Marvell's huge 2028 AI forecast just wowed the street — buy the post-earnings momentum
Guidance raises are one of the most reliable momentum triggers in stocks: when a company formally lifts its multi-year revenue outlook, analysts raise their models and price targets over the following weeks, not overnight. Marvell didn't just give a big 2028 number — it convinced the street its customer mix is genuinely diversified, which reduces the single-customer risk that has capped similar AI chip names. That combination of a higher forecast and a better narrative tends to produce follow-through buying after the initial pop fades. Trading the post-announcement consolidation breakout captures the institutional accumulation phase rather than the headline spike.
Idea
Guidance raises are one of the most reliable momentum triggers in stocks: when a company formally lifts its multi-year revenue outlook, analysts raise their models and price targets over the following weeks, not overnight. Marvell didn't just give a big 2028 number — it convinced the street its customer mix is genuinely diversified, which reduces the single-customer risk that has capped similar AI chip names. That combination of a higher forecast and a better narrative tends to produce follow-through buying after the initial pop fades. Trading the post-announcement consolidation breakout captures the institutional accumulation phase rather than the headline spike.
Advanced Analysis — institutional-depth research report
Verdict: solid thesis, but the entry door isn't open yet
The idea's momentum logic has real support: Marvell's latest quarter (ended August 1, 2026) grew revenue 13.3% quarter over quarter to $2.74B with net income jumping to $308M, and Broadcom's fiscal Q3 posted revenue up 33.4% to $29.6B with free cash flow of $13.7B — the strongest point for the trade. The strongest point against is that provisional September 30, 2026 ownership filings show net open-market insider selling at both companies (roughly -$513.6M at Broadcom across eight holders and -$23.8M at Marvell across five), directly countering the institutional-accumulation thesis. The backtest evidence is also narrow: only the nine-month Marvell window traded (49 trades, 12.6% return, 55% win rate, 8.6% max drawdown), the 12-, 24-, and 60-month windows failed to run on a data gap, and exits were filled on trigger bars rather than intrabar. Technically, neither leg has triggered — Marvell's ADX (22.2) clears 20 but on-balance volume has no live reading, and Broadcom's ADX (16.4) sits 3.6 points below 20 while RSI readings of 79.4 and 70.8 are already stretched. The verdict flips if the next ownership refresh firms up buying instead of selling, or if OBV confirms above zero on Marvell while Broadcom's ADX closes above 20. Until then, the disciplined answer is to wait for the setup's own confirmation rather than chase an overbought tape.
Trade now: close, but not triggered — here is what still has to line up
The strategy is long-only on Marvell and Broadcom, and as of the latest daily data (computed October 6, 2026) neither leg has fully triggered. Marvell closed at $287.01 and is the closer of the two: its 14-day ADX of 22.2 is above the required 20, and its 14-day RSI of 79.4 clears both the above-50 and above-45 tests. The one unconfirmed condition is on-balance volume above zero, which currently has no live reading — so Marvell is one confirmation away. Broadcom closed at $375.81 with a 14-day RSI of 70.8, but its ADX of 16.4 is 3.6 points below the 20 threshold, so the trend-strength condition is not there yet for either pair. The completed backtest window on the Marvell leg traded 49 times over 9 months, returning 12.6% with a 55% win rate and an 8.6% maximum drawdown — that is the evidence base for waiting rather than forcing the trade. Risk framing once an entry triggers: position sizing is fixed-risk at 2.4% of the account per trade with a hard 2.4% stop and a 4.8% take-profit, a 2:1 reward-to-risk profile, capped at 25% of the portfolio per name. Chart-based exits layer on top: Marvell's nearest support sits at $280 and nearest resistance at $290 (close $287.01), while Broadcom's nearest support is $370.35 and resistance $380 (close $375.81). The signal-based exit fires if the 14-day RSI drops below 40, or if a position is held past 75 trading days. Note that no robust parameter setup was established — variants could not be evaluated due to a data gap — so the published rules are used exactly as frozen. What "wait" means concretely: do nothing on Broadcom until its 14-day ADX closes above 20; do nothing on Marvell until on-balance volume confirms above zero while the ADX stays above 20 and the RSI holds above 50. Chasing now means buying Broadcom with trend strength still building (its 20-day and 200-day moving averages are at $373.03 and $367.93, so price is barely above both) and paying up on a Marvell whose RSI of 79.4 is deeply overbought. The invalidation is mechanical: if either leg triggers and then closes below its nearest support ($280 for Marvell, $370.35 for Broadcom) or hits the 2.4% stop, the idea is out for that leg.
placeholder unused
placeholder unused
placeholder unused
Start with the ownership posture, because it cuts directly against the institutional-accumulation narrative. As of the provisional September 30, 2026 filings, insiders at both companies show net open-market selling — roughly -$513.6M at Broadcom across eight reporting holders and -$23.8M at Marvell across five. The thesis assumes patient buyers building positions while analysts mark up models; the filings show the people with the best information were net sellers into the news. These ownership reports are marked provisional, so the picture could firm up or soften in later amendments. Marvell's fundamental base is far thinner than the headline story suggests. Yes, net income hit $308M in the latest quarter, but free cash flow actually fell 0.9% quarter over quarter to $479M and operating cash flow dropped 5.2% to $606M. The dividend history tells its own story: the annual payout was cut from $0.24 in 2025 to $0.18 in 2026 and the quarterly rate is a token $0.06, with zero growth over the tracked period. A company with a 2028 revenue narrative but…
Scores
- Conviction score breakdown: 55
- Thesis support: 70
- Trade readiness: 40
- Risk quality: 55
- Backtest evidence: 45
- Fundamentals trend: 65
Watch items
- MRVL — On-balance volume (OBV)
- MRVL — ADX (14)
- MRVL — RSI (14)
- AVGO — ADX (14)
- AVGO — On-balance volume (OBV)
- AVGO — RSI (14)
- MRVL — Close vs nearest support
- AVGO — Close vs nearest support
- AVGO — Insider net open-market activity
- MRVL — Insider net open-market activity
- MRVL — Next quarterly report
- AVGO — Next quarterly report