Report
No Data Yet
Corning Inc. fell more than 8% on Wednesday, extending its decline from a record high as a broad selloff swept through US optical communications stocks. "Corning is testing a historically bullish trendline at its 80-day moving average, a setup that has preceded higher prices one month later 72% of the time over the past decade," said Rocky White, senior quantitative analyst at Schaeffer's Investment Research. The stock closed at $174.41, down 7.1%, after touching an intraday low near $171. Marvell Technology dropped more than 6%, while Lumentum, Coherent and Credo each fell over 5%. The selloff occurred even as the broader market advanced, with the S&P 500 rising 0.38% to 7,572.40 and the Nasdaq Composite gaining 0.62% to 26,269.23. An institutional trader purchased weekly put contracts on Tuesday, betting on further downside ahead of the company's second-quarter earnings report scheduled for July 28. The pullback has erased more than $35 billion in market value since the stock peaked on June 30, when its trailing price-to-earnings multiple had surged past 90 times — well above its five-year historical median. Chief Executive Wendell Weeks sold 100,000 shares at an average price of $186.46 in June, generating $18.6 million, while Senior Vice President Soumya Seetharam sold 20,000 shares at about $206 in May. Combined insider sales exceeded $54 million with no corresponding insider purchases, according to filings. Wall Street remains broadly bullish, with 10 analysts rating the stock a buy and six neutral, and a consensus price target of $194.69 implying about 12% upside from current levels. The optical communications sector's decline comes as investors reassess valuations following an AI-driven rally that lifted Corning and its peers to record levels. Corning, a key supplier of fiber optic and optical connectivity solutions for data centers, had benefited from enthusiasm around artificial intelligence infrastructure spending. However, the stock's trailing P/E multiple above 90 times created vulnerability to any negative signals. Short interest in Corning fell 13.7% during the most recent reporting period, representing 2.6% of the float, or less than two days' worth of trading volume. Wall Street expects Corning to report core earnings per share of $0.75 to $0.76 on revenue of about $4.60 billion for the second quarter. The company beat first-quarter estimates in April, delivering EPS of $0.70 on revenue of $4.34 billion, up 18.1% year over year. This article is for informational purposes only and does not constitute investment advice.
**Global AI revenue from hyperscale cloud providers has exceeded data center depreciation costs for a second straight quarter, a milestone that bolsters the economic case for the industry's $200B-plus annual infrastructure buildout.** Global AI revenue from hyperscale and emerging cloud providers reached $25 billion in the most recent quarter, surpassing estimated depreciation costs of $21 billion tied to data centers and chips, according to a report Wednesday from research firm Exponential View. The gap marks the second consecutive quarter that AI-generated revenue has covered the capital-intensive depreciation burden, a threshold the firm said signals the industry is approaching financial sustainability. "The data shows that AI revenue is no longer just a promise — it's covering real costs," said Azeem Azhar, founder of Exponential View, in the report. "This doesn't mean every dollar spent is profitable, but it does mean the aggregate economics are starting to work." The $25 billion in revenue came from a mix of hyperscale providers — Amazon Web Services, Microsoft Azure, Google Cloud — and emerging cloud services that offer AI-specific compute. Depreciation costs of $21 billion reflect the rapid amortization of graphics processing units from Nvidia and Advanced Micro Devices, along with data center construction expenses that have pushed combined capital spending at the four largest US hyperscalers past $230 billion annually. The 19% revenue cushion over depreciation is thin but represents a marked improvement from early 2025, when AI revenue consistently fell short of depreciation by 10% to 15%. **Why the breakeven matters for investors** The milestone addresses a central question that has hung over the AI trade since the launch of ChatGPT in late 2022: whether the massive upfront spending on GPUs and data centers would ever generate sufficient returns. Microsoft, Amazon, Google and Meta Platforms have collectively committed more than $200 billion in annual AI-related capital expenditures, with Wall Street analysts at Goldman Sachs and Morgan Stanley warning earlier this year that a "productivity payoff" was needed by mid-2026 to sustain investor confidence. Exponential View's data suggests that payoff is beginning to materialize, albeit with narrow margins. The $4 billion surplus — revenue minus depreciation — represents roughly 2% of the hyperscalers' combined annual AI CapEx, meaning profitability from AI operations remains minimal even as top-line growth accelerates. Nvidia, whose data center revenue hit $36 billion in its most recent quarter, remains the primary beneficiary of the buildout, while companies such as AMD, Broadcom and Marvell Technology are competing for secondary supply positions. **What comes next** The sustainability of the trend hinges on whether AI revenue growth can outpace the depreciation curve. Hyperscalers are expected to continue raising CapEx through 2027 as they build out next-generation data centers equipped with Nvidia's Blackwell Ultra and Rubin architectures, as well as in-house chips from Amazon's Trainium and Google's TPU families. If AI revenue maintains its current trajectory, the cushion over depreciation could widen to 30% or more by late 2027, Exponential View projected. For investors, the report provides a data-driven counterweight to concerns that AI infrastructure spending is a bubble. Microsoft shares trade at 33x forward earnings, Amazon at 38x and Google at 24x — multiples that embed expectations of a multiyear AI-driven revenue acceleration. The breakeven milestone does not eliminate downside risk, but it shifts the burden of proof onto skeptics who argue the spending is unjustified. This article is for informational purposes only and does not constitute investment advice.
Micron and Marvell each tripled from March lows in the AI rally. Now both are down over 15% from June peaks — but what drove each decline differs sharply. "Micron's selloff is a China-memory narrative that doesn't touch Marvell's business at all," said Rachel Kim, semiconductor analyst at Edgen. "One is a structural competitive threat to a commodity product. The other is sector-wide profit-taking." Micron shares fell 8% to $903.50 on July 15 after Barron's reported that Apple is testing DRAM chips from Chinese producer ChangXin Memory Technologies (CXMT), which has become the world's fourth-largest DRAM maker. The selloff dragged the iShares Semiconductor ETF (SOXX) down 4% and pulled Marvell 7% lower, Intel 6% lower and AMD 6% lower — even though none of those three companies compete in memory chips. Micron's FQ3 2026 revenue reached $41.46 billion, up 346% year over year, with non-GAAP EPS of $25.11 and GAAP gross margin of 85%. The company guided for FQ4 revenue of $50 billion. For investors weighing which pullback offers a better entry, the distinction matters. Micron faces a genuine overhang from Chinese DRAM competition that could pressure pricing in commodity memory, even as its HBM business remains sold out through 2026. Marvell's decline, by contrast, reflects sector-wide de-risking after a 145% year-to-date gain — a rotation that could reverse as quickly as it started. TD Cowen maintains a $1,600 price target on Micron, implying roughly 77% upside from current levels. **The China Memory Overhang Is Micron-Specific** CXMT has climbed rapidly to become the world's fourth-largest DRAM producer, and Apple's decision to test its chips for devices sold in China points to a potential shift in the memory supply chain. Nio disclosed a $23.3 million investment in the Chinese memory maker, adding to the narrative that Chinese competitors are gaining scale. For Micron, which generated the bulk of its $41.46 billion in FQ3 revenue from DRAM and NAND, any erosion in pricing power in commodity memory could offset gains from high-bandwidth memory (HBM), the specialized chip that feeds data to AI accelerators. The company's HBM business remains a bright spot. Micron sold out its HBM capacity through 2026, and the product carries higher margins than commodity DRAM. But HBM represents a fraction of total DRAM shipments by volume. If CXMT's capacity expansion depresses pricing in the commodity segment — which still accounts for the majority of Micron's revenue — the margin mix could deteriorate even as AI-driven demand stays strong. **Marvell's Pullback Is About Rotation, Not Fundamentals** Marvell's 7% decline on July 15 came despite no direct exposure to the China memory story. The company designs custom silicon and networking chips for data center customers, a business that benefits from the same AI infrastructure buildout driving Micron's HBM demand. Marvell shares are up 145% year to date, and the stock has pulled back roughly 17% from its June record high. The selloff appears to be sector-wide profit-taking. The SOXX ETF, which holds all four names that fell on July 15, is a common vehicle for traders to reduce semiconductor exposure in a single trade. Marvell's fundamentals remain intact: the company has secured multiple design wins with hyperscale cloud providers for custom AI accelerators, and its networking business benefits from the 800-gigabit Ethernet transition in data centers. **Which Pullback Offers Better Value** For investors with a 12-month horizon, the two stocks present different risk-reward profiles. Micron trades at roughly 22 times forward earnings, a discount to its growth rate given the 346% revenue surge in FQ3. But the China competition overhang is not a near-term revenue threat — CXMT's chips are still being tested, and Apple has not committed to volume purchases. If the narrative shifts, Micron could recover quickly. Marvell trades at a higher multiple, reflecting its asset-light model and exposure to custom silicon — a market that is growing faster than commodity memory. The stock's pullback is driven by rotation rather than a change in fundamentals, which typically resolves faster than a structural threat. For traders who believe the AI infrastructure buildout has years left to run, Marvell's decline may be the more straightforward buying opportunity. This article is for informational purposes only and does not constitute investment advice.
The first wave of Q2 2026 earnings cleared a lowered bar, and now the market faces its real test: technology. Non-tech sectors — healthcare, consumer staples, financials — delivered results that met or exceeded reduced expectations, supporting the S&P 500 even as the Nasdaq Composite fell 0.66% to 25,949.60. The Dow Jones Industrial Average hit an all-time high of 53,289.30 before reversing to close at 52,879.27, down 0.33%. "The rotation into defensive sectors tells you the market is positioning for tech to disappoint," said Sarah Lin, equity analyst at Edgen. "The non-tech results were good enough, but the bar for semiconductors and AI-related names has moved well past what even a blowout quarter can clear." The Philadelphia Semiconductor Index dropped 5.5% to its lowest level in four weeks. Intel Corp. fell 8.2%, Micron Technology Inc. lost 7.3%, and KLA Corp., Marvell Technology Inc., Broadcom Inc. and Advanced Micro Devices Inc. all traded sharply lower. The VanEck Semiconductor ETF lost more than 5%. Nvidia Corp. slipped 1.8% after reports that Chinese AI startup DeepSeek is developing its own chip. Samsung Electronics Co. reported a 19-fold increase in operating profit for the second quarter, yet its stock sold off nearly 7% in Seoul. The reaction underscored how expectations have outpaced even exceptional results. South Korea's Kospi index gave back nearly 5% on the session. The selling pressure carried into U.S. markets, where the S&P 500 held at 7,516.76, down 0.27%, supported by gains in healthcare and consumer staples. Eli Lilly & Co. rose about 3%. Walmart Inc. advanced after announcing price cuts on products including ground beef and Coca-Cola. JPMorgan Chase & Co. and Microsoft Corp. also attracted buyers. Money leaving chips is rotating into sectors where the earnings bar is lower. Fiserv Inc. climbed 3.5% after reports the payments company held discussions with JPMorgan, Bank of America Corp. and other large U.S. banks about selling its debit card payments infrastructure business. In India, Tata Consultancy Services Ltd. met analysts' net profit estimates, supported by cost-cutting that offset weakness in its core IT services business. HCL Technologies Ltd., Wipro Ltd. and Tech Mahindra Ltd. are set to report this week. Accenture Plc earlier projected weaker-than-expected quarterly revenue, reinforcing demand concerns. Taiwan Semiconductor Manufacturing Co. is expected to release delayed June sales figures after Typhoon Bavi disrupted the schedule, offering a key indicator of global AI-driven demand. SK Hynix Inc. begins trading on the Nasdaq later this week, testing whether institutional money returns to chip stocks at current prices. The rotation into healthcare, staples and financials has been building for several sessions. The S&P 500 is sitting on a short-term retracement zone at 7,474.57 to 7,429.38, with the 50-day moving average at 7,410.62. The Nasdaq is pressing its 50-day moving average at 25,969.61. A break below those levels would signal the selling is broadening beyond tech. The guidance raise from non-tech companies signals management teams see stable demand in their end markets. But tech earnings will determine whether this remains a sector rotation or turns into a broader market correction. Investors will watch the June FOMC minutes on Wednesday for Chair Kevin Warsh's latest policy stance, and the SK Hynix listing later this week for a read on institutional appetite for semiconductor exposure. This article is for informational purposes only and does not constitute investment advice.
The Nasdaq Composite fell 0.7% at the open, while the S&P 500 slipped 0.3% and the Dow edged up 0.1%, as a rout in AI and chip stocks offset gains in energy shares. "The combination of spiking crude and a hot sector that's priced for perfection creates a natural trigger for profit-taking," said Michael Wilson, chief equity strategist at Morgan Stanley. "The question is whether this is a one-day shakeout or the start of a broader rotation." SK Hynix's US-listed shares plunged 9%, Micron Technology fell 5% and SanDisk dropped 6%, leading a broad semiconductor retreat. AMD and Intel each lost more than 4%, while optical networking names also slumped — Marvell Technology declined 5%, Corning fell 4% and Coherent shed 3%. The selloff extended to Asia earlier in the session, where Seoul's Kospi tumbled as much as 9%, with SK Hynix sinking 15% and Samsung Electronics losing 10%. The tech rout coincided with a 3.5% surge in Brent crude to $78.68 a barrel after Iran expanded military strikes to Gulf nations, raising fears over energy shipments through the Strait of Hormuz. Gold slid 1.5% to $4,060 as the oil spike revived expectations that central banks may need to keep interest rates elevated to combat inflationary pressures. **Oil Shock Rattles Rate Expectations** The renewed hostilities followed an Iranian attack on a commercial ship in the Strait of Hormuz early Sunday, with the Revolutionary Guards declaring the waterway "closed until further notice" — a threat the US Central Command countered by stating the strait remained open to lawful transit. The US military launched a fresh wave of strikes after several Gulf allies were targeted, marking the latest escalation in a conflict that has already reshaped global energy markets this year. "One can easily imagine the situation spiraling quite rapidly," said Fawad Razaqzada, a market analyst at Forex.com. "Of course, rhetoric can soften. We've seen that movie before. But for now, traders are forced to assume the worst." The oil spike complicates the outlook for the Federal Reserve, which has been navigating between sticky inflation and slowing growth. Higher crude prices feed directly into headline inflation measures, potentially delaying rate cuts that growth-oriented tech stocks have been pricing in. The US 10-year Treasury yield moved higher as traders adjusted expectations, though the exact level was not immediately available at the open. **Earnings Season Looms** The selloff comes at a critical juncture for the AI trade, with earnings season set to deliver reports from Taiwanese chip giant TSMC and Dutch equipment maker ASML this week. Wall Street banks including JPMorgan, Bank of America and Goldman Sachs are also scheduled to report, providing a broader read on the economy. IG analyst Fabien Yip said oil's return toward pre-war levels in June reflected markets pricing in a best-case outcome for the fragile US-Iran arrangement, and the "re-escalation exposes how fragile that assumption was." Near-term, she said, the risk premium should keep prices supported, though a repeat of the earlier spike appears unlikely given sluggish demand recovery and increased OPEC+ output. *This article is for informational purposes only and does not constitute investment advice.*
**The AI chip market is splitting into two tiers — Nvidia owns training, but Broadcom and Marvell are capturing the $1 trillion inference buildout with custom silicon that costs half as much.** The shift from buying Nvidia Corp.'s fastest GPUs to lowering the cost of operating AI at hyperscale is creating a growing opportunity for Broadcom Inc. and Marvell Technology Inc., whose custom chips cost roughly half as much as Nvidia's most advanced racks. "Hyperscalers continue to expand their investments in proprietary silicon to improve total cost of ownership and reduce dependence on merchant GPU suppliers," Morgan Stanley analysts wrote in a recent note. Building one gigawatt of AI infrastructure using ASIC racks designed by Broadcom or Marvell costs between $6 billion and $11 billion, according to Milk Road AI estimates. By comparison, racks built around Nvidia's GB300 processors cost roughly $19 billion, and the figure jumps to $25 billion under Nvidia's next-generation Vera Rubin architecture. Google, Amazon, Meta Platforms Inc. and Microsoft Corp. are all pursuing custom silicon alongside Nvidia GPUs because every percentage point of efficiency matters when spending tens of billions annually. The four hyperscalers are forecast to spend over $1 trillion on AI infrastructure next year, adding 19.5 gigawatts of incremental compute capacity — nearly triple the 6.7 GW they added in 2025. Google alone will add 6.8 GW, more than the hyperscalers combined two years ago. Broadcom, which partners with Google on its Tensor Processing Units and with Meta, and Marvell, which works with Amazon on its Trainium chips, benefit whether customers build proprietary processors or buy more off-the-shelf hardware. **The Economics of Custom Silicon vs. Merchant GPUs** The appeal of application-specific integrated circuits comes down to dollars. Nvidia's GPUs remain the gold standard for training frontier models, and its CUDA software platform provides a competitive moat that rivals have struggled to crack. But once models are deployed, inference workloads — answering prompts, generating images and powering AI applications — don't always require Nvidia's most powerful processors. They reward lower costs and higher efficiency instead. That is where Broadcom and Marvell enter the picture. Rather than selling chips under their own brands, they help cloud providers design silicon optimized for their own software stacks and infrastructure. Google continues developing TPUs with Broadcom while still purchasing enormous volumes of Nvidia GPUs. Amazon follows a similar dual-track strategy with Trainium alongside Nvidia deployments. That diversification gives cloud providers leverage in negotiations while matching the right chip to the right workload. **Nvidia's Response and the Investment Case** Nvidia is hardly standing still. The company continues expanding beyond GPUs into networking, rack-scale systems and software, and has opened technologies such as NVLink Fusion to custom silicon partners — including Marvell itself. Still, every new custom AI accelerator designed by a hyperscaler creates another opportunity for Broadcom or Marvell. They are standing directly in the path of a massive spending wave. For investors, the question is where the fastest incremental growth will occur. Nvidia commands the premium end of AI computing, particularly for training, and its data center revenue remains dominant. But custom silicon is becoming an essential part of every hyperscaler's long-term strategy. Broadcom appears best positioned today thanks to its deep relationships with Google and Meta, while Marvell continues strengthening its foothold with Amazon and other large customers. The AI infrastructure market is expanding so rapidly that multiple winners can emerge. This article is for informational purposes only and does not constitute investment advice.
**Five chip and data center executives publicly rebutted fears of an AI computing surplus, arguing energy supply — not demand — is the binding constraint.** The selloff in semiconductor stocks this month has been driven by concerns that AI infrastructure spending is peaking. Meta Platforms Inc. said it would sell idle AI computing capacity, and Samsung Electronics Co. warned on profit despite a 360% gain over 12 months. But executives who build and supply the infrastructure told CNBC the opposite is true. "AI demand is almost unlimited," Pat Gelsinger, former Intel Corp. chief executive and now general partner at Playground Global, said. "The real bottleneck is energy — not compute." Marc Boroditsky, chief revenue officer at Nebius, a cloud platform building data centers with Nvidia Corp. graphics processing units, said demand "far exceeds our ability to fulfill it, and that has been the case for some time." Andrew Feldman, chief executive of Cerebras Systems Inc., a maker of AI chips that went public this year, dismissed the Meta and xAI capacity sales as isolated cases. "Across the industry, demand for compute far exceeds available supply," he said. "We have a shortage of data centers and a shortage of many of the inputs needed to build compute." The most striking signal came from Lumentum Holdings Inc., a supplier of photonic and optical interconnect products. Chief Executive Michael Hurlston said the company's order book is filled for the next five years. "We are doing everything we can to expand capacity to meet what we now see as five years of demand," he said. Lumentum shares have risen about 600% over the past 12 months. **The Energy Constraint** The International Energy Agency projects global data center electricity demand will more than double to 945 terawatt-hours by 2030, up from about 415 TWh in 2024. AI is the primary driver, with data center power consumption growing at roughly 15% annually — more than four times the pace of other sectors. McKinsey & Co. estimates global data centers will require nearly $7 trillion in capital investment by 2030, with about $1.3 trillion earmarked for power generation, transformers, substations and transmission infrastructure. Gelsinger's framing aligns with a growing consensus among industry planners: the next phase of the AI race will be won not by chip architecture alone but by the power grid that sustains it. Countries including the U.S., China and Gulf states are racing to secure clean, reliable electricity for AI clusters, with India's Ministry of Power estimating data center electricity demand could reach 13.56 gigawatts by 2032. **From Token-Maxxing to Value-Maxxing** A separate concern has been whether enterprise AI spending can sustain its pace. Many companies have moved through a phase Boroditsky called "token-maxxing" — encouraging employees to use AI tools as much as possible, often on frontier models from OpenAI and Anthropic. As chief financial officers scrutinize costs, that approach is giving way to what he termed "value-maxxing." "CFOs tightening spending are actually looking for value," Boroditsky said. "We are seeing a more rational shift. Every technology cycle goes through this, and this rationalization will actually sustain demand." Cerebras's Feldman offered a complementary view: as enterprises mature in their AI deployment, different workloads will migrate to different tiers of compute. "You don't take a bus to buy groceries," he said. "Some workloads will move to one class of compute, simpler workloads to another." That suggests frontier models and open-source alternatives from companies like DeepSeek and Alibaba Group Holding Ltd. will coexist rather than cannibalize each other, keeping aggregate compute demand on an upward trajectory. Sungyun Park, chief executive of South Korean chip startup Rebellions Inc., which counts Samsung and SK Hynix Inc. among its backers, said the momentum behind AI infrastructure remains strong. He does not see the Meta and xAI moves as evidence of overinvestment by hyperscale cloud providers. For investors, the debate carries direct implications. Nvidia trades at about 22 times forward earnings after a 13.6% decline from its 52-week high, while Advanced Micro Devices Inc. has drawn analyst price targets as high as $615 on expectations its MI450 series and Helios platform will begin ramping in the third quarter. Marvell Technology Inc., whose optical connectivity products face extended lead times beyond six months, has analysts projecting more than 40% growth over the next three years. The risk is not that AI demand fades, the executives argued. It is that the industry cannot build power infrastructure fast enough to keep up. *This article is for informational purposes only and does not constitute investment advice.*
The AI data center buildout is minting a second tier of winners beyond Nvidia, as hyperscaler capital spending floods into custom silicon, optical interconnects, and PCIe connectivity. Five chip stocks — MACOM Technology Solutions, Marvell Technology, Arm Holdings, Astera Labs, and Cerebras Systems — have surged between 76% and 187% year to date. "We are seeing exceptional AI-related bookings, and as a result, we are significantly raising Marvell's revenue outlook for both fiscal 2027 and fiscal 2028," Chief Executive Officer Matt Murphy said on the company's May 27 earnings call. Marvell's data center segment contributed $1.83 billion in fiscal Q1, or 76% of total revenue, up 27% from a year earlier. Astera Labs posted $308.4 million in quarterly revenue, up 93.4% year over year, and has beaten consensus in every one of its 10 reported quarters. MACOM's fiscal Q2 revenue hit $289 million, up 22.5%, with management guiding the next quarter to as much as $339 million. Arm Holdings saw data center royalties more than double year over year as every major AI CPU — Nvidia's Vera, Google's Axion, Microsoft's Cobalt — now runs on its architecture. The divergence between these stocks and the broader semiconductor index is widening. Marvell has gained 164% year to date, Astera Labs 130%, and Arm 182%, while the Philadelphia Semiconductor Index has risen roughly 25%. The question for investors is whether the second-tier AI trade still has room to run. **Custom Silicon and the Toll-Booth Model** Marvell Technology has emerged as the custom chip partner of choice for hyperscalers building their own AI accelerators. Its portfolio spans 800G and 1.6T optical interconnects, 51.2T Ethernet switches, and custom XPU designs. Revenue reached $2.42 billion in fiscal Q1, up 27.6% year over year, and management guided fiscal Q2 to $2.7 billion. The stock carries 31 Buy ratings and seven Strong Buy ratings, with a consensus price target of $252. Arm Holdings operates a different model — collecting a license fee upfront and a royalty on every chip shipped. Nvidia's Vera, Google's Axion, and Microsoft's Cobalt are all Arm-based. Fiscal Q4 revenue reached $1.49 billion, up 20.1%, with licensing revenue of $819 million, up 29%. Chief Executive Officer Rene Haas said demand for Arm's AGI CPU "has exceeded expectations." The stock trades at 137 times forward earnings, with 27 Buy or Strong Buy ratings. **Revenue Jumps 93% and a $20B OpenAI Deal** Astera Labs makes the retimers, smart cable modules, and fabric switches that keep PCIe 6 and CXL links alive inside AI servers. In May, the company launched the Scorpio X-Series 320-lane Smart Fabric Switch targeting a $20 billion merchant scale-up market by 2030. Revenue jumped 93.4% year over year to $308.4 million, with operating income up 448%. Cerebras Systems went public in the second quarter of 2026, raising $6.4 billion, and locked a multiyear OpenAI deal for 750 megawatts of inference compute valued at more than $20 billion. First-quarter revenue printed at $193.4 million, up 94% year over year. Shares have pulled back 31% since the May IPO peak, trading below the analyst consensus target of $291. MACOM, the smallest of the five by market cap, builds the RF, microwave, analog and optical semiconductors that appear in hyperscaler switch cages and pluggable optics. As 800G migrates to 1.6T and copper interconnects hit their reach limits inside AI racks, MACOM's content per rack goes up. The stock trades at 45 times forward earnings, well below the pure-play AI silicon cohort. **Investor Implications** The five stocks trade at a wide range of valuations, but the sell-side remains broadly bullish: zero sell ratings across all five names. The risk is that hyperscaler capital spending, which drove the buildout, could slow if token demand growth decelerates — a scenario the market has not yet priced in. For now, the second-tier AI trade is accelerating, and the next earnings cycle will test whether the revenue momentum can sustain the multiples. This article is for informational purposes only and does not constitute investment advice.
**Semiconductor stocks surged Thursday, tracking an 8.8% rally in Chinese chip shares fueled by Changxin Memory Technologies' $4.34 billion Shanghai IPO plans.** US chipmakers rebounded sharply in premarket trading, recovering some of the ground lost after Samsung Electronics' forecast-beating results triggered a sector-wide selloff earlier this week. Micron Technology, Intel, Coherent and Marvell Technology each climbed more than 3% by 5:41 a.m. ET, while Applied Materials rose 3.8% and Advanced Micro Devices gained 2.2%. "The China semiconductor rally is providing a much-needed bid after the Samsung profit-taking shook confidence," said Rachel Kim, semiconductor analyst at Edgen. "Changxin's IPO is a reminder that memory demand remains structurally strong, even if near-term positioning got overextended." The CSI Semiconductor Index closed up 8.8% after Changxin Memory Technologies, the world's fourth-largest DRAM maker with about 7.7% global market share, said it would begin book-building July 15 for a Shanghai IPO aiming to raise 29.5 billion yuan, or roughly $4.34 billion. The move reignited enthusiasm for chip stocks across Asia and spilled into US premarket trading, with South Korea's Kospi closing 0.6% higher as SK Hynix rose 5.3% and Samsung gained 0.8%. The rebound comes just two days after a brutal selloff sparked by Samsung's preliminary second-quarter results. The South Korean giant reported operating profit of 89.4 trillion won ($58.44 billion) — nearly 19 times higher than a year earlier and exceeding the LSEG SmartEstimate of 87.3 trillion won. Revenue jumped 129% to 171 trillion won. Yet Samsung shares fell sharply as investors locked in profits following the stock's red-hot rally, raising broader questions about whether the artificial-intelligence-driven demand that has powered the chip sector's surge can sustain elevated pricing. **The Memory Price Question** Memory chip prices have risen sharply over the past year, and the Samsung selloff highlighted growing investor anxiety about demand durability. The PHLX semiconductor index remains down about 15% from its record high in late June, while Micron has dropped more than 20% since hitting its own peak on June 25. Despite the pullback, the index is still up roughly 75% year-to-date, and Micron has more than doubled. The so-called hyperscalers — Microsoft, Meta Platforms and Alphabet — are under increasing scrutiny as earnings season approaches. Their capital expenditure plans for AI infrastructure directly determine revenue visibility for chipmakers. A slowdown in spending growth could pressure the lofty valuations that semiconductor stocks command. Intel, which reports second-quarter results on July 23, guided for non-GAAP earnings per share of $0.20 on revenue of $14.3 billion at the midpoint — an 11% year-over-year increase and a sharp turnaround from the $0.10 per-share loss it posted a year earlier. Analysts expect full-year 2026 EPS of $1.09, a 161% surge from 2025. **What's at Stake for Investors** The semiconductor sector has added nearly half of the S&P 500's market value gains this year, according to JonesTrading data. But the speed of the rally has fueled debate about sustainability. Intel trades at 904 times trailing earnings, though its forward multiple of 137 reflects expectations of a dramatic earnings recovery. On Stocktwits, retail sentiment was neutral on Micron and the Roundhill Memory ETF, while bearish on Intel and AMD — a cautious posture ahead of earnings. The key question for the weeks ahead is whether the China-driven rebound has legs or represents a dead-cat bounce before another leg lower. Changxin's IPO will test investor appetite for memory exposure at a moment when the sector's valuation has rarely been higher. If the offering prices successfully, it could validate the bull case that AI-driven memory demand has further to run. If it stumbles, the selloff that began with Samsung may not be over. This article is for informational purposes only and does not constitute investment advice.
The Dow Jones Industrial Average set a new all-time high for the third consecutive session before reversing to close 0.3% lower at 52,834. "The intraday record shows the market's underlying strength, but profit-taking in high-flying chip names pulled the broader market back," said Sarah Lin, equity market analyst at Edgen. The S&P 500 fell 0.5% and the Nasdaq Composite dropped 1.2%. The Dow had risen as much as 0.5% in morning trading, crossing the 53,100 level for the first time, before giving up gains. Leading the Dow's advance were 3M, Amgen, Procter & Gamble, Google parent Alphabet, and Apple. The chip selloff was the day's dominant drag. Samsung Electronics dropped 7% in Seoul trading despite reporting second-quarter revenue of roughly 171 trillion Korean won ($112.7 billion) and operating profit of about 89.4 trillion won ($59 billion), both above consensus. The weakness spilled into US-listed peers: Intel fell 6.5%, Advanced Micro Devices dropped more than 3%, and Marvell Technology declined 6.5%. The VanEck Semiconductor ETF lost 3%. Energy was the S&P 500's best-performing sector, gaining more than 3%, as oil prices surged after the Treasury Department revoked a waiver allowing Iran to sell its crude. West Texas Intermediate futures jumped 4.9% to $71.90 a barrel, while Brent crude rose 5.1% to $75.70. Occidental Petroleum, Devon Energy, and APA each gained between 5% and 6%. The 10-year Treasury yield rose about seven basis points to 4.54%, while the US Dollar Index edged up 0.2% to 101.08. Gold futures fell 1.2% to $4,115 an ounce. Bitcoin traded near $63,700, little changed over the past 24 hours. In other single-stock moves, SpaceX shares fell about 7% on their first day as a member of the Nasdaq 100 index. Caterpillar dropped nearly 5% after announcing the acquisition of Skycatch, a mining data analytics firm, though the stock remains up more than 60% year to date. Walmart rose more than 1% after announcing thousands of price cuts across categories. Rivian sank 11% after the electric-vehicle maker announced a public offering of about 75 million shares. The reversal from intraday highs leaves the Dow up about 14% year to date, while the S&P 500 has gained roughly 18%. Investors now turn to the start of second-quarter earnings season next week, with big banks including JPMorgan Chase and Wells Fargo scheduled to report. This article is for informational purposes only and does not constitute investment advice.
**Three AI-focused companies — Marvell Technology, Palantir Technologies, and Fluence Energy — still have room to run after the S&P 500's 15% second-quarter rally, as structural demand in custom silicon, enterprise AI, and energy storage remains undervalued by current market prices.** The S&P 500 surged more than 15% in the second quarter of 2026, yet three AI-focused companies still offer significant upside as structural demand drivers in custom silicon, enterprise software, and energy storage remain underappreciated by current valuations. "AI infrastructure sits where powerful long-term themes meet real-world hardware, turning complex technology into billable capacity," said a senior technology strategist at a major investment bank. Marvell Technology looks compelling because the market is only now pricing in a durable second act in custom AI chips, according to industry analysts. Palantir Technologies has become the poster child of the enterprise AI trade, and after a sharp first-half pullback, its valuation may offer a more attractive entry point. Fluence Energy, meanwhile, has gotten torched over the last six months — its stock price has dropped 21.3% to $17.19 per share since January 2026 — creating a potential rebound opportunity as AI-driven energy demand accelerates. The broader market's 15% rally masks a divergence: the highest-flying AI stocks are seeing valuation compression as concerns bleed through, while companies with tangible revenue from AI infrastructure — custom chips, enterprise deployments, and energy storage — may still have room to re-rate higher. For investors, the question is whether the market has fully priced in the second wave of AI monetization. Marvell's opportunity in custom ASIC designs for hyperscale cloud providers represents a second act beyond its traditional networking and storage markets. The company's custom AI silicon business is gaining traction as Amazon, Google, and Microsoft seek alternatives to Nvidia's GPUs for inference workloads, a shift that could reshape the semiconductor competitive landscape. Palantir's enterprise AI platform has driven a surge in U.S. commercial revenue as companies move from pilot programs to production deployments. The stock's sharp first-half pullback followed a period when shares had more than tripled in 2025, and the correction may have brought valuations closer to levels that reflect the pace of actual revenue conversion rather than future expectations. Fluence Energy sits at the intersection of AI and clean energy, providing battery storage systems that help data centers manage power demand. The 21.3% decline in its stock price since January reflects broader headwinds in the clean energy sector, but the company's exposure to data center infrastructure buildout could provide a catalyst as AI compute demand continues to grow. This article is for informational purposes only and does not constitute investment advice.
**Marvell Technology is executing Broadcom's custom silicon playbook, targeting $10 billion in AI chip revenue by fiscal 2029.** Marvell Technology's data center revenue has swelled to 76% of total sales as the chipmaker locks in more than 50 custom AI design opportunities across over 10 hyperscaler customers, mirroring the strategy that turned Broadcom into a $1.71 trillion company. "We are seeing exceptional AI-related bookings, and as a result, we are significantly raising Marvell's revenue outlook for both fiscal 2027 and fiscal 2028," Chief Executive Officer Matt Murphy told investors on the Q1 earnings call. Revenue hit $2.418 billion in the fiscal first quarter, up 27.6% year over year, with data center climbing 11% sequentially. Free cash flow more than doubled to $483.1 million. Management guided second-quarter revenue to $2.70 billion, implying roughly 35% year-over-year growth. The custom XPU pipeline, which Reuters reported could top $10 billion by fiscal 2029, spans lead 3-nanometer programs with a top US hyperscaler and follow-on architecture work already underway. At roughly $251, Marvell trades at 67 times forward earnings — more than three times Broadcom's 20x multiple — pricing in the custom silicon ramp before it materializes in full. Any delay in the lead 3nm XPU program, expected to enter production in calendar 2026, could trigger a violent rerating. **The Custom Silicon Flywheel** Marvell has recycled proceeds from asset sales — the automotive Ethernet unit went to Infineon for $2.5 billion — into acquisitions that deepen its interconnect moat. The company closed Celestial AI, a photonic fabric startup, in February 2026 and picked up XConn Technologies for chiplet connectivity, then raised $2 billion through a Series A convertible preferred to fund the buildout. The strategy mirrors Broadcom's: sell the picks and shovels, let Nvidia fight the merchant GPU war. The NVIDIA NVLink Fusion partnership adds another layer. Marvell's high-speed optics, 51.2-terabit Ethernet switches, and custom silicon now touch every layer of the AI rack except the GPU itself. That breadth gives the company pricing power and customer stickiness that pure-play chip designers lack. **The Risks Priced Into 67x Earnings** The bear case is concentrated. Custom silicon revenue depends on a handful of hyperscalers, any of which could dual-source to Broadcom on the next node. Murphy acknowledged customers "may be pursuing multiple paths" on XPU supply. GAAP net income collapsed 80.4% year over year last quarter on a $331.8 million contingent consideration charge, while stock-based compensation climbed to $207.6 million. Insiders have logged 129 transactions, net selling. At 67 times forward earnings versus the semiconductor industry average of 65.7, the stock leaves little room for execution missteps. Gross margin guidance of 58.25% to 59.25% for the current quarter will be the first test — if it holds and revenue clears $2.7 billion, the thesis gains credibility. If it slips, the multiple compression could erase months of gains. Marvell's path to $10 billion in custom chip revenue by fiscal 2029 would compress today's forward multiple rapidly even without further expansion. The risk-reward is asymmetric at $251, but only for investors willing to bet that hyperscaler demand for custom silicon outruns the competitive response from Broadcom. Watch Q2 FY27 gross margins and the 3nm XPU timeline — those two data points will determine whether Marvell becomes Broadcom's equal or its cautionary tale. This article is for informational purposes only and does not constitute investment advice.
UBS analyst Timothy Arcuri raised Marvell Technology Inc.'s price target to $340 from $230, implying a 39% upside from the current $245.29. "Marvell's Teralynx T100 switch silicon and growing exposure to the CXL interconnect market represent underappreciated growth drivers," Arcuri, who reiterated a Buy rating, said. The new $340 target comes after Marvell in June launched the Teralynx T100, a 102.4 Tbps switch chip built on 3nm process technology for AI data centers. UBS estimates the global CXL market will reach about $45 billion by 2027 and expand to $70 billion to $100 billion by 2030, with Marvell generating roughly $1 billion in CXL-related revenue by 2027. The upgrade counters a recent selloff that erased about $23.4 billion in market value on July 2, when shares dropped 9.84% to $245.29. The stock has fallen roughly 20% since joining the S&P 500 on June 22. Marvell reported first-quarter fiscal 2027 revenue of $2.418 billion on May 27, up 28% from a year earlier, with data center revenue making up 75.7% of total sales. The company guided second-quarter revenue to about $2.7 billion. Chief Executive Officer Matt Murphy said Marvell now expects its custom-chip business to exceed $10 billion in fiscal 2029 and raised its 2028 revenue outlook to roughly $16.5 billion from $15 billion. The Teralynx T100, sampling to customers this quarter, consumes under 1,000 watts — up to 25% less power than competing products — and supports a 512-port radix for scale-out deployments. Marvell also partnered with Nvidia Corp. in March on NVLink Fusion, with Nvidia investing $2 billion in the company. The UBS upgrade signals that Wall Street sees Marvell's AI networking and custom silicon bets as multiyear revenue drivers. Investors will watch the company's second-quarter earnings report, expected in late August, for updates on Teralynx T100 customer adoption and CXL design wins. This article is for informational purposes only and does not constitute investment advice.
Semiconductor stocks now account for a record 19.7% of the S&P 500, almost four times their weighting of about 5% in June 2020, as the AI boom concentrated capital into a handful of chipmakers. But that dominance is showing cracks. "The market is undergoing a 'mega rotation,' with capital aggressively shifting from lagging mega-cap tech into cyclical and value sectors," Craig Johnson, chief market technician at Piper Sandler, said. The rotation accelerated after Micron Technology Inc.'s blowout earnings failed to sustain momentum in the broader tech sector, he added. The Magnificent Seven — Nvidia Corp., Amazon.com Inc., Microsoft Corp., Meta Platforms Inc., Alphabet Inc., Apple Inc. and Tesla Inc. — were among the weakest performers in recent sessions. Microsoft and Apple dropped on price increases for some devices driven by higher memory costs, while the Nasdaq Composite closed at 25,297, down 0.2% on Friday. The Dow Jones Industrial Average, by contrast, hit intraday records during the week, signaling capital was moving into overlooked areas. Exchange-traded funds attracted more than $1 trillion in inflows year-to-date through late June 2026, about 45% above the record pace from the same period last year, according to Citadel Securities strategist Scott Rubner. The inflows have reinforced a self-perpetuating cycle: as semiconductor shares outperform, their index weight rises, prompting passive funds to allocate even more capital to the same names. **Valuation signals flash caution** Several indicators now point to elevated valuations across the semiconductor sector. Bank of America's proprietary Bubble Risk Indicator reached 0.91 for the PHLX Semiconductor Sector and 0.82 for the Technology Select Sector, on a scale where 1.0 represents extreme bubble-like conditions. The S&P 500's price-to-sales ratio has climbed to 3.22, well above its long-term average of 1.84, according to LSEG data. The Buffett Indicator — comparing total US stock market capitalization to gross domestic product — stands at 231.8%, a level that historically has preceded below-average returns. Before the dotcom crash, semiconductor stocks represented just over 8% of the S&P 500, less than half their current share. The concentration has broadened beyond Nvidia to include Broadcom Inc., Taiwan Semiconductor Manufacturing Co., ASML Holding NV, Advanced Micro Devices Inc., and memory-chip makers Micron and SanDisk Corp. **Memory stocks emerge as next AI beneficiary** The AI-induced memory shortage has emerged as a key theme, with Micron and SanDisk benefiting from rising demand for high-bandwidth memory (HBM) used in AI accelerators. But even these names have shown volatility — Micron fell 6.7% and SanDisk dropped 10.5% on Friday as investors questioned whether the rotation had further to run. Rob Haworth, senior investment strategy director at US Bank Asset Management, said the rotation is a critical narrative for investors to watch. "This key commodity for hyperscalers and technology hardware companies is weighing on these stocks," he said, referring to higher memory costs. "It indicates still solid investor sentiment despite the challenge of higher costs for some of the largest technology companies." David Morrison, senior market analyst at Trade Nation, noted that the Dow's resilience relative to the S&P 500 and Nasdaq could be a positive sign that "investors are still keen to be fully invested in equities but are rotating out of overheated semiconductor stocks into some overlooked sectors offering better value." **Software stocks stage a comeback** The rotation is not limited to memory. Software stocks that lost nearly $1 trillion in market value during a six-day selloff triggered by Anthropic's Claude plug-in launch in January have begun recovering. ServiceNow Inc., Salesforce Inc. and Adobe Inc. have each risen 10% or more in recent sessions as investors rotated into laggards. JPMorgan Chase & Co. identified software companies with proprietary enterprise data and established workflows as AI beneficiaries rather than casualties. Goldman Sachs Group Inc. CEO David Solomon echoed that view, saying "plenty of companies will pivot and do just fine." For investors, the broadening of the AI trade creates a dilemma. Nvidia shares trade at elevated multiples after a multiyear rally, while memory and software stocks offer discounted entry points after steep pullbacks. The question is whether the rotation reflects a healthy broadening of the AI investment theme — or the early stages of a more significant unwind in the sector that powered the bull market. This article is for informational purposes only and does not constitute investment advice.
**AI infrastructure stocks created 6 times more value than big tech hyperscalers over the past four years, a UBS research report found.** AI infrastructure companies generated 600% value creation since 2022, compared with 100% for hyperscalers such as Amazon, Microsoft and Google, according to UBS's research arm — a shift in where AI investment returns are concentrated. "The value creation in AI infrastructure has dramatically outpaced the hyperscalers themselves, marking a structural change in how the AI supply chain captures economic returns," the UBS research team wrote in a report published July 3. The finding comes as AI infrastructure stocks have surged in 2026. Sandisk gained 857%, Micron Technology rose 304% and Intel climbed 278%, according to data compiled by Investor's Business Daily. Western Digital, Marvell Technology, Seagate Technology and Dell Technologies each more than doubled. The seven S&P 500 components have benefited from surging demand for memory chips, data center equipment and custom AI processors as cloud providers race to build out capacity. The UBS analysis suggests the $200 billion-plus AI infrastructure buildout is generating returns disproportionately for suppliers rather than the cloud platforms footing the bill. That could fuel further capital rotation into semiconductor and data center stocks, while raising questions about the return on investment for hyperscalers spending tens of billions annually on AI. ## Infrastructure Suppliers Capture the Upside The UBS report quantifies what investors have sensed: the companies building the physical layer of AI — memory chips, networking gear, servers and data center power systems — are capturing a growing share of the AI economy's value. Memory chipmakers have been the standout beneficiaries. NAND and DRAM prices have surged 200% and 300%, respectively, over the past year as hyperscalers compete for limited supply, according to JPMorgan Chase strategist Meera Pandit. Micron's revenue jumped 345% in the May quarter, while Sandisk's sales rose 251% in the March period. Both companies guided sharply higher for the current quarter. The supply shortage is so acute that Nvidia has invested more than 250,000 GPUs into CoreWeave's data centers alone, the neocloud provider disclosed. ## Can the Cycle Sustain? The memory chip industry has historically been defined by boom-and-bust cycles, with periods of supply scarcity followed by oversupply and price collapses. Wall Street expects memory chip sales to peak in 2028, after which Micron's adjusted earnings could decline 27% in fiscal 2029 and Sandisk's by 54%, according to analyst estimates cited by The Motley Fool. Micron trades at 24 times earnings, while Sandisk trades at 67 times — a premium that reflects the magnitude of the current upcycle but leaves little room for error if demand normalizes. The UBS report did not address the cyclical risk directly, but the valuation gap between infrastructure suppliers and hyperscalers suggests the market is pricing in sustained growth rather than a repeat of the 2022-2023 downturn, when memory stocks fell 50% to 60%. For investors, the UBS analysis supports a rotation that has already reshaped the S&P 500's leadership. AI infrastructure names now command a combined market capitalization exceeding $3 trillion, with Micron alone valued at $1.29 trillion. The question is whether the 600% value creation figure represents a peak or a waypoint. Nvidia, the largest AI infrastructure supplier, trades at 21 times forward earnings — a discount to its growth rate that suggests the market sees room for further gains even as the cycle matures. This article is for informational purposes only and does not constitute investment advice.
**Anthropic's early-stage push to develop proprietary AI chips with Samsung as a potential manufacturing partner sent semiconductor stocks into a broad selloff across US and European markets.** Anthropic's early-stage push to develop proprietary AI chips and its talks with Samsung Electronics as a potential manufacturing partner triggered a broad selloff in semiconductor stocks, with the Philadelphia Semiconductor Index sliding 4.3% and the Nasdaq 100 turning negative after an initial gain driven by weaker-than-expected US jobs data. The project remains nascent, with no detailed design or manufacturing work begun, Anthropic told The Information. The company said Amazon's Trainium chips, Google's tensor processing units and Nvidia's graphics processors will remain core to its computing strategy. Anthropic is considering Samsung's 2-nanometer manufacturing process — which packs more transistors per square millimeter to improve performance per watt — and the Korean conglomerate's advanced packaging facilities, according to The Information. The company recently hired Clive Chan, an early member of OpenAI's custom chip team, as part of a deliberate engineering buildout. The move mirrors a strategy adopted by OpenAI, which tapped Broadcom to design its first custom inference chip, called Jalapeño, unveiled last month. The selloff hit memory makers hardest. Sandisk tumbled 12%, Western Digital fell 7.5% and Micron Technology dropped 4.3%. AMD slid 3.9%, Intel lost 2.9% and Nvidia declined 1.3%. European chip stocks also fell, with ASML, ASM International and BE Semiconductor Industries each dropping more than 3% and Nokia losing 4.1%. **Why AI Labs Are Building Their Own Silicon** The move by Anthropic follows a playbook adopted by Google, Amazon, Meta and Microsoft, all of which have developed proprietary silicon to reduce dependence on third-party suppliers. Nvidia, despite the competitive noise, has not lost ground — The Information's own estimates put the company's AI chip market share at 74%, higher than before the inference-chip arms race began. For Samsung's foundry business, winning a marquee AI client like Anthropic would be a significant victory. The Korean company has struggled with leading-edge process yields relative to TSMC's N2 node, a concern analysts have repeatedly raised. Google is separately considering using Samsung for part of a future tensor processing unit, according to The Information, which would represent another win for Samsung's contract manufacturing business if confirmed. **Investment Impact** The selloff reflects a market recalibrating the competitive dynamics of the AI chip supply chain. Nvidia shares fell 1.3% on the news. Broadcom, which generates custom chip design revenue from OpenAI, and Taiwan Semiconductor Manufacturing both traded higher in the session, as investors appeared to view the competitive threat as distant. The broader implication is that as more AI labs bring chip design in-house, the pricing power and volume commitments of incumbent semiconductor manufacturers face a longer-term threat. Samsung, SK Hynix and Micron all participated in Anthropic's $65 billion fundraising round in May, giving the AI company deep ties to the memory chip industry it may now be seeking to disrupt. For investors, the key question is whether Nvidia's 74% market share can withstand a wave of custom silicon from its own customers — a dynamic that could take years to play out but is already moving stock prices. This article is for informational purposes only and does not constitute investment advice.
**Wall Street is rewarding the companies supplying the AI boom rather than the technology giants funding it.** The artificial intelligence trade has undergone a decisive rotation, with investors abandoning the Magnificent Seven in favor of the semiconductor and hardware suppliers powering the multitrillion-dollar infrastructure buildout. "The easy phase of the AI investment story is over," Nigel Green, CEO of deVere Group, said. "Investors were happy to fund the biggest infrastructure buildout in corporate history while the narrative was simple. Now they want evidence." The Magnificent Seven — Apple, Microsoft, Alphabet, Amazon, Meta, Nvidia and Tesla — have shed $2.3 trillion in combined market value this month alone. Microsoft and Meta have fallen more than 30% from their 52-week highs, while Oracle has dropped nearly 60% from its all-time peak. Memory maker Micron Technology posted 85% margins on surging data center demand, and chip suppliers including Intel, Marvell, AMD and Sandisk have drawn fresh buying interest. The rotation reflects a fundamental reassessment of who captures the economics of artificial intelligence. CNBC's Jim Cramer, who continues to own several Magnificent Seven stocks, said the suppliers of AI infrastructure — not the hyperscalers writing the checks — are now in the strongest position to benefit. Within five years, Green predicts the Magnificent Seven will become the "Magnificent Three," with only a handful of mega-cap companies truly capturing AI's economic upside. **The Spending Divide Widens** Microsoft, Amazon, Alphabet and Meta are collectively spending hundreds of billions of dollars building artificial intelligence infrastructure, and the pace is accelerating. But markets have reached a stage where ambition alone no longer justifies the outlay. Apple's decision to raise prices because of soaring memory and storage costs signaled that even the world's most powerful technology companies may no longer control the economics of their own ecosystem, Green said. The market is already voting on this divide. While Magnificent Seven stocks have come under pressure, semiconductor and memory companies have continued to surge as investors increasingly back the suppliers rather than the buyers of AI infrastructure. Micron's 85% margins on data center memory products illustrate the pricing power flowing to the pinch points of the boom. **Downstream Fallout Spreads** The repricing has extended well beyond the mega-cap names. More than 75 data center projects worth a combined $130 billion were canceled in the first quarter of 2026, according to Data Center Watch, as community opposition and grid connection delays slowed buildouts. Alternative investment firms that backed the theme have been among the hardest hit: Blue Owl, TPG, Ares Capital Management, KKR and Blackstone have fallen between 40% and 60% from their all-time highs. Independent power producers that signed data center electricity deals have also pulled back. Constellation Energy has fallen 37% from its 52-week high, Vistra Energy is down more than 26%, and Oklo, the Sam Altman-backed nuclear startup, has dropped over 70%. **What Comes Next** The second-quarter earnings season, approaching in July, will force markets to confront the question they have spent the past year avoiding: where are the returns on AI spending? If one major hyperscaler blinks and changes its spending plans, it could trigger further turbulence. A recent report warned that a reversal of excessive tech investments could risk a global financial crash. For now, the money is rotating. Cramer identified Micron, Intel, Marvell, AMD and Sandisk as the suppliers best positioned to benefit from the current cycle. Nvidia, which has been both a Magnificent Seven member and the dominant AI chip supplier, occupies a unique position — its products remain essential to the buildout, but its stock has not been immune to the broader selloff. Green said investor nerves are likely to intensify before they settle. "Markets are being asked to finance one of the largest capital expenditure cycles in history while accepting that the ultimate winners remain uncertain," he said. "Of course, this creates volatility, anxiety, and also periodic crises of confidence. We should expect more of them." This article is for informational purposes only and does not constitute investment advice.
**Memory chip and hardware supply chain stocks delivered the best quarterly performance in history as AI infrastructure spending cascaded beyond Nvidia.** The US memory chip and hardware supply chain index surged 159% in the second quarter, more than quadrupling from 69.59 points at the start of the year to 277.74 points, as investors rotated from AI hyperscalers into semiconductor enablers. "The rotation out of AI hyperscalers into AI enablers has shifted investors' euphoria into semis, driving spectacular rallies," Barclays analyst Anshul Gupta said in a note. The index closed at 268.03 on June 30, up 3.83% on the day, capping a June gain of 27.26%. Among the best performers, Micron Technology rose 304% in the first half of 2026, while Intel gained 278% and Advanced Micro Devices added 171%, according to LSEG data. Sandisk, spun off from Western Digital in February 2025, led the S&P 500 with an 858% surge. The rally reflects a fundamental shift in AI spending patterns: as hyperscalers like Amazon, Alphabet, Meta and Microsoft pour capital into data centers, the companies supplying memory, storage and networking gear are capturing an increasing share of that investment. The VanEck Semiconductor ETF rose 71% in the second quarter, its best quarterly performance since inception in 2000. ## Memory Makers Lead the Charge Micron, one of three major producers of computer memory, saw its market value swell by roughly $920 billion in the second quarter alone. The company reported last week that revenue in the latest quarter more than quadrupled from a year earlier, driven by skyrocketing memory prices from AI chip demand. Its gross margin jumped to 84.9% from 39% a year earlier. Micron CEO Sanjay Mehrotra announced new customer agreements that he expected would "significantly enhance the durability and predictability" of the company's financial performance, according to a company statement. Intel, the legacy maker of central processing units, added $480 billion in market cap during the second quarter as its stock jumped 216%. The company is building US chip factories while benefiting from renewed demand for CPUs as more AI workloads move to devices. AMD, Intel's rival in both CPUs and graphics processing units, added $615 billion in value after its stock nearly tripled. ## Beyond Chips — The Infrastructure Boom Spreads The rally extended well beyond memory and processor makers. Marvell Technology, which makes networking gear for data centers, climbed about 200% in the second quarter. Arm, which supplies chip designs to other semiconductor companies, rose 134%. Seagate Technology and Western Digital, the dominant players in data storage, gained 250% and 271% respectively in the first half, according to LSEG. Dell Technologies, a major supplier of AI servers, rose 243% in the first six months of 2026. Applied Materials and Lam Research, which manufacture the equipment used to produce chips, gained 181% and 153% respectively. The breadth of the rally suggests the AI infrastructure buildout is entering a new phase. While Nvidia remains the dominant AI chipmaker, its stock gained only 15% in the second quarter — a fraction of the returns delivered by the companies supplying the ecosystem around it. For investors, the question is whether valuations can sustain the pace. Micron trades at 8.1 times forward earnings, the lowest forward P/E among the top 20 S&P 500 performers this year, reflecting lingering concern over the cyclical nature of the memory business. Sandisk's forward P/E has actually declined to 12 times despite its 858% gain, as earnings estimates have risen even faster than the stock price. Jefferies analyst Blayne Curtis raised his price target on Sandisk to $3,000 on Friday, implying 32% upside from its closing price of $2,273.73. This article is for informational purposes only and does not constitute investment advice.
Marvell Technology received a Buy rating and $296 price target from 24/7 Wall St., with a model projecting $1 trillion in market value by late 2032. "Marvell is one of the most explosive semiconductor stories of the cycle, with custom AI design activity at an all-time high," the 24/7 Wall St. analyst team said. The $296.46 target implies 11.13% upside from Friday's close of $266.77. KeyBanc raised its target to $385 with a bull case of $450, while BofA went to $365 and Stifel to $350. UBS lifted its target to $340 on accelerating CXL demand, with analyst Timothy Arcuri projecting CXL-related revenue of $1 billion in 2027. Marvell trades at 66x forward earnings with data center revenue representing 76% of the $2.42 billion Q1 mix. The $1 trillion milestone, projected for Dec. 31, 2032, hinges on custom silicon scaling to over $10 billion in revenue by fiscal 2029. Shares have surged 234% over the past year and 214% year to date, tripling off the January low of $79.01. The stock sits 26% below the 52-week high of $329.88 after a 14% one-week pullback. CEO Matt Murphy guided Q2 revenue to $2.7 billion at the midpoint, representing 35% year-over-year growth, and said the company is "significantly raising Marvell's revenue outlook for both fiscal 2027 and fiscal 2028." S&P 500 inclusion took effect June 22, adding passive flows to the AI bid. The bull case centers on custom silicon. Murphy said AI custom design activity is at an all-time high, with more than 50 opportunities across over 10 customers. Custom chip revenue could exceed $10 billion by fiscal 2029, and interconnect guidance was raised by over 70%. Recent acquisitions of Celestial AI and XConn add photonic fabric and chiplet IP for 1.6T optics. Risks include a forward P/E of 66x and trailing P/E of 92x, leaving no room for execution slippage. Data center represents 76% of sales, and a 10% drop on June 9 tied to ByteDance ASIC headlines showed how fast sentiment can shift on vertical integration fears. GAAP net income fell 80.6% year over year in Q1, though this reflects a $331.8 million contingent consideration charge tied to the Celestial AI deal. The bear-case 1-year price is $224.08, a 16% drawdown. The guidance raise indicates management expects AI demand to sustain its rapid pace. Investors will watch the Q2 earnings report for confirmation that the 35% growth trajectory holds and bookings remain at record levels. This article is for informational purposes only and does not constitute investment advice.
**Marvell Technology's 4.97% surge reflects growing conviction that its custom AI silicon and networking portfolio can capture hyperscaler infrastructure spending.** Marvell Technology's custom AI chips and networking gear are emerging as key growth drivers for hyperscaler customers, pushing the stock up 4.97% on June 28. "The upside case is powerful, but the stock's future depends on execution, customer concentration, and whether Marvell can turn AI infrastructure demand into durable revenue," Rick Orford, an affiliate of The Motley Fool, said. The gain, based on June 12 market prices and reported June 28, came as investors reassessed Marvell's position in the AI infrastructure buildout. The company's custom chip business competes with Broadcom and Nvidia in designing application-specific integrated circuits (ASICs) for cloud giants seeking alternatives to general-purpose GPUs. Marvell's networking portfolio — data center switches and optical interconnects — provides a second revenue stream tied to the same hyperscaler expansion. Customer concentration creates the central risk: a small number of hyperscalers account for a large share of custom chip orders. If Marvell delivers on its design wins, the AI chip and networking businesses could become major growth engines. If execution falters, the stock's current valuation may prove difficult to justify. **Custom Chips vs. General-Purpose GPUs** Marvell's custom chip strategy targets the same hyperscaler demand that has driven Nvidia's data center revenue above $100 billion annually. But unlike Nvidia's general-purpose GPUs, Marvell builds tailored ASICs for specific customer workloads — a model that offers lower margins but deeper customer lock-in. The approach mirrors Broadcom's custom chip business, which has secured design wins with Google and Meta for their internal AI accelerators. The trade-off is structural. General-purpose GPUs command higher prices and gross margins because they serve a broad market. Custom ASICs sacrifice margin for guaranteed volume and multi-year supply agreements. For hyperscalers, the calculus is about total cost of ownership: a custom chip that delivers comparable inference performance at 30 percent to 40 percent lower power consumption can save millions annually at data center scale. **Networking as a Second Growth Vector** Marvell's networking business adds another layer of hyperscaler exposure. As cloud providers expand data center footprints, demand for high-speed switches, optical components, and interconnect technology grows in tandem. The company's portfolio in data center switching and electro-optics positions it to benefit from the same infrastructure cycle driving its chip business. The networking segment also diversifies Marvell's revenue base. While custom chip design wins are concentrated among a few hyperscaler customers, networking products sell across a broader set of enterprise and cloud buyers. That diversification matters if any single chip customer delays or reduces orders. **Supply Chain and Execution Risks** Execution remains the critical variable. Custom chip programs require long development cycles, with revenue recognition lagging design wins by 12 to 18 months. Marvell must also navigate supply chain dependencies on TSMC for advanced node manufacturing and on packaging partners for HBM (high-bandwidth memory) integration — the same bottlenecks affecting the broader semiconductor industry. Any disruption at TSMC's fabrication facilities or in the CoWoS (chip-on-wafer-on-substrate) packaging supply chain could delay Marvell's product timelines. The company's reliance on a single foundry for its most advanced chips mirrors an industry-wide concentration risk that has become a focal point for investors since the pandemic-era supply shortages. **What This Means for Investors** Marvell shares trade at a premium reflecting the AI growth premium baked into the stock. If the company executes on its design wins and diversifies its customer base, the current valuation could prove justified. But with customer concentration and execution risk weighing on the outlook, the stock's next leg higher depends on tangible revenue conversion from the AI infrastructure pipeline. Investors will be watching Marvell's next earnings report for updates on design win progress and revenue guidance tied to its custom chip programs. This article is for informational purposes only and does not constitute investment advice.