In the AI infrastructure boom, Nvidia (NASDAQ: NVDA) sells the brains of the AI factory while Micron (NASDAQ: MU) supplies the memory that keeps those brains fed with data, and that difference shapes which stock will benefit more from the current phase of this historic spending wave. In my view, Nvidia is the clearer winner because a greater fraction of every dollar of hyperscaler capex is spent on its accelerators than goes toward memory chips of the type that Micron manufactures. Micron still looks like a powerful second-derivative play, since AI servers can't be built without the high-bandwidth memory it supplies.
The money flow this year is wildly high. The hyperscalers themselves say they plan to spend hundreds of billions of dollars in 2026 alone to expand AI data centers, GPU clusters, networking, and power infrastructure, a sharp jump from already elevated 2025 levels. One estimate puts combined capex for Amazon, Microsoft, Alphabet, and Meta Platforms at around $700 billion, with roughly two-thirds of that directed toward AI infrastructure rather than traditional cloud. Within that budget, the largest line item is the AI server stack itself, where accelerated servers built around high-end GPUs drive most of the component revenue growth. And of course, the hyperscalers are not the only tech players building data centers now.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Nvidia sits directly in the center of this buying spree. Its data center business now revolves around entire racks of AI computing power, not just single chips. Systems like the GB200 Grace Blackwell Superchip and GB200 NVL72 tie together dozens of CPUs and GPUs into logical accelerators that can train and serve trillion-parameter models more efficiently than the prior-generation Hopper platforms. Hyperscalers are lining up to deploy these systems in their AI clouds, with massive companies committing to offer GB200 NVL72 instances to customers who want to run large language models at scale. All this sounds dense, but the basic point is that Nvidia products are in steady demand.
Nvidia's roadmap also continues to push the limits of performance and memory. Architectures like Blackwell and its new Vera Rubin processors combine vast computing throughput with enormous pools of high bandwidth memory (HBM), turning racks into "AI factories." That keeps Nvidia at the absolute center of procurement decisions when cloud providers are calculating how many clusters they will need to handle their training and inferencing workloads in 2026 and beyond.

#memory #high #signal

Minda Corporation will invest close to Rs2.70bn ($28.2m) to build a new manufacturing facility in Pune, Maharashtra, to produce thin film transistor (TFT) displays for automotive use, executive director Aakash Minda told CNBC.
The project has secured approval under the Indian government's Scheme for Promotion of Manufacturing of Electronic Components and Semiconductors.
According to the report, it is designed to cut the company's reliance on imported displays while strengthening its manufacturing presence within India.
At the outset, the plant will concentrate on automotive displays for instrument clusters, infotainment systems and ******* pits.
Minda already produces clusters and other display-related components and intends to supply the displays made at the new facility to its own operations alongside external customers.

#clusters

Micron Technology (NASDAQ: MU) has been on a tear recently, and its run in this artificial intelligence (AI) era is about more than just selling more memory. Micron's run is about solving one of the quiet bottlenecks in artificial intelligence (AI) infrastructure, which is power, and that is where I think the story bleeds directly into utility stocks in a way the market has not fully priced in yet.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Micron's latest numbers show just how central it has become to AI. In the third quarter of its fiscal 2026, total revenue hit $41.5 billion, up 346% year over year and marking the fifth straight quarterly record. DRAM revenue alone was $31.3 billion, up 343% and now 76% of total sales, while data center revenue topped $25 billion on an annualized run rate north of $100 billion. ****** ysts have pushed estimates higher, largely on the back of AI-driven demand for high-bandwidth memory (HBM) and high-performance dynamic random-access memory (DRAM).
Underneath those numbers is a very specific technology angle. Micron's HBM3E memory, which sits right next to Nvidia's H200 GPUs and AMD's next-generation accelerators, delivers more than 1.2 terabytes per second of bandwidth while using about 30% less power than competing offerings. This means that AI clusters can either cut their electricity bills or pack more GPUs into the same power envelope, which is exactly what hyperscalers care about now that power availability has become a defining constraint for scaling AI.
Micron just raised its planned U.S. investment to more than $250 billion through 2035, aiming to put about 40% of its DRAM output on American soil to supply AI data centers and to support more than 90,000 jobs. Its solid-state drive (SSD) business is also framed in power terms now.

#power #billion #memory #NVIDIA

Groq, a San Francisco-based AI infrastructure company, has raised $350m in a Series A round that values the company at $3.5bn.
The round was led by technology investment company Disruptive, with Nvidia expected to participate.
Alongside the $650m secured in June 2026, the latest fundraise takes Groq's recent funding to $1bn.
Completion of the Series A remains subject to customary closing conditions.
Groq aims to use the funding to expand its data centre network and support customers seeking access to medium and larger clusters of Nvidia accelerated computing for training and inference work.

#round

The artificial intelligence (AI) memory supercycle has transformed semiconductor investing, fueled by explosive demand for high-bandwidth memory (HBM) and advanced DRAM to power hyperscalers' training clusters, inference deployments, and data center expansion.
Surging demand for graphics processing units (GPUs) and expanding agentic AI workloads are tightening memory chip supply -- driving higher average selling prices and record profits for manufacturers.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Micron Technology (NASDAQ: MU) and Sandisk (NASDAQ: SNDK) have emerged as two popular plays in the AI memory boom. As of Aug. 6, Micron stock has gained roughly 715% over the past year and about 212% year-to-date. Sandisk has posted even more dramatic gains, nearing 3,000% returns over the last 12 months and roughly 439% so far in 2026.
With that said, I think a stronger opportunity may be hiding in plain sight with South Korean rival SK Hynix (NASDAQ: SKHY). Let's dig into why.

#signal

Oracle's (NYSE: ORCL) stock has rallied 27% over the past two weeks. Let's see what catalysts drove its stock higher, and if it can maintain its momentum over the next few months.
The biggest catalyst for Oracle's stock was the expansion of its partnership with Alphabet's (NASDAQ: GOOG) (NASDAQ: GOOGL) Google on July 30. Google will integrate its advanced Gemini AI models into Oracle's AI Agent Studio, enabling Oracle's Cloud Infrastructure (OCI) customers to build and deploy AI agents with Google's AI software.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
OCI has been Oracle's primary growth engine over the past year. In fiscal 2026 (which ended on May 31), its cloud infrastructure revenue surged 77% to $18.1 billion, accounting for 27% of its top line, as more companies used its high-peformance GPU clusters to power AI applications.
OCI is much smaller than Amazon (NASDAQ: AMZN) Web Services (AWS), Microsoft (NASDAQ: MSFT) Azure, and Google Cloud, but it integrates its services into those larger cloud infrastructure platforms so that its enterprise customers can run their workloads seamlessly across multiple cloud providers without being locked into a single ecosystem. Its expanded partnership with Google complements that strategy and increases its exposure to the AI market.

#cloud #Google #infrastructure #Stock

Emerald Wealth Partners, an independent ******* et and wealth management firm based in Zurich, released its Q2 2026 investor letter for the "Focused Equity Strategy." A copy of the letter can be downloaded here. The Strategy reported a 0.8% gross return for the second quarter of 2026. Equity markets rallied after the first quarter's Iran shock faded and oil prices reversed their spike, but the advance remained narrow and concentrated in cyclical semiconductor and memory stocks benefiting from supply constraints. The letter warned that investors were treating peak-cycle earnings as durable, while FOMO, IPO demand, and a US cyclically adjusted P/E ratio near 40 pointed to late-cycle market behavior. Long-term fundamentals are currently undervalued by the market, leading to underperformance in the strategy, notably due to the portfolio's avoidance of low-quality momentum-driven semiconductor stocks, which boosted the benchmark this quarter. It remains focused on enterprise infrastructure, digital platforms, custom silicon, and networking businesses that hold customer relationships, workflow data, distribution, and efficiency advantages, which could allow them to monetize AI rather than be displaced by it. Please review the Strategy's top five holdings for its key selections.
In its second-quarter 2026 investor letter, Emerald Wealth Partners Focused Equity Strategy highlighted Broadcom Inc. (NASDAQ:AVGO). Broadcom Inc. (NASDAQ:AVGO), a leading American company that designs and develops various semiconductor devices and infrastructure software solutions, contributed 64 bps to the strategy's performance during the quarter. On August 4, 2026, Broadcom Inc. (NASDAQ:AVGO) closed at $418.16 per share. One-month return of Broadcom Inc. (NASDAQ:AVGO) was 7.58% and its shares gained 38.62% over the past 52 weeks. Broadcom Inc. (NASDAQ:AVGO) has a market capitalization of $1.99 trillion.
Emerald Wealth Partners Focused Equity Strategy stated the following regarding Broadcom Inc. (NASDAQ:AVGO) in its Q2 2026 investor letter:
"Broadcom Inc. (NASDAQ:AVGO) added 64 bps, its 22.3% share price climb helped by the news OpenAI has selected the company to design its custom inference chip. Broadcom is the dominant designer of custom AI silicon: it has long co-designed Google's Tensor Processing Units, which power Gemini and some of Anthropic's inference capacity through Google Cloud. Few companies possess the engineering depth to develop bespoke accelerators at hyperscale. AI spending is shifting from training models to running them, where custom chips deliver materially better performance per watt and a lower cost per token than Nvidia's general purpose GPUs. Under Hock Tan, one of the industry's ablest capital allocators, the company has ******* embled two further pillars that sit at the center of the AI buildout: It dominates the Ethernet switching silicon that interconnects AI clusters — content that grows with cluster size, whether those clusters run on Nvidia

Emerald Wealth Partners, an independent ***** et and wealth management firm based in Zurich, released its Q2 2026 investor letter for the "Focused Equity Strategy." A copy of the letter can be downloaded here. The Strategy reported a 0.8% gross return for the second quarter of 2026. Equity markets rallied after the first quarter's Iran shock faded and oil prices reversed their spike, but the advance remained narrow and concentrated in cyclical semiconductor and memory stocks benefiting from supply constraints. The letter warned that investors were treating peak-cycle earnings as durable, while FOMO, IPO demand, and a US cyclically adjusted P/E ratio near 40 pointed to late-cycle market behavior. Long-term fundamentals are currently undervalued by the market, leading to underperformance in the strategy, notably due to the portfolio's avoidance of low-quality momentum-driven semiconductor stocks, which boosted the benchmark this quarter. It remains focused on enterprise infrastructure, digital platforms, custom silicon, and networking businesses that hold customer relationships, workflow data, distribution, and efficiency advantages, which could allow them to monetize AI rather than be displaced by it. Please review the Strategy's top five holdings for its key selections.
In its second-quarter 2026 investor letter, Emerald Wealth Partners Focused Equity Strategy highlighted Broadcom Inc. (NASDAQ:AVGO). Broadcom Inc. (NASDAQ:AVGO), a leading American company that designs and develops various semiconductor devices and infrastructure software solutions, contributed 64 bps to the strategy's performance during the quarter. On August 4, 2026, Broadcom Inc. (NASDAQ:AVGO) closed at $418.16 per share. One-month return of Broadcom Inc. (NASDAQ:AVGO) was 7.58% and its shares gained 38.62% over the past 52 weeks. Broadcom Inc. (NASDAQ:AVGO) has a market capitalization of $1.99 trillion.
Emerald Wealth Partners Focused Equity Strategy stated the following regarding Broadcom Inc. (NASDAQ:AVGO) in its Q2 2026 investor letter:
"Broadcom Inc. (NASDAQ:AVGO) added 64 bps, its 22.3% share price climb helped by the news OpenAI has selected the company to design its custom inference chip. Broadcom is the dominant designer of custom AI silicon: it has long co-designed Google's Tensor Processing Units, which power Gemini and some of Anthropic's inference capacity through Google Cloud. Few companies possess the engineering depth to develop bespoke accelerators at hyperscale. AI spending is shifting from training models to running them, where custom chips deliver materially better performance per watt and a lower cost per token than Nvidia's general purpose GPUs. Under Hock Tan, one of the industry's ablest capital allocators, the company has ***** embled two further pillars that sit at the center of the AI buildout: It dominates the Ethernet switching silicon that interconnects AI clusters — content that grows with cluster size, whether those clusters run on Nvidia GPUs

Corvex (NASDAQ:MOVE) secured a multi-year agreement to provide NVIDIA Blackwell GPU infrastructure, funding the expansion through debt, customer prepayments, and existing cash rather than issuing new shares.
Corvex (NASDAQ:MOVE) signed a multi-year agreement to provide NVIDIA Blackwell GPU clusters to a leading AI company.
The expansion is being funded without issuing additional equity, relying instead on debt financing, customer prepayments, and cash on hand.
The company deployed high-density, liquid-cooled GPU infrastructure in approximately two weeks within an existing air-cooled facility.
Revenue from the agreement has already been recognized as cluster deliveries progressed, with full run-rate revenue expected to begin midway through the current quarter.

#corvex #NVIDIA #multi #year

Arete Research, an independent research firm, has upgraded Texas Instruments Incorporated (NASDAQ:TXN) to Buy from Neutral and lifted its price target to $405 from $303. This 34% raise in the PT comes days after the company posted a strong quarter – a surprise from an ***** og cyclical sector. Texas Instruments' Q2 revenue reached $5.46 billion, up 23% year-over-year, beating the $5.24 billion consensus. The company's EPS rose 52% year-over-year, beating the $1.92 Street estimate with $2.14. The most crucial piece of information tucked away in the report is the twofold growth in data center income.
According to an Arete Research ***** yst, the surging AI demand is anticipated to cause three years of ***** og semiconductor shortages. The ***** yst believes that Texas Instruments is well-positioned with respect to capacity, which should yield significant market share gains during this cycle. Arete projects revenue of approximately $34 billion with earnings of $17 per share by fiscal 2028. These structural tailwinds, discussed by the ***** yst, align with the company's operational and financial performance. Let's break them down.
The AI buildout faces one severe bottleneck – power shortage. And power is exactly what Texas Instruments, nicknamed by the market as "boring ***** og," sells. Hyperscalers are shifting toward 800-volt DC architectures to support denser GPU clusters. As a result, the need for ***** og content per rack increases. Texas Instruments serves this critical conversion layer through high-voltage gallium-nitride (GaN) power devices, real-time motor control, and sensing technology. Additionally, the company's March collaboration with Nvidia on humanoid robotics expands this portfolio. These demands fuel the quarterly ***** og revenue, which rose 26% to $4.37 billion. Texas Instruments is not relying on an industrial market recovery but is working on strengthening its position as a significant supplier for technology's most capital-intensive infrastructure cycle.
A more durable driver stems from timing rather than the AI narrative. Texas Instruments is completing a six-year, roughly $24 billion fab expansion that constrained free cash flow. Spending is dropping significantly. The 2026 capex projected at $2 billion to $3 billion stands in contrast to the comparatively high $4.55 billion in 2025. The fabs are complete, and revenue is recovering now, allowing management to anticipate free cash flow per share exceeding $8 in 2026, compared to $3.23 in 2025. Cash generation is tripling alongside declining capital expenditure, altering the investment profile investors have avoided during the build.

#instruments #analog #Research #cash

Welcome to the fresh season. The Shuffle Up series reflects how I tier positions as I get ready for my drafts. The salaries are unscientific in nature and meant to illustrate where talent clusters and drops off. Players listed at the same salary are considered even.
Join or create a Yahoo Fantasy Football league for the 2026 NFL season
I am going to draft receivers proactively, as I do every year. I want to build a roster where the receivers on my team essentially start themselves. It's probably harder to find an impact free agent at this position than it is at any other spot. If you're going to have a dynamic WR room, you need to build it now.
$44 Ja'Marr Chase
$43 Puka Nacua

#receivers #players #join

Welcome to the fresh season. The Shuffle Up series reflects how I tier positions as I get ready for my drafts. The salaries are unscientific in nature and meant to illustrate where talent clusters and drops off. Players listed at the same draft salary are considered even.
Join or create a Yahoo Fantasy Football league for the 2026 NFL season
Today's **** ignment is perhaps the most important one -- the running backs. If there's one spot you want to run pure in any fantasy season, it's here.
$47 Jahmyr Gibbs
$47 Bijan Robinson

#players #gibbs #bijan

Corning Incorporated (NYSE:GLW) delivered rapid optical growth on July 28, then suffered a share-price decline of more than 20% after its guidance fell short of elevated expectations. Coherent Corp. (NYSE:COHR) also traded sharply lower. Coherent belongs in this story for a specific reason: the two companies monetize different components of the same AI data-center links.
Corning Incorporated (NYSE:GLW) supplies the passive layer, including optical fiber, cable and dense connector systems. Coherent supplies active devices and modules, including lasers, photodiodes and 800G and 1.6T transceivers that convert electrical signals into light and back again. As clusters add accelerators, network builders need both more fiber paths and faster optical endpoints. Corning's order pace can therefore provide a partial read-through on the volume and timing of deployments that create transceiver demand for Coherent Corp. (NYSE:COHR). It cannot establish Coherent's market share, pricing or margins.
Corning reported second-quarter core sales of $4.74 billion, up 17% year over year, and core earnings of $0.78 a share, up 30%. Optical Communications sales increased 32% to $2.07 billion. Within that segment, Enterprise Networks grew 65%, and the company said generative-AI product sales grew significantly faster.
The disappointment came from the next quarter. Corning projected $4.9 billion to $5.0 billion of third-quarter core sales, about 16% year-over-year growth but slightly below Wall Street's expectation, according to Reuters. Optical growth also eased from 36% in the first quarter to 32% in the second. Corning had gained roughly 64% in 2026 through the prior close, leaving little room for even modest deceleration.
Coherent's own latest quarter, ended March 31, argues against treating Corning's guidance as proof of weaker Coherent operations. Revenue rose 21% to $1.81 billion, GAAP gross margin reached 37.7%, and management cited exceptionally strong data-center and communications demand while expanding capacity. Its product roadmap includes 1.6T transceivers, lasers and photodiodes for AI networks. Different customer mixes, active-component content and manufacturing yields can make its results diverge from Corning's.

#corning #optical

The preliminary fiscal Q4 2026 update from Super Micro Computer Inc. (NASDAQ:SMCI), which was announced on July 21, serves as a case study in the distinction between a discounted valuation and a 'value trap'. The server manufacturer announced more than $60 billion in new orders during the quarter, a record backlog, and guided gross margins to 15-17%, nearly doubling its previous 8.2-8.4% prediction, citing a better customer and product mix.
Revenue is still expected to be near the low end of its $11-12.5 billion guidance range, below the roughly $11.67 billion ****** yst consensus, though shares rose as much as 20% on the news, with margin and order data clearly outweighing the top-line miss for investors focused on where the business is going.
For the broader technology ecosystem, Super Micro Computer Inc. (NASDAQ:SMCI)'s order increase is an important indicator of downstream artificial intelligence hardware demand. Since Super Micro Computer Inc. (NASDAQ:SMCI) bases its high-performance server clusters on NVIDIA GPU architectures and has historically contributed roughly 9% of NVIDIA's total revenue, the $60 billion order intake provides solid proof that hyperscaler AI infrastructure spending remains strong. At a time when macro experts have questioned the ability of major cloud providers to continue multibillion-dollar capital expenditure cycles, Super Micro's record backlog indicates that customer demand for liquid-cooled AI computing racks is increasing rather than decreasing.
However, Super Micro's own history is why the stock's price can't be evaluated the same way a clean order-book beat generally is. Back in March 2026, federal prosecutors unveiled an indictment charging co-founder and board member Yih-Shyan "Wally" Liaw, along with two other individuals ****** ociated with the company, with collaborating to smuggle $2.5 billion in NVIDIA-powered AI servers to China in breach of US export regulations. Shares plunged more than 28% in a single day as a result of the announcement, and one ****** yst reported by Yahoo Finance at the time described the company as "uninvestable."
That history is reflected in how cheap the company has become, despite the solid order data: Super Micro Computer Inc. (NASDAQ:SMCI) trades at a forward P/E ratio of approximately 9x, less than half the hardware sector median of about 24x, and a PEG ratio of around 0.4, both of which would ordinarily indicate serious undervaluation.

#micro #smci #billion #company

OpenAI has raised its projected spending on computing infrastructure to around $750 billion through 2030, up from roughly $600 billion earlier this year, according to The Wall Street Journal. New deals with cloud-computing providers are driving the higher figure, as OpenAI continues its push to obtain the vast computing resources its AI models require.
On Wednesday, OpenAI said it would invest $20 billion to begin construction on a data center called Project Camellia in Effingham County, Georgia. Unlike its other facilities, where the company leases chip capacity from providers such as Oracle and Amazon Web Services, the Savannah Gateway Industrial Hub site marks OpenAI's first venture as the principal designer and builder of its own data center.
Sachin Katti, OpenAI's vice president of compute strategy, said the company has contracted with utility Georgia Power to receive 3.2 gigawatts of power between 2028 and 2032. OpenAI has already acquired the land for the project and is in the process of selecting a partner to build and operate the site, Katti said.
The company has also hired Brent Mayo, a key figure in building Elon Musk's xAI data center infrastructure, according to the Journal. After departing xAI earlier this year, Mayo had been instrumental in standing up that company's Colossus supercomputer campus in Memphis, where he managed the rapid procurement and commissioning of large-scale AI chip clusters. In the head of data-center build and delivery role, he will be responsible for keeping cloud partners on track with their construction timelines, while also contributing to the new Georgia facility. Mayo reports to Uday Ruddarraju, who was promoted this month to become OpenAI's chief technology officer of computing capacity.
The $750 billion projection represents the latest escalation in a spending trajectory that has drawn scrutiny inside the company. Chief Financial Officer Sarah Friar has privately raised concerns that OpenAI may not be able to honor future computing contracts if revenue growth does not keep pace with commitments. That came after Chief Executive Sam Altman stated publicly that OpenAI intended to spend $1.4 trillion on computing capacity, a figure that unsettled observers; Friar subsequently stepped in to correct the record, telling investors the actual projected outlay through 2030 was approximately $600 billion.

#computing #center #company

It's been a tumultuous year for Meta Platforms (NASDAQ: META) investors. Despite the company's strong underlying ad revenue and user engagement metrics, the stock has been under meaningful pressure.
While the stock at one point cratered by as much as 20%, shares are now down just 2% year to date -- showing some degree of resilience as capital continues rotating toward companies that are viewed as clearer artificial intelligence (AI) infrastructure winners. With its next earnings reported scheduled for July 29, I predict Meta CEO Mark Zuckerberg will take that opportunity to announce a major strategic shift that could reframe Meta's role in the AI economy.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Meta's stock has been under pressure this year for one reason: the scale of its AI investments. The company is deploying unprecedented sums to build data centers, design custom silicon, and procure GPU clusters to support its AI initiatives.
Those capital expenditures have compressed its free cash flow in the near term. Investors are growing increasingly concerned about the potential timeline for Meta to deliver meaningful returns on this capital, especially amid questions about how AI will be monetized across the broader tech sector.

#year #Stock #flashing

The explosive growth of artificial intelligence (AI) has ignited a powerful supercycle in the memory semiconductor landscape. As large language models (LLMs) and generative AI systems scale to trillions of tokens, the bottleneck is shifting from raw compute to the speed and capacity of data movement.
Nvidia (NASDAQ: NVDA) is a dominant force in this ecosystem, not merely as the leading designer of graphics processing units (GPUs) but as the primary driver of demand for specialized high-bandwidth memory (HBM). The company's GPUs support the majority of hyperscale training clusters and inference workloads, forcing memory producers to align their roadmaps with Nvidia's performance targets.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Three companies possess the technology and manufacturing expertise to produce HBM4 at scale: SK Hynix (NASDAQ: SKHY), Samsung, and Micron Technology (NASDAQ: MU). Each company is aggressively expanding capacity and refining its manufacturing processes to meet Nvidia's specifications.
Nvidia CEO Jensen Huang is taking the memory supercycle incredibly seriously. Over the last couple of years, Huang has quietly dropped some breadcrumbs that can be traced to Nvidia's favorable memory suppliers. Let's take a look at what Huang has to say about the memory bottleneck, and explore which AI memory stock you may want to put on your radar right now.

Credo Technology Group Holding Ltd (NASDAQ:CRDO) is one of the best performing AI stocks over the last 3 years, with a 3Y CAGR of 153%. On July 6, William Blair **** yst Sebastien Naji named Credo his top semiconductor pick for the next six months. That followed several constructive June calls. Evercore ISI **** yst Mark Lipacis initiated coverage on June 22 with an Outperform rating and a $325 target, describing Credo as an AI-connectivity leader moving from a mainly copper portfolio toward a combined copper-and-optical offering.
The same day, Stifel **** yst Tore Svanberg maintained a Buy rating and raised his target to $350 from $250 after management meetings, citing the company's vertically integrated, system-level approach across both technologies. Bank of America kept Buy on June 26. The common thread is that larger AI clusters require more high-speed, energy-efficient connections. Credo's active electrical cables, digital signal processors and optical products address that need. Risks include customer concentration, rapid interface transitions, and expectations that already **** ume substantial growth.
Credo Technology Group Holding Ltd (NASDAQ:CRDO) provides high-speed connectivity solutions, including active electrical cables, digital signal processors, retimers, SerDes chiplets, and optical components for hyperscale, cloud, and AI networks.
While we acknowledge the potential of CRDO as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: 33 Stocks That Should Double in 3 Years and Cathie Wood 2026 Portfolio: 10 Best Stocks to Buy.

Former Florida Gov. Jeb Bush (R) on Wednesday raised the alarm over Iranian drones believed to be in Cuba.
"Iran has been a consistent threat to the united security interests of our country, certainly of Israel," Bush said during a United Against Nuclear Iran event, serving as the group's chair. "And now, we see that it's not just Israel and the United States, but the Gulf region as well. I also want to point out that the press reports that there are 300 of these [drones] in Cuba."
He said the U.S. has capabilities to protect the country before adding that "drones organize in clusters and can get through the defense capabilities, as we've seen in the Middle East as well."
"I think it's important to recognize that Iran has consistently been working with Cuba, Venezuela prior to the departure of Maduro from the regime, creating instability in not just in Cuba, but certainly in the region," Bush said, referring to deposed Venezuelan President Nicolás Maduro.
He praised the current administration's efforts against Iran's military capabilities, saying "we're on the verge of creating a much more peaceable world by standing up against this aggression."

Nebius Group NV (NASDAQ:NBIS) is one of the best NASDAQ stocks to invest in for long term. On June 24, Nebius announced the release of Nebius AI Cloud 3.6, which introduces upgrades to developer experience, security, and storage performance. A key highlight is the introduction of Nebius Echo, an AI-powered agent integrated into the web console that allows users to manage infrastructure using NL. The update also streamlines workflows with improved search, notification systems, and one-click integration with SkyPilot.
To support production workloads in regulated environments, the platform has added enterprise-grade security features, including a new Key Management Service with customer-managed encryption keys and Workload Identity Federation. Additionally, FinOps teams can now use a new budgeting tool to set spending targets and alerts, while engineers gain more granular control over Kubernetes and containerized image deployments.
photo by Business-laptop-campaign-creators on Unsplash
The release further addresses performance bottlenecks through major storage enhancements, including local SSD support for GPU servers and an intelligent object storage class for cost-effective data archiving. These updates, combined with the launch of the Nebius Builder Program and new professional certifications, are designed to help teams scale AI projects more efficiently while maintaining operational control.
Nebius Group NV (NASDAQ:NBIS) is a technology company that provides infrastructure and services to AI builders worldwide. It offers Nebius AI, an AI-centric cloud platform that provides full-stack infrastructure, including large-scale GPU clusters, cloud services, and developer tools.

Cerebras (NASDAQ: CBRS), a producer of AI chips, went public at $185 per share on May 14. Its stock opened at $350, but it now trades at about $205. That's still 11% above its IPO price, but investors who chased its post-IPO gains are now underwater. Let's see why Cerebras' stock fizzled out -- and if it's worth buying today.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Cerebras doesn't produce small GPUs like Nvidia (NASDAQ: NVDA). Instead, it builds massive AI processors on a single silicon wafer without cutting them into individual chips. Cerebras chips are as big as dinner plates, while Nvidia's GPUs are the size of postage stamps.
Cerebras claims its bigger chips bypass the networking bottlenecks, data latency, and power constraints ******* ociated with connecting traditional GPU clusters. They also outperformed traditional GPU clusters in inference tasks (when applications accessed trained data). It generates its revenue by selling its wafer-scale processors and CS-3 systems, as well as providing customers with cloud-based access to its own wafers to run inference tasks.
Cerebras recently secured a multi-year $20 billion deal with OpenAI to deploy 750 megawatts of its wafer-scale inference systems. It's also integrating its CS-3 systems into Amazon (NASDAQ: AMZN) Web Services (AWS), the world's largest cloud infrastructure platform.

You may want to be extra cautious with your fresh fruit and veggie platter this holiday weekend: An explosive diarrhea-causing parasite is behind a surge of summer food poisoning cases right before Independence Day.
Cyclosporiasis, a stomach-churning illness caused by a parasite often found on produce, has sickened hundreds across more than 17 U.S. states since May 1, according to the Centers for Disease Control and Prevention (CDC) and state health data. One state, Michigan, is contending with a cluster of infections that has already more than tripled its total case count compared with 2025.
While state health authorities, the CDC and the Food and Drug Administration (FDA) are working to determine a common thread among the illnesses, a singularly related outbreak hasn't been identified. Instead, the cases thus far appear to be part of a seasonal surge, though officials are searching for a potential throughline.
"CDC has no evidence of a single, multistate Cyclospora outbreak linking cases happening right now and being reported in press," the agency said in a June 2 statement to USA TODAY. "The data currently found on CDC's website is a surveillance count of cases across the United States of people with cyclosporiasis. This number includes clusters of cases currently under investigation by FDA and cases that have not been linked to a common source."
Here's what to know about avoiding the parasitic infection while picnicking this Fourth of July and beyond, including where it's been detected so far.

In the early days of online shopping, Amazon (NASDAQ: AMZN) was a simple website that sold books. In the years that followed, the company expanded its marketplace into a more comprehensive e-commerce platform. That eventually helped pave the way for the launch of its cloud infrastructure platform, Amazon Web Services (AWS). This chain of events quietly turned Amazon into an essential digital infrastructure provider -- driving trillions of dollars in market value.
Elon Musk's ****** e Exploration Technologies (NASDAQ: SPCX) is following a similar path. While ****** eX began with rockets that made it cheaper to get payloads into orbit, the company now also offers global internet connectivity through its Starlink business and is building large artificial intelligence (AI) data centers.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
SpaceX's long-term goal is to create a comprehensive suite of tools that power the entire AI economy. Recent steps, including its merger with xAI and its acquisition of Cursor AI, are speeding this process up.
SpaceX oversees the complete sequence required to deliver all aspects of the AI infrastructure value chain. The company's rockets handle the launches that place equipment into orbit. Starlink's broadband satellites provide a global connectivity network that can link AI systems with end users. And on the ground, ****** eX is deploying large clusters of servers dedicated to training AI models.

Explosive growth in the use of artificial intelligence (AI) is driving an expanding need for specialized computing resources, and traditional cloud infrastructure providers are struggling to supply those resources in sufficient quantity. ****** e Exploration Technologies (NASDAQ: SPCX) -- which is best known for its reusable rockets and its Starlink satellite network -- is aggressively expanding beyond the aerospace sector and into the world of accelerated computing capacity.
Through a series of targeted investments and strategic partnerships, ****** eX (as the company is known) is positioning itself to supply access to high-performance GPU clusters, and building a foothold in the neocloud economy.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Neoclouds are specialized data centers built around dense clusters of GPUs -- largely Nvidia's industry-leading processors -- rather than general-purpose servers. They streamline access to the huge parallel-processing power that's required for AI training and inference, sparing their clients the capital outlays of building and operating their own data center infrastructure. By specifically optimizing their clusters to handle AI workloads, neoclouds help accelerate model development and lower barriers to entry for smaller research teams. This is particularly useful now as there are a host of bottlenecks limiting the pace at which new data centers can be brought online.
Over the last year, ****** eX deployed meaningful capital into AI infrastructure, buying substantial quantities of Nvidia GPUs. The company has since inked agreements to supply AI infrastructure capacity to prominent clients such as Anthropic, Alphabet's Google Cloud, and Reflection AI. The total value of those three contracts could be about $82 billion over the next three years.

Broadcom Inc. (NASDAQ:AVGO) is one of the fastest-growing high-bandwidth memory stocks to buy. The stronger topic fit occurred on June 9, 2026, when Broadcom, Apollo, and Blackstone launched the AI XPV Platform to support more than 20 gigawatts of AI compute capacity, leveraging Broadcom's XPUs and networking solutions through 2028.
The platform starts with a $35 billion tranche for more than 1 gigawatt of Anthropic compute infrastructure expected to begin deployment at Fluidstack sites in mid-2026. Broadcom is not an HBM supplier, but custom AI XPUs typically depend on high-bandwidth memory and high-speed networking to keep large training and inference clusters from becoming data-starved. The update also lands soon after Broadcom's fiscal Q2 2026 report, where revenue rose 48% year over year to $22.2 billion and AI semiconductor revenue grew 143% to $10.8 billion. The LSEG VMware renewal is useful software news, but the XPV platform is much closer to the HBM-driven AI hardware cycle.
Broadcom Inc. (NASDAQ:AVGO) designs and supplies semiconductor and infrastructure software products, including custom AI accelerators, networking chips, storage connectivity, broadband components, wireless products, and VMware-based infrastructure software.
While we acknowledge the potential of AVGO as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: 33 Stocks That Should Double in 3 Years and Cathie Wood 2026 Portfolio: 10 Best Stocks to Buy.

As graphics processing unit (GPU) clusters grow larger and more power-hungry, electrical connections are reaching their limits when it comes to speed, heat, and energy use. In separate announcements, both Goldman Sachs and Nokia (NYSE: NOK) recently flagged photonics as the next critical layer in artificial intelligence (AI) infrastructure. Optical networking, which uses light to move data, offers an alternative path forward.
Let's ****** yze this enormous commercial opportunity and explore what it could mean for Nokia as the Finnish company quietly transitions from yesterday's leader of mobile devices to an AI networking powerhouse.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Photonics is the process of generating, controlling, and detecting light (photons) to transmit information. In AI data centers, silicon photonics takes this process a step further by integrating lasers, modulators, and detectors directly onto chips using existing semiconductor manufacturing lines. This integration can deliver measurable advantages over traditional electrical wiring, including higher bandwidth, lower latency, and reduced power consumption.
As AI training demands ever-larger clusters of GPUs, electrical signals are struggling to keep up without excessive heat and energy waste. Photonics solves this by enabling dense, energy-efficient optical links between chips and racks. This is essential, as compute power is no longer the main constraint of AI development -- connectivity is. Without scalable optical solutions, the next generation of AI applications will be limited by underlying infrastructure rather than silicon's capacity capabilities.

Cerebras Systems (NASDAQ: CBRS) and **** e Exploration Technologies (NASDAQ: SPCX) are two of the most talked-about artificial intelligence-linked initial public offering stories of 2026. But investors should not put them in the same bucket.
Cerebras is an AI-focused chipmaker that offers an unusual alternative to Nvidia's widely used chips. **** eX is a rocket, satellite internet, and AI infrastructure company after its merger with xAI.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
So, which looks like the better buy-and-hold pick for the next 10 years?
Cerebras' primary product is its wafer-scale engine, a powerful AI processor built using an entire silicon wafer. Makers of traditional GPUs and CPUs slice those wafers up to manufacture dozens or hundreds of smaller chips, and AI workloads generally require clusters of such chips connected together in servers and data centers. But in those data centers, performance can be limited by how quickly data can move across chips, memory systems, and servers.

We just covered Donald Trump Stock Portfolio: 10 Best AI and Tech Stock Picks in 2026. Marvell Technology (NASDAQ:MRVL) ranks #5 (see Donald Trump Stock Portfolio: 5 Best AI and Tech Stock Picks in 2026). The stocks identified in this article are based on Trump's financial disclosure filings released by the U.S. Office of Government Ethics. According to a statement from the Trump Organization cited by Reuters, Trump's investment holdings are maintained through fully discretionary accounts managed by third-party financial institutions, which have sole authority over investment decisions.
Stock Performance Since Trade Date: +180%
Marvell Technology (NASDAQ:MRVL) is another favorite stock of Donald Trump that has shown spectacular performance since his purchase in February 2026. It makes custom AI chips, networking solutions, and semiconductor products designed for hyperscalers and data centers. Its customers include the world's largest cloud and technology companies including Google, Amazon, and other major hyperscalers that are increasingly turning to custom ASICs as a more cost-effective way to scale computing power for AI clusters.
Trump isn't the only fan of Marvell Technology (NASDAQ:MRVL). Recently, Nvidia CEO Jensen Huang publicly endorsed Marvell as a potential trillion-dollar company, causing the stock to surge 45% in a single week. Nvidia also made a $2 billion strategic investment in Marvell, and the two companies are collaborating on NVLink Fusion, which makes Marvell's custom XPUs compatible with Nvidia's broader AI infrastructure stack including Vera CPUs, ConnectX NICs, BlueField DPUs, and Spectrum-X switches.
But beyond the Huang factor, there are strong fundamental growth catalysts for Marvell Technology (NASDAQ:MRVL). The company recently unveiled the first switch operating at a record speed of 102.4 Tbps designed for AI clusters, delivering energy savings of up to 25% compared to competitors. It also acquired Polariton Technologies to strengthen its photonics-based optical solutions, and has secured over 10 XPU attach program wins with custom revenue expected to grow 20% year-over-year in FY2027, doubling again in FY2028, and nearly tripling by FY2029 towards more than $10 billion.

Arista Networks, Inc. (NYSE:ANET) is one of the best growth stocks to buy, according to billionaire Ray Dalio's Bridgewater **** ociates. On June 10, Wolfe Research reiterated an Outperform rating on Arista Networks (NYSE:ANET) and a $175 price target.
The positive stance follows meetings with the company's CFO, Chantelle Breithaupt, during which it became clear the company is staring at a favorable data-switching environment. According to the research firm, the company is well-positioned to capitalize on its role as cloud providers, NeoClouds and LLM builders remain focused on GPU utilization rates to reduce token costs.
Arista Networks is increasingly seeing traction as customers shift from bundled Nvidia solutions to best-of-breed networking options. Similarly, the emergence of non-NVIDIA accelerators for inference applications and the increased deployment of Ethernet are positive for the company. That's because the company continues to excel in both high-entropy inference and low-entropy training environments. Therefore, it remains well-positioned with large customers, including Anthropic and Google.
Arista Networks, Inc. (NYSE:ANET) builds cloud networking solutions, including high-speed Ethernet switches and routers. They dominate the data center market by providing the high-bandwidth, low-latency infrastructure required by cloud **** ans like Microsoft and Meta Platforms, and are a critical player in powering high-performance AI training clusters.
While we acknowledge the potential of ANET as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.

Broadcom Inc. (NASDAQ:AVGO) is one of the best AI networking stocks to buy according to ******* ysts. The company gave investors a fresh AI infrastructure catalyst on June 9, when it announced the AI XPV Platform with Apollo and Blackstone as initial anchor investors.
The platform is designed to enable more than 20 gigawatts of compute capacity through 2028 using Broadcom's XPUs and networking solutions customized for frontier AI labs, including Anthropic and OpenAI. It is launching with an initial $35 billion tranche, led by Apollo in partnership with Blackstone, to support Anthropic's planned expansion of more than 1 gigawatt of compute infrastructure at Fluidstack-based sites, starting in mid-2026.
Image by drobotdean on Freepik
That makes Broadcom one of the cleanest fits for an AI networking list because the story is not only about custom accelerators. The company's AI infrastructure roadmap also depends on networking solutions that connect accelerators, servers, and large-scale AI clusters for training and inference. Broadcom said the platform creates a scalable framework for future XPU-based compute capacity and networking aimed at enabling frontier model training and inference at lower cost and power.
Broadcom's latest results reinforced the momentum. On June 3, the company reported fiscal second-quarter revenue of $22.19 billion, up 48% year over year, while AI semiconductor revenue reached $10.8 billion, up 143%, driven by demand for custom AI accelerators and AI networking.