market is really pricing in is $MRNA’s entire oncology platform: the current price implies that personalised mRNA + Keytruda can eventually be replicated across lung, kidney, and bladder cancer, as well as a range of other tumour types.

As of today, though, melanoma is still the only indication with a positive Phase 3 readout.

There are actually two forces layered into this green candle.
✅One is a genuine platform re-rating
✅The other one is a purely mechanical short squeeze. $MRNA was already one of the most heavily shorted large-cap stocks (ORTEX estimates short interest at around 13.5% of free float, with shorts down roughly $4.8B mark-to-market intraday), and forced covering and a fundamental catalyst were compressed into the same day, making it very hard in real time to separate how much of the move came from each.

So I’m taking neither side.
✅There is no near-term data point that can immediately invalidate the platform story, and shorting a stock simply because it is “too expensive” is dangerous when the catalyst is real, and shorts are actively being squeezed
✅At the same time, the current price has already prepaid for the assumption that “other tumour types will work too,” and that part of the valuation will ultimately have to be justified by subsequent cross-tumour data.

What I’m mainly watching next is the price-volume relationship:
✅If volume falls materially while the price holds near the highs, that suggests new long-only capital is accepting this platform valuation.
✅If volume falls and the price also retraces quickly, that suggests today’s clearing price was still largely a squeeze price.

In the short term, the squeeze is fading while RSI is already above 90. The real battleground comes down to two things: whether this valuation can hold once short covering runs its course, and whether the next cross-tumour readout can support the expectations that were pulled forward into today’s price. If that readout falls apart, you have to move fast.

[The Kobeissi Letter @KobeissiLetter]

BREAKING: Moderna stock, $MRNA, extends gains to +212% on the day in after hours trading following positive phase 3 results of first the mRNA treatment to prevent cancer.

The stock has now added over +$50 billion in market cap today.

[TK Edge Lab @quantX02ge]

AWS grew 37% in Q2, its fastest in ~18 quarters; backlog jumped ~2.5x YoY to $496B, and AI plus custom silicon each cleared a $25B run rate at a 39.4% segment margin. The other side: capex was lifted to $220B, trailing FCF flipped to -$7.6B, and the Q3 revenue guide landed below the Street.

[TK Edge Lab @quantX02ge]

Today's 78,756 caps a steady accumulation: ARK also took 106,941 on Aug 13 and 35,000 on Aug 18, buying through a ~29% drawdown from the IPO. The thematic read: she's stepping around memory like MU and Hynix, betting Cerebras and Groq cut HBM dependence in inference. CBRS still runs a 40.6% gross margin and an adjusted loss, though.