Marvell Stock (MRVL) Slides Into the Mid-$260s After Jensen Huang's Trillion-Dollar Call Sparked a 32.5% Rocket to $291

Marvell just lived a full market cycle in five trading days. It rocketed 32.5% in a single session to an all-time high near $291 after Nvidia's Jensen Huang called it "the next trillion-dollar company" on a Computex stage, then got dragged straight back down when Broadcom's earnings cracked the entire AI-chip trade. That's the whole story of MRVL in one sentence: the best the business has ever looked, the most stretched the valuation has ever been, colliding inside a sector-wide selloff. The stock is shedding more than 6% on June 5 into the mid-$260s, and the violence cuts both ways because that's exactly how a 1.99-beta momentum name trades when the narrative and the multiple go to war.

The Best Business, The Worst Timing

The setup is a paradox. Marvell's fundamentals are firing on every cylinder — record data center revenue, raised guidance, a $2 billion Nvidia investment, and the most valuable endorsement in the entire semiconductor industry. And yet the stock is falling, hard, because it priced all of that and then some before the chip complex rolled over. When a stock trades at 29 times forward sales and 47 times forward earnings, there's no cushion. Any wobble in AI sentiment gets amplified, and Broadcom handed the market exactly that wobble at the worst possible moment for MRVL.

This is the danger of buying a great company at a perfect price. The Huang call on June 2 turned Marvell into the market's purest AI-connectivity bet overnight, and the FOMO crowd chased it to a record. Two days later, Broadcom's mediocre guidance reminded everyone that even AI darlings have to clear ever-rising bars, and the same momentum that lifted MRVL 32.5% reversed on it. The business didn't change between the high and the selloff. The price did. That's the lesson the tape is teaching this week.

Where MRVL Trades Now

Put numbers on the move. MRVL closed near $284.27 on June 3, up roughly 258% over the prior year and more than 269% year-to-date, after touching an all-time high of $291.30. Its market cap swelled to around $254 billion at the peak. On June 5 the stock is down more than 6% alongside Micron and Super Micro as the AI-chip selloff bites, pulling it into the mid-$260s and trimming the market cap back toward the $230 billion zone. The 52-week range tells the story of the run: a low of $61.44 against that $291.30 high, nearly a five-bagger from trough to peak.

The stock sits miles above its own trend. The 50-day moving average is around $153 and the 200-day is near $100, so even after a 6% drop, MRVL trades at roughly 75% above its 50-day and close to triple its 200-day. That's a parabolic chart, and parabolic charts mean-revert violently when the catalyst fades. With a beta of 1.99, the stock moves nearly twice as hard as the market in both directions, which is why a sector selloff hits it harder than the index and a sector rally lifts it faster. This is not a name that sits still.

The Huang Endorsement Lit The Fuse

The rocket fuel came from the most powerful man in AI hardware. At Computex 2026 in Taipei on June 2, Nvidia CEO Jensen Huang shared a stage with Marvell CEO Matt Murphy and called Marvell "the next trillion-dollar company," praising its role in AI infrastructure and connectivity. The market's reaction was instant and simple: Huang blessed it, so buy it. MRVL ripped 32.5% on the day, then tacked on nearly 8% more the next session and roughly 4% after that, hitting back-to-back all-time highs as the endorsement turned into a self-reinforcing momentum spiral.

The endorsement landed on top of an already-strong setup. Murphy's keynote, titled "The Future of AI Scaling Depends on Connectivity," reframed Marvell as the company solving the bottleneck that comes after raw compute — moving data between chips fast enough to keep AI accelerators fed. Huang's "trillion-dollar" line wasn't just a compliment; coming from Marvell's most important partner and customer, it read like a signed check on the company's growth trajectory. The market treated it that way, and the stock went vertical.

Then Broadcom Broke The Trade

The party ended fast. On June 4, Broadcom reported earnings that left its full-year AI chip targets unchanged, and in a tape priced for relentless upward revisions, "unchanged" read as a miss. AVGO tumbled more than 12.5%, and the shock rippled across every AI and semiconductor name with a connectivity or custom-silicon story attached. Micron, Marvell, and Super Micro all came under pressure, with the selling extending into June 5 as the chip complex repriced two days running and spilled into Asian markets, where Korean chip names got crushed.

For Marvell, the timing was brutal. The stock had just printed an all-time high on pure AI euphoria, leaving it maximally exposed when the euphoria reversed. A name trading at 29x sales has the longest way to fall when sentiment cracks, and MRVL's 6%+ drop reflects exactly that vulnerability. The Broadcom news had nothing to do with Marvell's own fundamentals — it was a read-through on AI capex sentiment — but that's how sector selloffs work. When the trade breaks, the most expensive, most loved, highest-beta name in the group takes the hardest hit. Marvell was all three.

The Business Underneath Is Elite

Strip away the price action and Marvell's actual results are exceptional. The company posted record Q1 fiscal 2027 revenue of $2.418 billion, with data center now making up 75% of the total — a stunning shift from less than 10% of the business a decade ago and 50% just two years back. Operating cash flow hit $638.8 million in the quarter. This is a company that has remade itself into an AI-infrastructure pure play, and the transformation shows up in every line.

The growth trajectory is the headline. Marvell went from $2.3 billion in revenue in 2016 to $4.5 billion five years later, and now runs near $11.4 billion — roughly 2.5x growth in the most recent stretch alone. Management guided revenue to about $11.5 billion this fiscal year and $16.5 billion for fiscal 2028, the latter raised from a $15 billion guide issued just three months earlier. Interconnect revenue, the components that link chips together, is now expected to climb 70%, up from prior guidance of 50%. When a company is raising its outlook this aggressively this often, the underlying demand is real, not a story. The business deserves a premium. The question is how big a premium.

The Nvidia Tie That Binds

Marvell's relationship with Nvidia is its greatest asset and its sharpest risk, wrapped together. Nvidia made a $2 billion strategic investment in Marvell and the two are collaborating on NVLink Fusion, the technology that scales up AI systems by linking accelerators. That partnership is why Huang's endorsement carried so much weight — he wasn't praising a competitor, he was talking up a partner whose connectivity tech makes Nvidia's GPUs more powerful at scale. The connectivity bottleneck Huang described is precisely the problem Marvell is built to solve, which positions it to ride Nvidia's coattails through the entire AI buildout.

The flip side is dependence. A business this tied to one dominant partner inherits that partner's cycle, and when AI capex sentiment wobbles — as it did on the Broadcom print — names levered to the Nvidia ecosystem swing with it. The $2 billion stake aligns the two companies, but it also means Marvell's fortunes are increasingly hostage to hyperscaler spending decisions and Nvidia's roadmap. The blessing that sent the stock to $291 is the same linkage that makes it vulnerable when the AI trade gets questioned. Concentration cuts both ways.

The Connectivity Story

Marvell's edge is breadth across the connectivity stack. It leads in optical interconnect and silicon photonics, builds custom silicon — the XPU accelerators hyperscalers design for their own data centers — and makes high-end switches like the Teralynx T100, a 102.4 terabit-per-second chip that sits at the center of AI networking fabric. As Murphy put it, the company is still "in the early innings" of the actual buildout, and the addressable market backs that up: AI data center networking is forecast to exceed $100 billion by 2030, growing at a compound rate north of 30%.

That diversification across optics, custom silicon, and switching is what separates Marvell from single-product chip names. As AI clusters scale from thousands to hundreds of thousands of accelerators, the bottleneck shifts from compute to the plumbing that connects everything — and Marvell sells the plumbing. The bull case is that connectivity becomes the next great semiconductor land grab and Marvell owns a leading share of a market compounding at 30%-plus. If that TAM materializes, today's $11.5 billion revenue base looks small. That's the dream the stock was pricing at $291.

The Valuation Is The Whole Problem

Here's where the bears live. MRVL trades at roughly 29 times forward enterprise-value-to-sales and near 47 times forward earnings, multiples that sit far above typical semiconductor-sector medians. Its PEG ratio is around 1.0, which means investors are paying for every dollar of expected growth with no margin of safety — the entire thesis hinges on Marvell delivering the explosive growth the price already assumes. One independent fair-value model pegs the stock at a 321% premium to intrinsic worth. When the multiple gets this rich, the stock stops trading on results and starts trading on sentiment, and sentiment is fickle.

That's why the Broadcom selloff hit so hard. A stock at 29x sales has priced years of flawless execution into the present, so any crack in the AI narrative forces a violent repricing. If data center demand cools even modestly, or hyperscaler capex plans get trimmed, or a competitor takes share, the premium evaporates fast. The business can keep growing 20%-plus and the stock can still fall sharply, because the valuation isn't betting on growth — it's betting on growth accelerating beyond what's already a heroic guide. That gap between an elite business and a perfect price is the single biggest risk in owning MRVL here.

The Competition

Marvell doesn't have the field to itself. In custom silicon, it goes head-to-head with Broadcom, the far larger player whose own earnings just triggered this selloff — and Broadcom's scale advantage is real, with the resources to outspend Marvell on R&D. In the broader AI accelerator space, Nvidia is both partner and the gravitational center everyone orbits. In optical networking, Marvell competes against high-growth specialists like Lumentum and Coherent, both of which are riding the same connectivity wave. And across the AI-chip complex, it trades in lockstep with names like Micron and Super Micro, which is why all three fell together this week.

The structural risk that valuation bears flag is scale. Marvell is significantly smaller than Broadcom and Nvidia, which raises the question of whether it can invest enough to keep pace with peers that have multiples of its capital to deploy. The company's answer is focus — it doesn't try to do everything, it dominates the connectivity niche where its optics and custom-silicon expertise give it an edge. So far that focus has worked, with data center revenue exploding and guidance climbing. But in a capital-intensive industry where the giants spend freely, staying ahead requires Marvell to keep executing flawlessly while outspent. That's a high bar, and the valuation assumes it clears it every quarter.

Wall Street Is Split

The analyst spread tells you how uncertain the right price is. Stifel's Tore Svanberg holds the Street-high target at $321, raised from $230, on a Buy rating, valuing the stock at around 55 times fiscal 2027 earnings and citing the market's growing acceptance of Marvell's place in the AI supercycle. But the rest of the Street sits far lower. Oppenheimer lifted its target to $200 from $170 with an Outperform, Bank of America went to $200 from $125 with a Buy, and Goldman Sachs raised to just $180 from $125 while keeping a cautious Neutral. The consensus average target sits near $224 — below where the stock has been trading.

That's a remarkable picture: a stock that ran to $291 while the average analyst target sits at $224, meaning the market priced the shares above where most of Wall Street thinks they belong. The bulls and bears aren't arguing about whether Marvell is a great business — they agree it is. They're arguing about what you should pay for it, and the targets range from $180 to $321, a span of nearly 80%. When the analyst spread is that wide and the stock trades above consensus, you're looking at a name where price discovery hasn't settled and volatility is the only certainty. The most aggressive independent calls run to $400 and beyond; the most bearish valuation models scream overvaluation. Nobody agrees, which is itself the signal.

The Chart: A Parabolic Move Cooling

Technically, MRVL is a parabola taking its first real breath. The stock broke out hard after consolidating gains, blowing through every moving average on its way to $291.30, and now sits roughly 75% above its 50-day at $153 and nearly triple its 200-day at $100. Parabolic advances like this one are thrilling on the way up and treacherous on the way down, because there's no nearby support to catch a falling knife — the moving averages that would normally act as a floor are tens of percent below the price.

The 6%+ drop on June 5 is the first meaningful test of whether this breakout holds or rolls over. A high-beta name that's tripled off its lows can give back 20% or 30% in a sector selloff and still be in a long-term uptrend, because the gains were so large to begin with. The key question is whether buyers step in on the dip to defend the breakout or whether the air pocket below recent prices pulls the stock down toward its faster-rising moving averages. With beta at 1.99 and the AI-chip trade under pressure, expect outsized swings in both directions until the sector finds its footing.

The Chart: A Parabolic Move Cooling

Technically, MRVL is a parabola taking its first real breath. The stock broke out hard after consolidating gains, blowing through every moving average on its way to $291.30, and now sits roughly 75% above its 50-day at $153 and nearly triple its 200-day at $100. Parabolic advances like this one are thrilling on the way up and treacherous on the way down, because there's no nearby support to catch a falling knife — the moving averages that would normally act as a floor are tens of percent below the price.

The 6%+ drop on June 5 is the first meaningful test of whether this breakout holds or rolls over. A high-beta name that's tripled off its lows can give back 20% or 30% in a sector selloff and still be in a long-term uptrend, because the gains were so large to begin with. The key question is whether buyers step in on the dip to defend the breakout or whether the air pocket below recent prices pulls the stock down toward its faster-rising moving averages. With beta at 1.99 and the AI-chip trade under pressure, expect outsized swings in both directions until the sector finds its footing.

The Forecast

The base case is elevated volatility with a wide range, because that's what a stretched-valuation, high-beta name does when its sector is repricing. As long as the AI-chip selloff runs, MRVL stays under pressure and can keep giving back its Computex spike, with the mid-$200s as the near-term battleground and the consensus $224 target as a magnet below current levels. The business is strong enough that a stabilization in chip sentiment brings buyers back fast, but the valuation is rich enough that there's real downside if the selloff deepens.

The bear case is a continued AI-capex scare — more Broadcom-style guidance disappointments, hyperscaler spending worries, or a broader risk-off — that compresses MRVL's multiple from 29x sales toward something more normal, pulling the stock toward its 50-day at $153 if the parabola fully unwinds. The bull case rests on Marvell out-executing its own raised guidance: delivering on the $16.5 billion fiscal 2028 target, growing interconnect 70%, and proving the $100 billion networking TAM is real, which justifies the Stifel $321 target and the dream of trillion-dollar scale. The catalysts to watch are AI-chip sentiment day to day, the health of the broader semiconductor tape, and the next earnings report on August 20, which will test whether the fundamentals can grow into the valuation. For now, MRVL is the market's purest AI-connectivity bet caught in its first real gut-check — elite business, perfect price, and a chip selloff forcing the reckoning.

That's TradingNEWS