Most of the market right now is obsessing over oil above $100 and rate-hike fears, but Broadcom handing investors a 2028 roadmap tells you its custom AI chip pipeline is contracted far into the future, not dependent on this quarter's mood. Side-by-side co

Verdict: a strong thesis still needs its entry — wait for AVGO to confirm

The idea's core claim — that Broadcom's custom AI chip pipeline is contracted into 2028, per the Yahoo Finance piece on the company's 2028 guidance — is backed by real numbers: fiscal 2025 revenue up 23.9% to $63.9B, a 39.9% operating margin (98th peer percentile), and $26.9B of free cash flow (99th percentile), with the latest quarter showing revenue of $22.2B and record net income of $9.3B. Marvell's fiscal 2026 added support with revenue up 42.1% to $8.2B and a swing to $2.7B of net income, though its earnings are lumpy (diluted EPS fell 4.0% even as revenue surged). The strongest point against: ownership filings for the period ended June 30, 2026 show net open-market insider selling at both names — roughly $283.3M at AVGO across 23 holders and $19.5M at MRVL — while the technical entry triggers are not yet live, with AVGO's RSI at 43.5 below the required 45 and its close of $360.83 still 2.9% below the 21-day EMA at $371.31. The completed backtest gives the rule set credibility (6 trades in 60 months, 66.7% win rate, 62.0% return, 17.1% max drawdown), but with six trades in five years and exit fills on daily bars, each trade carries heavy weight and fills are approximate. A fresh ownership filing showing insider buying at either name, or AVGO confirming all three entry conditions on the same daily bar, is the fact that would flip this from watch-list to actionable. Until then, the disciplined move is to wait — the setup is defined, but it isn't on.

Trade now: close, but not triggered — the AVGO setup needs one more day of strength

Neither entry rule is live today, so the concrete answer is: wait, with defined levels in hand. AVGO closed at $360.83, about $10.48 below its 21-day EMA at $371.31 — so the pullback-then-reclaim condition is only half done (price is below the EMA; it needs to close back above it). The 14-day RSI at 43.5 still needs to cross above 45, roughly 1.5 points away, and MACD must cross back above its signal line. MRVL is in the mirror position: $226.96 already sits above its 21-day EMA at $223.07, but with RSI at 53 it is 8 points past the cross threshold, so that trigger is not in play today. Once an entry fires, risk is tightly boxed: the fixed stop sits at a 2.4% loss on the position and the fixed take-profit at a 4.7% gain, roughly a 2:1 reward-to-risk, with position sizing capped at 25% of the book and fixed-risk sizing at about 2.4% equity risk per trade. A secondary stop closes the trade if price crosses back below the rank-2 support level, and a secondary target sells into the nearest resistance — for AVGO that is $360.20 (already nearly touched) with support at $357.11 and $350.03; for MRVL resistance sits at $230.43 with support at $220. What does waiting mean concretely? Do nothing until AVGO prints a day where the low dips under the 21-day EMA while the close finishes above it, the 14-day RSI crosses above 45, and MACD crosses its signal — all on the same daily bar. The same checklist applies to MRVL, priority two. The evidence base here is the completed backtest: over 60 months the pair produced 6 trades, a 66.7% win rate, a 62.0% return, and a worst drawdown of 17.1%; the 24-month window delivered 4 trades at a 75% win rate. That history supports using these exact live levels rather than improvising entries — the setup is defined, it just isn't on yet. One caution on process: the parameter-sensitivity evaluation ran out of its time budget, so no robust setup was established nearby; trade the published rule set as written rather than assuming a tweaked version would be better.

The bull case: contracted AI demand meets elite margins

The core of the idea — that Broadcom's custom AI chip pipeline is contracted into 2028, per the Yahoo Finance piece on the company's 2028 guidance — is exactly what the latest fiscal year numbers show. For the fiscal year ended November 2, 2025, Broadcom grew revenue 23.9% to $63.9B, posted a 39.9% operating margin (98th percentile among 854 Information Technology peers), and generated $26.9B of free cash flow (99th percentile among 791 peers). Those are not hype-cycle numbers; they are the profile of a business converting AI demand into cash at scale. The trend has continued into fiscal 2026. Quarterly revenue stepped from $14.9B in the quarter ended February 2, 2025 to $22.2B in the quarter ended May 3, 2026, with operating margin reaching 48.6% in that latest quarter and gross margin rising to 69.5%. Net income of $9.3B in that quarter is the strongest in the company's reported series. Momentum like this makes the 'durable business, not a headline chase' framing credible. The Marvell comparison supports the idea's second leg. Marvell's fiscal year ended January 31, 2026 showed revenue up 42.1% to $8.2B and a swing to $2.7B of net income from an $885M loss the prior year, with free cash flow of $1.4B (98th percentile among peers). The custom-chip niche is generating real, growing profit at both companies — consistent with the idea's read that this is a durable franchise, and per the Yahoo Finance Broadcom-vs-Marvell comparison piece, it is being treated that way by the market. The trade design matches the thesis. Over 60 months of daily bars, the…

Scores

• Conviction score breakdown: 64
• Thesis support: 78
• Trade readiness: 45
• Risk quality: 60
• Backtest evidence: 55
• Fundamentals trend: 82

Watch items

• AVGO — RSI (14)
• AVGO — Close vs 21-day EMA
• AVGO — MACD (12,26,9)
• AVGO — Close vs rank-2 support
• MRVL — RSI (14)
• MRVL — Close vs 21-day EMA
• MRVL — Close vs support
• AVGO — Insider net open-market activity
• MRVL — Insider net open-market activity