Signal hierarchy: X trade > Portal memo > Cockpit. Sep 8: IFX sold — AJ exited Infineon Sep 7 and is not a buyer at €58.58. ASML added at €1,652 on Sep 1 (avg €1,689), now +5% and being held. Portfolio at a new all-time high pre-market. Catalysts: ORCL Thu Sep 10, Aug CPI Fri Sep 11.
Sources: X @alojohhardcore Aug 27–29, Portal weekly Aug 22, NVDA memo Aug 27, MRVL memo Aug 28, XPEV memo Aug 24, COHR initiation Aug 25, Cockpit Aug 28.
The reaction, not the print, is the investment case. Revenue rose 85.5% year over year to $29.6B, adjusted EPS rose 96.2% to $3.32, and operating margin expanded 237bps to 67.9%. The stock rallied 3.4% to $379.69 — then surrendered the entire move, trading down to $361.20. AJ reads this as a market treating earnings compounding at ~100% as "borrowed rather than earned."
A multiple dislocation, not an estimate dislocation. This is the analytical hinge: AJ's FY2027 estimate of $18.48 sits 6% below the consensus average of $19.57. He is not asking anyone to accept a more aggressive forecast than the sell side already carries — the street is higher, and the equity still trades at 20.9x. "The market is not disputing the estimates. It is declining to capitalise them."
The gross-margin bear case is the wrong debate. Gross margin fell 336bps to 75.0%, which is what the market reacted to. But operating margin expanded in the same period — the compression is mix (custom accelerator silicon vs 90%+ margin software), not price. Broadcom added ~$13.6B of quarterly revenue while adding almost nothing to its cost base; incremental operating margin was ~71%, higher than reported gross margin.
The inversion frames the downside. For a buyer at $361 to make nothing over twelve months, the forward multiple must compress from 20.9x to 13.5x — a level unseen since before AI revenue was material. Implied PEG at that point: 0.24x. "The bull case requires the market to pay what it has recently paid. The bear case requires the market to pay what it paid before the growth existed."
Track record earns the forward numbers a hearing. Three consecutive quarters of materially better accuracy than consensus — 0.1% error last quarter (consensus 2.3%), 0.10% the quarter before (consensus 1.64%), 1.0% before that (consensus 2.4%). AJ is careful to note this validates near-term order-flow modelling, not a multi-year forecast.
AJ's highest conviction name. FY2028 EPS $16.16 is 23% above Street at $13.13. Prior forecast track record 0.5% error. Price corridor implies $371–$420 from $221 entry (68–90% upside).
X confirms conviction. Pre-earnings: "high chance of ripping/rerating." Post-earnings: "Bullseye." Guidance "so crazy it could reset the entire sector." Next ER Nov 19, buy point ~Oct 30.
Called this drop perfectly. Pre-earnings: "near bottom of cockpit." Stock fell 10.3%. Post-ER: corridor won't move much. Trading at 37.1x vs median 25.4x — only accumulate near $166.
NEW COVERAGE. 65% 12-month upside. Optical networking and AI datacom play in the best Cockpit quadrant (Upside + Inexpensive).
Deep value China EV with VW optionality. AJ 91% below consensus yet sees 39–90% upside. VW partnership ~50% of gross profit.
Contrarian DRAM thesis anchored to NVDA earnings. "Reverse X feed undefeated." ~80% of NVDA revenue needs DRAM. MU up 4% pre-market confirming thesis.
Long-awaited entry into semi equipment. Had post-Q2 memo ready. Waited for pullback. "Been waiting for this one for quite a while."
Rankings from X trade alerts Aug 27–29. Exact rankings not disclosed — order reflects trade activity and commentary. All estimates.
| # | Ticker | Signal | Action | Context |
|---|---|---|---|---|
| 1 | NVDANVIDIA | BUY | Core Hold | "Bullseye" on earnings. 12M $371–$420. FY28 EPS 23% above cons. |
| 2 | MUMicron | BUY | Added $910 | "Added meaningfully." 80% of NVDA rev needs DRAM. |
| 3 | ASMLASML Holding | BUY | Added More | Added alongside AMD. EUV lithography monopoly. |
| 4 | AMDAdvanced Micro Devices | BUY | Added More | Added alongside ASML. #2 AI GPU. Benefits from NVDA guidance. |
| 5 | AMATApplied Materials | NEW | New $466 | "Been waiting for this one." Semi equipment = pickaxe play. |
| 6 | COHRCoherent Corp | INITIATE | New Coverage | $275 → $454 (+65%). Optical AI play. FY28 EPS 17% above cons. |
| 7 | XPEVXPeng | BUY | Held | 39–90% upside at $11.36. VW = 50% GP. |
| — | METAMeta | SWING | Exit $588/$598 | Swing trading vehicle. Buy $569 → sell $588. Margin trades. |
| — | MRVLMarvell | HOLD | Watch Only | Called -10.3% drop. Accumulate near $166 only. |
| — | EOSEos Energy | CLOSED | Short Closed | "Rotten company, aka a scam." Closed at $3.26. |
"—" = swing or watch-list without confirmed sizing. Core #1–#7 estimated from activity.
Corridor = NTM EPS × P/E range. Green zone = attractive. Red zone = stretched. Bar shows price position within corridor.
Each cell shows quadrant assignment. UI UE DE. Highlighted = changed from prior day.
AJ Investment Research Cockpit & Corridor replica. Corridors = AJ NTM EPS × observable P/E multiple, reconciled against AJ's Cockpit → Details view so the ±1.5σ/±1σ/median bands match his published numbers. Each name is quoted in its own listing currency, as AJ does: € for ASML/Infineon/Aixtron, HK$ for Xiaomi, ₩ thousands for Samsung/SK Hynix, $ otherwise. PEG = Fwd P/E ÷ EPS growth rate. Green = AJ est; blue = consensus. Not investment advice.
"Currently in pre-market we are at another new all-time high. Not posting this to brag. It's more an indicator of travel for the type of portfolio we have." The AI-semi concentration is working.
But note how that came out. In explaining why he is holding ASML, AJ wrote: "we won't be selling (except if we do another overall portfolio trim like last week)." A portfolio-wide trim is a material sizing event, and it was never posted to X as its own alert — it surfaced only as an aside a week later. The ASML add itself is the same story: it went out on Sep 1 and only became visible here as a quoted screenshot on Sep 8.
This is the app migration biting. AJ said on Sep 7 that the Member App now carries all alerts and that "X is already optional". These two events are the first concrete evidence that X-only tracking now misses real trades. Treat the X feed as a lagging, partial mirror — and treat an absence of alerts here as weak evidence that nothing happened.
Worth noting for anyone tracking AJ through X: he now says the AJ Member App carries everything the feed does — all trade alerts, plus posts and threads — and argues it is the better channel: more reliable (subscribers report X surfacing posts a day late), easier to search, easier to use. His words: "X is already optional."
Combined with the instant trade-alert emails launched Aug 30, the practical implication is that X is becoming a lagging mirror rather than the primary feed. A trade could appear in the app or an email well before it shows up here. Treat an absence of X alerts as weaker evidence than it used to be.
S&P 7,719 (+0.1% wow), forward P/E 19.3x (−0.2x). The index now sits below its three-year median and average and below the −1.0 standard deviation band. AJ's framing: buyers are paying near a valuation low relative to the past 30, 90 and 180 days and the past two years. On a ten-year lookback 19.3x still screens slightly rich vs a 19.0x average, but he argues the shorter window is the relevant one — leaning on decade-old multiples "would have kept investors out of the market."
Nasdaq +0.4% on the week and now +3.1% over three months — a notable flip from −2.1% a week earlier. Top: Micron +9%, Meta +7%, Nvidia +6%. Bottom: Palantir −6%, with Amazon, Broadcom and Microsoft all −3%.
Corridor entry rankings — Broadcom sweeps. AVGO is now #1 best entry on both the 1-year and the 90-day observable multiple, ahead of Amazon and Meta (1Y) and Alphabet and AMD (90D). Least attractive on both: Palantir, with Oracle and Microsoft filling out the bottom.
Long-term upside ranks Tesla first — but AJ appends his December 2025 SpaceX warning — then Nvidia and AMD. Least: Apple, Microsoft, then Meta/TSMC. PEG: cheapest growth is Micron, Nvidia, Broadcom (the last flagged for revenue-concentration worries, "in particular in Anthropic"). Most expensive: Apple, Tesla, then Alphabet/Palantir.
AJ sharpened the CPI thesis into a single clean claim. Consensus for August is 3.4%; July printed 3.4%. If consensus is right again, headline inflation is unchanged — and because he has already argued the energy contribution will be larger this month on the 2025 base effect, an unchanged headline necessarily implies easing pressure from non-energy items.
His words: "That's good news vis-a-vis stocks and interest rates." The setup is that the market is reading the oil print as an inflation problem when the arithmetic points the other way. August CPI lands Friday, September 11.
Seasonal caveat. Over the last 12 years the week after Labor Day has been up 5 times and down 7, with an average return of −0.5% and a median of −0.7%; the extremes were +3.6% (2022) and −4.2% (2024). AJ expects "a bit of directionless trading Tuesday/Wednesday" ahead of the data.
The Cybercab launch event boosted TSLA volume but far below prior launch levels. AJ's read: "Enthusiasm has objectively cooled for this stock." At Friday's $351 close, Tesla trades at a level first reached in October 2021 — roughly five years of no price progress. The weekly separately notes the pattern: front-running into the event, then the typical sell-the-news reaction after a disappointing one.
On management guidance he is blunt: "Since 2021, none of the financial guidance Elon ever provided turned out to be accurate. By now, he completely stopped providing any financial guidance. Why? He doesn't have to anymore."
Hold this against the corridor. Tesla still screens as the most attractive on long-term upside in the weekly — but AJ flags the Corridor Method as less reliable for TSLA than for any other name, driven by SpaceX-related distortions, and points back to his December 2025 SpaceX warning. Treat the upside ranking as a methodology artifact, not a recommendation.
AJ spelled out the methodology behind the Cockpit's y-axis for the first time. His growth term is NTM+1 earnings ÷ NTM earnings — not this year's growth, but the growth the market expects over the twelve months starting a year from now.
The reasoning: "Investors should not focus on the next twelve months. Investors should focus on what investors believe the next twelve months in twelve months can deliver." Forward P/E already captures NTM; pairing it with NTM growth would double-count the same window. Rolling NTM intervals are used throughout rather than calendar years, since what matters is a consistent 1-year period, not January-to-December.
This matters for reading the Cockpit: a low PEG on AJ's definition means the market is cheaply pricing growth that is still two years out — which is exactly the AVGO argument below.
AJ reframed AVGO's valuation as a hurdle rather than a target. On his updated post-guidance estimates, for the stock simply not to rise from $361, Broadcom's forward P/E would have to compress from 20.9x to 13.5x. That is the bar the bear case has to clear.
PEG cut from 0.96 to 0.38x. Cheap growth on AJ's NTM+1 definition. And if the multiple did slide to 13.5x by September 2027 — the flat-stock scenario — the implied PEG at that point would be 0.24x, which he calls "extremely cheap."
The honest caveat. Asked what the market might discount, AJ conceded the outlook is "very FY2028 loaded" — the earnings carrying the thesis are far out. His counter is management credibility: "Hock Tan is credible and we just have to wait and see till the market starts to price it in."
Price action tells the same story. AVGO popped 3.4% on the print before profit-taking erased it. Over the last twelve months it has outperformed the S&P by just 0.5pp — despite a 12pp drawdown on Q2 — and has twice built and then surrendered large outperformance. Subscribers noted it now screens 1st in the 90-day Cockpit for tactical entry.
The oil scare is misread. "Oil is freaking investors out" — but the accurate framing is that oil is worrying investors that other investors would be selling. August 2026 delivered oil at $83, comparable to June's $82. In June, oil contributed 86bps (0.86pp) to headline CPI.
The base effect is what matters. June 2025 average oil was $67; August 2025 was $64 — $3 lower. So August 2026 compares against a lower 2025 bar than the June 2026 print did. The August 2026 oil contribution to CPI will therefore likely exceed June's 86bps.
Why that is bullish, not bearish. July CPI printed 3.4%. Consensus (incomplete) points to another 3.4% for August. If headline holds flat while the energy contribution rises, the arithmetic implies non-energy inflation drivers actually eased — a genuinely positive signal that the market is currently mispricing as an oil problem.
AJ's read: "I don't want to sound too bullish but there is a reasonably likely pathway for a favorable development next week." August CPI drops Friday, September 11. Separately, AJ reposted his May thesis: the administration stays focused on pulling pump prices down ahead of the Midterms.
Credit signal is broadly reassuring for US equities. IG spreads unchanged at 79bps, BBB at 98bps, broad HY tightened 14bps to 263bps — near bottom decile since financial crisis. All-in cost of debt fell for both cohorts (IG 5.43→5.38%, HY 7.19→6.96%). August was easing, not tightening.
Lone deterioration: CCC. Widened 40bps to 1,031bps — through the 1,000 distress threshold, fourth consecutive widening at ~24bps/mo. Single B minus BB actually compressed 5bps, meaning stress has not migrated up the quality curve. Contained default cycle in weakest decile, not a systemic tell.
Escalation trigger to watch: BB spreads through 200bps (currently 153). That is the single level that flips the equity call. Everything else on the page is secondary.
Actionable splits: Screen leveraged equity holdings by rating & maturity wall, not by leverage ratio. Danger zone: net debt > 5x, rated CCC/B-, maturities in 2027–28. Best positioned: profitable mega-cap with strong balance sheets. Small caps face headwind from CCC minus BB quality spread widening 53bps.
AJ's biggest macro call: NVDA guidance "so crazy it could reset the entire sector." AI infrastructure spend accelerating faster than anyone modeled. Adding aggressively across the AI semi complex.
"About 80% of the revenue Nvidia guided for can't happen without DRAM." MU is the direct derivative play. "Reverse X feed is undefeated" — when X is bearish, AJ buys.
AJ launched instant email alerts — subscribers now get an email within minutes of an opening, adding, trimming, or exiting a position. No more waiting to catch X posts. Trades still open in the AJ Member App.
PEG ratio added alongside Corridor Method. Current: 17 stocks in Upside+Inexpensive, 8 in Upside+Expensive, 1 in Downside+Expensive (MSFT only). Zero value traps. Rivian added to Details view coverage.
Multiple META swing trades: buy $569 → sell $588, exit $598 "all on margin." Tactical income overlay on core AI positions. 3–7% returns in 1–3 days on high-liquidity names.
Sources: AJ portal (ajinvestmentresearch.com) and X @alojohhardcore, Aug 22 – Sep 7. Weekly Sep 6, AVGO post-Q3 memo Sep 3, US Credit Update Aug 30, CPI thesis Sep 2 & Sep 6, TSLA/Cybercab Sep 6, ORCL Sep 10 beat call Sep 3, NVDA memo Aug 27, MRVL memo Aug 28, XPEV memo Aug 24, COHR init Aug 25, Cockpit Sep 8 (Overview + Details, both lookbacks), AJ Dashboard (stale — last updated Apr 26), Companies coverage Sep 8, X trade alerts Aug 27 – Sep 7 (IFX sold Sep 7, ASML added Sep 1). A full capture of AJ's four surfaces as of Sep 8 is archived at AJ Snapshot · Sep 8. Sourcing caveat: AJ has moved alerts to his Member App and calls X "optional" — the Sep 1 ASML add and a portfolio-wide trim both reached X only indirectly, so absence of an alert here is weak evidence that nothing happened.