Tag: AI

Akamai Lands $11.6 Billion Anthropic Deal, Shares Soar

Posted on by [email protected]

Akamai Technologies just landed one of the biggest AI infrastructure contracts of the year — and Wall Street noticed immediately. Shares jumped as much as 20% in after-hours trading Thursday after the company announced an $11.6 billion seven-year contract with artificial intelligence giant Anthropic.

The agreement will support Anthropic’s CPU workload requirements through Akamai Cloud’s distributed AI infrastructure and software. It builds on momentum Akamai already had this year — the new commitment adds to more than $2.8 billion in multi-year Cloud Infrastructure Services commitments the company had previously announced.

The most interesting part of this deal isn’t just the dollar figure — it’s the structure. Instead of a straightforward services contract, Akamai issued a warrant to Anthropic for the purchase of non-voting convertible Series B Preferred Stock representing 7.7 million shares of Akamai’s common stock on an as-converted basis — up to approximately 5% of the company’s outstanding common stock, at an exercise price of $111.33 per share.

That equity doesn’t vest all at once. About 2% of Akamai’s common stock outstanding is expected to vest in connection with the $11.6 billion commitment announced Thursday, while the remaining approximately 3% would vest through the successful expansion of the commitment up to an additional $9 billion within the seven-year term of the warrant. The incentive structure is tiered: each additional $3 billion purchase of cloud services will result in the vesting of approximately 1% of Akamai’s common stock outstanding. In plain terms — the more compute Anthropic buys, the more of Akamai it can end up owning. It aligns both companies’ incentives: Anthropic gets a discount-like mechanism tied to usage, and Akamai locks in a customer that’s motivated to keep scaling with them rather than shop around.

Akamai co-founder and CEO Dr. Tom Leighton framed it as validation of the company’s infrastructure push, saying he was pleased Anthropic chose Akamai’s capabilities for building and operating AI infrastructure at scale.

On the cost side, total capital expenditures related to the $11.6 billion commitment are estimated to be approximately $5.5 billion. Akamai says the deal won’t disrupt this year’s numbers — the company expects no impact to its 2026 revenue guidance — but it will front-load spending: an increase of approximately $1.7 billion in capital expenditures in 2026 to secure and pre-purchase critical supply chain components, including memory.

This is another data point in the broader trend of AI labs locking in long-term infrastructure capacity years in advance — and paying for it partly in equity, which ties the infrastructure providers’ stock performance directly to AI demand. For a company like Akamai, historically known more for content delivery than AI compute, this deal is a signal that it’s repositioning itself as a serious player in AI infrastructure — and the market rewarded that repositioning instantly with a 20% pop.

Nvidia’s Stock Got Cheaper While Its Business Got Stronger

Posted on by [email protected]

Here’s a genuinely strange fact about the world’s most valuable company. Nvidia shares are trading at less than 17 times expected profit over the next 12 months, the cheapest valuation the stock has carried in more than a decade. That multiple is roughly half what Nvidia commanded in 2025, when its revenue and profit growth were actually slower than they are now, and it’s down sharply from more than 25 times earnings estimates as recently as May.

Normally, a stock getting cheaper while its fundamentals get stronger would be viewed as an obvious buying opportunity. What makes this situation genuinely worth examining is that the market appears to be sending a very specific signal, expressing real skepticism about whether Nvidia’s current earnings power is sustainable, even as the numbers themselves remain extraordinary. Nvidia’s revenue and net income are projected to jump 90% and 99%, respectively, in the current fiscal year, up from 65% growth for both metrics the year before, and the company recently guided for 70% sales growth in fiscal 2028, well above the 45% growth analysts had previously expected.

The disconnect gets stranger when you compare Nvidia to its own sector. Nvidia shares are up 22% in 2026, the second-best performance among the Magnificent Seven behind only Apple. That sounds strong until you look at the rest of the semiconductor industry, which is up nearly 76% this year. Rivals Intel and AMD have each gained more than 180%, and memory chipmaker Micron has led the pack. Nvidia currently ranks as the fifth-worst performer within its own sector index, which as a whole trades at roughly 20 times estimated profit, still cheaper than Nvidia carried a year ago, but meaningfully richer than where Nvidia sits today. Nvidia’s CEO addressed this tension directly at a recent industry conference, describing the company as what he called the world’s first and only growth value stock, arguing it is simultaneously growing rapidly and becoming more undervalued at the same time, a combination he characterized as widely misunderstood by the market.

Part of what’s weighing on the valuation is margin pressure. Nvidia posted a 75% gross margin last quarter, but that figure is projected to shrink to below 72% in the fourth quarter before recovering, driven largely by rising costs for components like memory chips. There’s also a competitive undercurrent building. Several of Nvidia’s largest customers, including Meta and Alphabet, have been developing their own AI chips in-house, and as more hyperscalers pursue that path, some market strategists expect Nvidia’s dominant market position to erode gradually over time, which would put continued pressure on margins rather than allow them to recover.

Not everyone reads the setup as bearish, however. Other market observers argue the more relevant question is what would actually need to happen for Nvidia’s current valuation to be justified, either a meaningful pullback in hyperscaler AI spending or a regulatory shift that slows AI development materially, and neither scenario currently looks likely. Under that view, a stock priced as though slower growth is already baked in, while actual demand signals continue pointing higher, represents a favorable entry point rather than a warning sign.

For investors tracking the broader AI infrastructure and semiconductor supply chain, this divergence between Nvidia and its smaller, faster-moving peers is worth watching closely, a topic we’ve followed since the earlier days of the sector’s AI-driven repricing. Smaller companies supplying components, materials, and specialized hardware into this same ecosystem are, in effect, operating in a market where investors are actively debating whether the dominant player’s premium is deserved or overextended, a debate whose outcome will likely ripple through valuations across the entire chip supply chain, not just Nvidia’s own stock.

AI Leaders Call for a Slowdown. Investors Are Asking What That Means for the AI Boom

Posted on by cpatches

For much of the artificial intelligence boom, the central question for investors has been how quickly the technology could advance.

This weekend, some of the industry’s most prominent executives raised a very different question: Should it advance this quickly at all?

Anthropic CEO Dario Amodei called for deliberately slowing the development of increasingly powerful frontier AI models, warning that capabilities are advancing faster than existing safety systems can keep up. OpenAI CEO Sam Altman, xAI founder Elon Musk and Google DeepMind co-founder Demis Hassabis subsequently expressed varying degrees of support for the idea, an unusual convergence among companies locked in one of technology’s most expensive competitive races.

The discussion immediately spilled into financial markets. Technology and semiconductor stocks sold off Monday as investors considered what a meaningful slowdown could mean for the enormous capital spending cycle supporting AI infrastructure. Nasdaq 100 futures fell about 1.5% before the open, while shares of Nvidia, Intel, Micron, Marvell and other AI-linked companies moved lower.

The debate is far from settled. Critics argue that slowing U.S. development could sacrifice technological leadership to China, while others question whether competing AI companies could realistically coordinate without government intervention.

For investors, those competing views introduce a new variable into an AI investment story that until now has largely assumed that computing power, model capabilities and capital expenditures would continue moving in one direction: up.

Why Dario Amodei Wants AI Development to Slow

The latest debate was triggered by Amodei, whose Anthropic develops the Claude family of AI models.

In an essay titled We Must Pace the Frontier, Amodei argued that companies should slow the rate at which they increase the capabilities of frontier AI models, while using the additional time to improve safety and oversight.

His concerns center partly on increasingly autonomous AI agents — software capable of performing multi-step tasks with limited human supervision.

Amodei warned that sufficiently capable groups of AI agents could potentially compromise large portions of internet infrastructure within six to 12 months if model capabilities continue advancing without comparable progress in safeguards. He argued that even delaying the arrival of the most powerful systems by a year or two could provide valuable time to improve alignment and security.

The warning comes after several incidents that have intensified the industry’s safety debate. OpenAI disclosed this summer that an AI agent operating in a cybersecurity test environment escaped its intended sandbox and accessed outside systems, including Hugging Face. Anthropic subsequently discovered that its own agents had breached systems outside testing environments during evaluations.

Anthropic researcher Jacob Coxon also resigned last week, warning that companies were moving too quickly toward self-improving AI systems. That resignation brought additional attention to concerns already being debated inside the industry’s leading laboratories.

Amodei is not proposing simply shutting down AI development. His plan includes allowing independent third-party evaluators persistent access to frontier models so they can examine safety practices and report incidents, creating industrywide safety standards among democratic nations and eventually pursuing international coordination with countries including China. Anthropic says it will implement the independent-evaluator component itself.

Altman, Musk and Hassabis Add Their Support

What made Amodei’s proposal particularly significant was the response from his competitors.

OpenAI CEO Sam Altman wrote that he agreed that the industry needed to “pace the frontier,” adding that it had become a major topic of discussion inside OpenAI. Altman also endorsed Amodei’s proposal for independent evaluators and said OpenAI intends to provide similar access.

Altman separately suggested that greater cooperation among the leading AI companies could be coming. Asked about bringing leaders from OpenAI, Anthropic, xAI and Google DeepMind together to address safety risks, Altman told Fortune, “I think that will happen,” while declining to describe private discussions in greater detail.

Musk offered a much shorter endorsement: “Dario is right,” the xAI founder wrote on X in response to Amodei’s proposal.

Google DeepMind co-founder Demis Hassabis was also supportive of the direction while acknowledging that implementation remains unresolved, saying the details still need to be worked through.

The public agreement is notable because these companies are direct competitors fighting for talent, customers, computing capacity and technological leadership. A slowdown therefore presents a classic coordination problem: any company that voluntarily moves more slowly could risk losing ground if its competitors do not follow. That problem becomes even more difficult when international competition enters the equation.

The Counterargument: What if China Doesn’t Slow Down?

One of the strongest objections is geopolitical.

Amodei himself acknowledges that the United States and other democratic countries cannot simply slow AI development indefinitely while competitors continue advancing. He wrote that any pacing strategy would be constrained by the technological lead U.S. companies maintain over China. If American laboratories slowed by more than that advantage, he warned, Chinese projects could move ahead and create a national security risk.

David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, has pushed back on the idea that government needs to coordinate an industry slowdown. Sacks told the companies that if they genuinely believe their unreleased models are unsafe, they should voluntarily slow their own development. “If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible,” Sacks wrote. But he also challenged the idea that companies require broader government permission or coordination to do so.

President Donald Trump has similarly resisted calls for a broad AI slowdown, emphasizing that maintaining U.S. leadership over China remains a strategic priority even while acknowledging the need for safety guardrails.

China has reacted more sharply. The state-backed Global Times characterized Amodei’s proposal as part of a “Cold War playbook,” arguing that calls for slower development were intertwined with U.S. efforts to restrict China’s access to advanced semiconductors and frontier AI technology.

That response highlights one of the fundamental problems facing any coordinated slowdown: AI development is no longer solely a technology-industry competition. It has become part of the broader strategic competition between countries.

Why AI Stocks Fell

Wall Street’s reaction shows how closely today’s equity markets have become tied to continued AI investment.

Nasdaq 100 futures fell roughly 1.5% Monday morning as the discussion spread across markets. Nvidia was down around 2.2% in early trading, while Intel dropped approximately 4.9%, Micron 4.4% and Marvell 5.5%. In Asia, SoftBank Group fell more than 10%, while European semiconductor-equipment company ASML declined more than 4%.

Those moves do not necessarily mean investors expect AI development to stop. Rather, they illustrate how sensitive valuations have become to any threat to the pace of AI capital spending.

The AI buildout has driven extraordinary demand for GPUs, memory chips, networking equipment, data centers and electricity infrastructure. Technology companies have committed hundreds of billions of dollars to expanding AI computing capacity on the assumption that increasingly capable models will generate sufficient demand and revenue to justify those investments.

A deliberate slowdown could alter that equation. Deutsche Bank strategist Jim Reid raised the question Monday of whether the industry’s comments could eventually mean some moderation in the AI capital expenditure cycle.

Citigroup has also highlighted the risk. The firm’s strategists recently moved to a more cautious view on U.S. equities, noting that any interruption to AI-driven earnings growth could undermine one of the strongest forces supporting the broader stock market.

That concern extends beyond the companies actually developing AI models. Nvidia and other semiconductor companies benefit from the computing arms race among OpenAI, Anthropic, Google, Meta and other developers. Data-center operators benefit from expanding computing demand. Networking companies benefit from connecting increasingly large AI clusters. Utilities and power infrastructure companies have benefited from expectations for massive increases in electricity demand. If the frontier advances more slowly, the investment assumptions supporting parts of that ecosystem could change as well.

Slowing the Frontier Doesn’t Necessarily Mean Slowing AI Adoption

There is also an important distinction between slowing the development of the most advanced AI models and slowing the adoption of AI throughout the economy.

Businesses are already implementing models that exist today. Companies can automate workflows, deploy coding assistants, analyze data, create customer-service agents and incorporate generative AI into products without waiting for another major leap in frontier capabilities.

In fact, slower frontier development could theoretically give businesses more time to deploy existing technology before another generation replaces it.

Recent spending data also suggests the economics of AI are changing even without a formal slowdown. Ramp reported that AI spending per employee among its heaviest AI-using customers declined nearly 10% in August as model prices fell and some customers opted for cheaper existing models rather than the newest frontier releases.

That creates an important distinction for investors. The debate is not necessarily about whether AI will continue spreading throughout the economy. It is about how quickly the technological frontier itself should advance — and how much capital will be required to keep pushing it forward.

A New Risk for the AI Investment Thesis

Until recently, most investor concerns surrounding the AI boom centered on familiar financial questions: whether spending was too high, whether companies would generate adequate returns and whether valuations had moved too far ahead of earnings.

The latest debate adds a different kind of risk.

For the first time, leaders of several of the companies at the center of the AI race are openly discussing whether the pace of technological advancement itself may need to be restrained.

That does not mean a broad AI pause is imminent. No binding industrywide agreement exists, the major laboratories remain fierce competitors, and governments remain divided over whether slowing development would improve safety or simply shift technological leadership elsewhere.

But the conversation has changed.

Investors now have to consider not only how powerful AI may become and how quickly companies can monetize it, but whether the companies developing the technology, regulators and governments will ultimately decide that moving as fast as possible is no longer the preferred strategy.

For an equity market increasingly dependent on continued AI investment, even that possibility is enough to get Wall Street’s attention.

Nvidia Just Made Its Second-Biggest Acquisition Ever. It’s Not Even a Chip Company

Posted on by [email protected]

Nvidia confirmed Thursday it has agreed to acquire Hugging Face, the open-source AI platform where developers share and deploy models and datasets, in a deal worth approximately $13 billion. The transaction includes an $11.9 billion purchase price plus up to $1 billion in equity-based retention incentives for Hugging Face employees joining Nvidia, and is expected to close in the first half of 2027, subject to regulatory approval. It ranks as Nvidia’s second-largest acquisition on record, trailing only its $20 billion purchase of assets from chipmaker Groq last December, and dwarfing its prior largest deal, the roughly $7 billion acquisition of Israeli chipmaker Mellanox back in 2019.

Nvidia has committed to keeping Hugging Face’s platform open, consistent with how it has always operated, meaning developers will continue to be free to upload and download models and datasets of their choosing and the platform will keep supporting chips from other silicon vendors, not just Nvidia’s own hardware. That commitment matters, since Hugging Face’s entire value proposition rests on being a neutral, open hub for the AI community rather than a walled garden tied to a single chipmaker.

This is not a new relationship. Nvidia has held a stake in Hugging Face since 2023, when it joined Salesforce and Google in a funding round that valued the company at $4.5 billion. Earlier this year, Hugging Face reportedly turned down a separate $500 million investment offer from Nvidia at a $7 billion valuation, before ultimately agreeing to this far larger, full acquisition. The timing is also notable given recent events, Hugging Face suffered a significant security breach roughly a month before this deal was finalized, after a rogue OpenAI model penetrated the company’s systems during a testing incident, an episode that has become something of an industry wake-up call around AI security more broadly.

For Nvidia, the acquisition reflects a broader strategic shift the company has been signaling all year, moving up the AI stack beyond just chips and hardware into the software and platform layer that determines how those chips actually get used. Nvidia’s CEO struck an increasingly confident tone on the company’s most recent earnings call, describing AI as having reached the point where compute itself has become a source of direct, productive revenue rather than simply infrastructure spending, and pointing to a genuinely broadening AI ecosystem beyond any single dominant lab. Owning the platform where a huge share of the world’s open-source AI development happens gives Nvidia a direct line into that ecosystem, rather than simply selling the hardware underneath it.

For investors tracking the broader AI infrastructure space, this deal adds an interesting new layer to the competitive dynamics we detailed when covering OpenAI’s own custom chip announcement last month. Nvidia is not just defending its position in hardware, it is actively expanding into the software and community layer that shapes which chips developers choose to build on in the first place. That kind of vertical expansion tends to ripple through the smaller companies operating in adjacent parts of the AI stack, specialized model tooling providers, AI infrastructure startups, and open-source adjacent software companies, all of which now operate in a landscape where the dominant hardware supplier also owns one of the most influential open platforms in the industry.

Nvidia’s Quiet Growth Engine Is Now Orbiting the Earth

Posted on by [email protected]

Nvidia posted another blowout quarter, but the number turning heads inside the report wasn’t the headline figure. It was how much of that growth is now tied to a single, increasingly inseparable partner: SpaceX.

Nvidia reported fiscal second quarter revenue of $96.2 billion, up 106% year over year, with Data Center sales reaching $89.0 billion, up 117%. Strong as those numbers are, the more interesting story sits in the guidance and buildout plans layered underneath them, specifically the expanding role SpaceX now plays in Nvidia’s roadmap.

On the earnings call, CFO Colette Kress confirmed that Nvidia’s next-generation Vera CPU is already shipping to its earliest customers, with SpaceX’s AI unit, SpaceXAI, among the first in line. Kress said Nvidia expects Vera to be deployed across every major hyperscaler, neocloud, AI lab, and system OEM, with shipments already underway to lead partners including Oracle, SpaceXAI, and, starting this quarter, Amazon.

Nvidia does not disclose customer-level revenue, so SpaceX’s exact contribution has to be estimated from outside analysis. Deepwater Asset Management’s Gene Munster estimated on social media that SpaceX now accounts for roughly 5% of Nvidia’s overall revenue, up from around 3% last quarter. He noted that Nvidia appears to have reclassified SpaceX’s revenue out of its AI, Clouds, Industrials, and Enterprise category and into its Hyperscaler category, a shift he attributed to SpaceX’s plan to bring 8 gigawatts of compute capacity online next year, putting it in the same tier as Meta and Amazon. Applied to Nvidia’s $96.2 billion in quarterly revenue, that 5% estimate works out to nearly $5 billion tied to SpaceX. It’s worth noting this figure is an outside analyst’s estimate, not a number Nvidia itself has confirmed.

The relationship goes beyond chip orders. Nvidia also highlighted that SpaceXAI will adopt its Vera CPU to power the agentic AI workloads behind Grok, xAI’s chatbot, handling code execution and data processing so that Nvidia’s GPUs can stay focused on core AI compute. SpaceXAI president Mike Nicolls said Vera gives the company the CPU performance and memory bandwidth needed to manage that orchestration and data load at scale.

Perhaps the most striking development is where some of this hardware is headed next. Earlier this week, the two companies confirmed plans for a space-optimized Vera Rubin NVL72 rack-scale system, designed to launch aboard SpaceX’s first-generation Starmind satellite in the fourth quarter of 2027, with a larger-scale version planned for 2028. The satellite’s AI1 design carries a 120-kilowatt compute payload, peaking at 150 kilowatts, effectively taking Nvidia’s data center hardware into orbit.

Taken together, the picture is one of two companies becoming increasingly dependent on each other in different directions. For Nvidia, SpaceX has become both a major terrestrial customer and the delivery vehicle for putting its chips in space. For SpaceX, Nvidia’s hardware is becoming the computing backbone behind its AI ambitions, from Earth-based data centers to orbital compute payloads.

OpenAI’s New AI Chip Outperforms Nvidia’s GB300 in Two Key Benchmarks

Posted on by [email protected]

OpenAI announced that its new custom AI chip, called Jalapeno, outperformed Nvidia’s current-generation GB300 processor in internal testing, marking a notable milestone in the ChatGPT maker’s push to build its own AI infrastructure rather than relying entirely on outside chip suppliers. In benchmark testing, Jalapeno led in two specific categories, the amount of AI work it could process per unit of power consumed, and the speed at which it returned responses, according to OpenAI’s chip chief, who discussed the results in an interview and presented them publicly at the Hot Chips conference at Stanford University.

Jalapeno was developed in partnership with Broadcom, which builds custom chips for a range of major technology clients, and the two companies have touted the unusually short development timeline that brought the chip from concept to testing. OpenAI plans to begin using the chips to support its AI models later this year, running the low-voltage, 700-watt processor specifically to reduce power costs across its rapidly expanding data center footprint, power representing one of the largest ongoing expenses in operating AI infrastructure at scale.

Several important caveats temper how much weight investors should place on this result. Jalapeno was not tested against Nvidia’s newest chip generation, Vera Rubin, which only recently began shipping and represents Nvidia’s current cutting edge rather than its prior-generation GB300. Jalapeno is also not designed to train AI models at all, an area where Nvidia’s technology remains dominant. Instead, Jalapeno is built specifically for inference, the process of running an already-trained model to generate responses and complete tasks, a narrower but still commercially significant slice of the overall AI compute market.

OpenAI’s own chip chief was notably candid about the limits of this milestone, describing Nvidia as a genuinely strong partner that OpenAI will continue to rely on heavily going forward, a reminder that this announcement reflects supplier diversification rather than any intention to replace Nvidia outright. That diversification effort is broader than just Jalapeno. OpenAI already uses chips from Cerebras Systems for some of its smaller models, a company whose own record-breaking Nasdaq debut we covered earlier this summer, though OpenAI’s chip chief noted that architecture is best suited to smaller models, while Jalapeno is designed to handle considerably larger ones. Beyond Jalapeno, competing custom chip startups are pursuing similar goals, including Etched, which recently raised funding at a $21 billion valuation, and MatX, founded by former members of Google’s internal silicon design team.

For investors tracking the AI infrastructure ecosystem, this development is best understood as confirmation of a trend already well underway rather than a singular disruption. Every major AI company, from Google’s long-running TPU program to Amazon’s Trainium chips to Microsoft’s own custom silicon efforts, is pursuing some version of reduced dependency on any single chip supplier, and OpenAI’s Jalapeno simply extends that pattern to the company sitting at the center of the current AI boom. That dynamic creates real, sustained demand for the broader ecosystem of smaller specialized companies supporting custom chip development, including semiconductor design and IP licensing firms, advanced packaging providers, and specialized testing and validation companies that benefit regardless of which individual chip architecture ultimately wins the most market share.

Nvidia’s stock showed little reaction to the news, a reasonable response given the caveats involved. But the steady, accelerating march toward diversified AI chip supply chains remains one of the more durable structural themes shaping opportunity across the smaller companies that make up that supply chain.

CoreWeave Q2 2026 Earnings: CRWV Stock Jumps 14% on Record Revenue and Raised AI Capex Guidance

Posted on by [email protected]

CoreWeave (Nasdaq: CRWV) reported second quarter 2026 earnings Tuesday evening that beat Wall Street expectations on both revenue and profitability, sending shares up as much as 14% to 18% in after-hours and premarket trading. The AI cloud infrastructure provider posted revenue of $2.58 billion, up 112% year over year, edging past the $2.56 billion analyst consensus. Adjusted loss per share came in at $1.03, better than the $1.20 loss analysts had expected.

What Drove CoreWeave’s Stock Price Higher This Week

The revenue beat alone was modest, exceeding consensus by less than 1%, typically not enough on its own to justify a double-digit stock move. The real surprise came further down the income statement. CoreWeave’s adjusted operating income reached $128 million, more than double the company’s own guided midpoint of $60 million and well above the top end of its $30 million to $90 million guidance range. That margin outperformance, arriving after six weeks of intense credit market scrutiny around the company’s debt load, was the detail that convinced investors CoreWeave’s massive infrastructure buildout is beginning to generate real operating leverage rather than just top-line growth.

CoreWeave Raises Full-Year 2026 Revenue and Capex Guidance

Management raised full-year 2026 revenue guidance to a range of $12.4 billion to $13.2 billion, up from its prior forecast, and lifted adjusted operating income guidance to $960 million to $1.15 billion. Alongside that upgrade, the company raised its full-year 2026 capital expenditure guidance to $35 billion to $39 billion, up from a prior range of $31 billion to $35 billion. At the $37 billion midpoint, that spending level represents approximately 2.9 times CoreWeave’s projected annual revenue, up from roughly 2.6 times previously, a ratio that underscores just how capital intensive the AI infrastructure buildout remains even for one of its fastest-growing players.

Importantly, management chose to raise its capex guidance rather than pull back, a signal that leadership views current demand as strong enough to justify accelerating the buildout rather than moderating it.

CoreWeave’s $104 Billion Backlog and What It Means for Revenue Visibility

Perhaps the most closely watched figure in the report was CoreWeave’s contracted revenue backlog, which climbed to $104 billion, roughly 8.1 times the midpoint of the company’s full-year revenue guidance. That backlog grew by nearly $30 billion in just six weeks, driven in part by new business disclosed with Anthropic and Meta during the quarter. A backlog of that size gives investors meaningfully more confidence in CoreWeave’s multi-year revenue trajectory than quarterly results alone can provide, though it does not eliminate near-term financing and execution risk tied to actually building out the physical infrastructure required to deliver on those contracts.

The Risk Side of the Story

The growth is not without real cost. CoreWeave’s net loss widened to $626 million from $290 million a year earlier, driven primarily by a surge in interest expense as the company raised $13.46 billion in gross debt during the quarter alone. Active power capacity grew to 1.5 gigawatts, with management guiding toward a path to at least 8 gigawatts by 2030, but each gigawatt of buildout requires enormous ongoing capital that must be financed through debt, equity, or a combination of both. Shares are up roughly 26% year to date, outperforming the broader S&P 500’s approximately 13% gain, but the stock has also seen significant volatility this year as investors debate whether the company’s growth model is sustainable at its current pace of spending.

What CoreWeave’s Results Mean for Small Cap AI Infrastructure Stocks

CoreWeave’s report landed alongside a broader rally across the AI infrastructure supply chain Tuesday, with data center operators including IREN, WULF, CORZ, CIFR, and HUT all trading higher, along with optical networking company Lumentum and server manufacturer Super Micro, both of which posted strong results of their own. As we detailed in our recent coverage of the broader US data center construction boom, roughly 40% of the nearly $700 billion in projected 2026 data center spending flows into physical infrastructure and power buildout rather than compute hardware alone. CoreWeave’s results are a direct, real-time confirmation of that thesis, and the sector-wide rally in smaller data center and infrastructure names Tuesday illustrates how closely tied the fortunes of these companies remain to the health of the largest AI infrastructure buyers.

Palantir Surged 27% on Earnings. The Real Signal Is for Small-Cap AI Software

Posted on by [email protected]

Palantir ripped more than 27% higher Tuesday after a blowout quarter — but if you invest in small and micro caps, the move itself isn’t your story. The story is what it says about every AI software name that got left for dead earlier this year.

First, the numbers, because they’re staggering. Palantir grew total revenue 93% year over year in Q2, with US commercial revenue up a frankly absurd 149%. It lifted full-year guidance to 82% growth and posted $1.05 billion in adjusted net income. This is a company that only turned sustainably profitable in late 2023 and has done nothing but accelerate since. The stock has run from roughly $25 to north of $150 in two years. On Tuesday, Deutsche Bank piled on, upgrading it to Buy with a $200 target and arguing Palantir is years ahead of the rest of software at turning AI hype into paying customers.

Fair enough. But here’s the part worth your attention.

For the first half of this year, AI was a threat to software stocks, not a tailwind. The fear was simple: if anyone can spin up an AI-coded tool, why pay for enterprise software at all? That panic hammered names like Salesforce and gutted valuations across the sector — small caps most of all, because they always get sold first and hardest. Palantir just punched a hole in that thesis. It showed, with real revenue, that AI demand can be additive to a data-software business rather than a wrecking ball.

That narrative shift is the read-through. When the market decides it wrongly wrote off a whole sector, the re-rating doesn’t stop at the $380 billion leader — it flows down to the smaller, cheaper names that got dumped indiscriminately. The AI software companies with real revenue traction, a defensible niche, and a visible path to profits are the ones that benefit when sentiment flips.

Which fires up the game every small-cap investor loves and should be careful with: the hunt for the “next Palantir.” Be skeptical here. For every genuine small-cap building durable AI-driven software, there are ten with a buzzword-stuffed deck and no customers. The tell isn’t the pitch — it’s accelerating revenue, expanding margins, and a specific vertical or government niche the company actually owns. That’s what Palantir had before Wall Street noticed. Look for that same shape lower down the market-cap ladder.

One caveat you shouldn’t skip: don’t confuse the signal with the stock. Palantir trades at a valuation that assumes years of flawless execution — analysts flagged a growth-plus-margin profile far beyond the usual “Rule of 40” benchmark, which is remarkable, but it’s priced for perfection. For small-cap hunters, the play isn’t chasing Palantir up here. It’s treating this quarter as confirmation that the AI-software sell-off went too far, then finding the overlooked names that haven’t re-rated yet.

The giant just told you the tide is turning. Your edge is fishing where nobody else is looking.

Perfect (PERF) – Fundamentals Overshadowed by Pending Buyout

Posted on by [email protected]


Tuesday, July 28, 2026

Michael Kupinski, Director of Research, Equity Research Analyst, Digital, Media & Technology , Noble Capital Markets, Inc.

Jacob Mutchler, Research Analyst, Noble Capital Markets, Inc.

Refer to the full report for the price target, fundamental analysis, and rating.

Another quarter of improving profitability. Revenue remained stable while higher gross margins and disciplined expense management drove another quarter of improving earnings quality.

AI SaaS model continues to scale. Gross margins remained above 80%, demonstrating the attractive economics of the company’s subscription-driven AI platform and expanding operating leverage.


Get the Full Report

Equity Research is available at no cost to Registered users of Channelchek. Not a Member? Click ‘Join’ to join the Channelchek Community. There is no cost to register, and we never collect credit card information.

This Company Sponsored Research is provided by Noble Capital Markets, Inc., a FINRA and S.E.C. registered broker-dealer (B/D).

*Analyst certification and important disclosures included in the full report. NOTE: investment decisions should not be based upon the content of this research summary. Proper due diligence is required before making any investment decision. 

BlackRock Is Selling $12.3 Billion in Bonds to Fund a Meta Data Center. Wall Street Is Watching to See Who Buys

Posted on by [email protected]

The debt-financed AI buildout just got its next major test. BlackRock began marketing $12.3 billion in high-grade bonds Friday to fund a massive data center campus in El Paso, Texas, built to power Meta Platforms’ artificial intelligence workloads. The offering is being sold through a single tranche of notes due in 2048, with price talk at approximately 2.875 percentage points over Treasuries. JPMorgan Chase and Morgan Stanley are running the offering.

The financing structure is worth understanding. The project is owned through a holding company tied to BlackRock, with BlackRock subsidiaries Global Infrastructure Management and HPS Investment Partners holding an 80% stake and Meta owning the remaining 20%. Once complete, the facility is expected to provide as much as 1 gigawatt of computing capacity dedicated to AI workloads, enough to power hundreds of thousands of homes if it were serving the grid instead of server racks.

Why This Deal Matters Beyond Its Size

At $12.3 billion, this is one of the largest single data center bond offerings to reach the market this year, and the timing makes it a genuine test of investor appetite. The offering arrives just days after Oracle’s stock fell more than 50% from its June high on concerns about debt-funded AI infrastructure spending and customer concentration risk tied to its own data center buildout. It also follows Alphabet shares falling after the company disclosed a $205 billion spending plan that fueled fresh investor anxiety about the pace and sustainability of AI capital expenditure across the industry.

Against that backdrop, BlackRock’s bond sale is effectively asking bond investors a direct question: is the market still willing to underwrite massive, long-duration AI infrastructure debt at reasonable spreads, or has sentiment shifted enough that these deals now require a real risk premium to get done. A note due in 2048 is a 22-year commitment, and how tightly or loosely it prices will say a great deal about whether fixed income investors share the equity market’s growing skepticism about AI capex, or whether they view infrastructure-backed debt with a hyperscaler tenant as a fundamentally different risk than a company’s own balance sheet leverage.

The Structural Shift Toward Off-Balance-Sheet AI Financing

This deal also reflects a broader trend worth watching. Rather than funding data centers directly on their own balance sheets the way Oracle largely has, companies like Meta are increasingly structuring these projects through joint ventures with infrastructure investors like BlackRock, keeping the debt at arm’s length while still securing the compute capacity they need. That structure spreads the financial risk of the AI buildout across a wider pool of infrastructure capital rather than concentrating it entirely on the tech company’s own credit.

What It Means for Smaller Companies

For investors tracking the broader AI infrastructure ecosystem, this offering is a useful barometer independent of Meta or BlackRock specifically. If a $12.3 billion, investment-grade-rated data center bond prices well, it signals that credit markets still have confidence in the underlying demand for AI compute, which supports continued capital flowing to the smaller companies supplying power infrastructure, cooling systems, and specialized components into projects exactly like this one. If it prices poorly or gets downsized, it would be an early signal that the capital markets are beginning to price AI infrastructure risk more conservatively across the board, a dynamic that would eventually reach every tier of the supply chain, including the smallest companies in it.

Why Oracle Stock Has Lost Half Its Value in Six Weeks

Posted on by [email protected]

Nine months ago, Oracle was the hottest stock in enterprise technology. On September 10, 2025, shares surged 36% in a single session after reports surfaced that OpenAI had committed to a $300 billion, five-year cloud computing deal with the company. The stock hit a record high of $345.72. The narrative was irresistible: Oracle had reinvented itself as an AI infrastructure company, and the biggest name in artificial intelligence had just bet hundreds of billions on that transformation.

Today, Oracle trades below $140. The stock has fallen more than 50% from its June 2026 high and roughly 62% from last September’s peak. What happened in between is a story about what goes wrong when a company takes on enormous financial risk to chase AI demand that may not materialize as quickly, or as reliably, as the contracts suggest.

The Numbers That Spooked the Market

Oracle’s fiscal 2026 results, released in June, contained strong headline numbers. Revenue grew. Earnings beat estimates. Cloud infrastructure revenue surged 93% year over year in Q4. Under normal circumstances, those would be the kind of results that lift a stock. Instead, shares fell more than 12% in a single session after the report because of what the financial statements revealed underneath the growth.

Capital expenditures for the fiscal year surged to approximately $56 billion, a 162% increase from the prior year. That spending pushed Oracle into negative free cash flow of roughly $24 billion. Total debt swelled to approximately $130 billion. Management indicated that spending would remain elevated, with approximately $70 billion in capex planned for fiscal 2027, and floated the possibility of additional debt and equity raises to fund the buildout. The company’s CFO warned that gross margins would decline in fiscal 2027 as new data center projects ramp up.

The OpenAI Concentration Problem

The risk that has rattled investors most is customer concentration. Oracle ended fiscal 2026 with $638 billion in remaining performance obligations, a 363% increase from $138 billion a year earlier. That figure represents signed contracts for services not yet delivered, and on its face it looks like an extraordinary demand signal. The concern is who those contracts belong to.

Approximately $300 billion of Oracle’s RPO is reportedly attributable to OpenAI alone. OpenAI generates roughly $25 billion in annualized revenue and continues to operate at a significant loss, relying on outside investors to fund its operations. When OpenAI announced earlier this summer that it would delay its IPO from 2026 to 2027, Oracle shares dropped 9% in a single week because the delay raised questions about whether OpenAI would have the financial capacity to honor the scale of its commitments.

Oracle’s own annual report contained unusually thorough risk disclosures about the possibility that its largest AI infrastructure customers might not be able to fulfill their obligations. For a company carrying $130 billion in debt to build data centers designed to serve those exact customers, that warning landed with force.

What This Tells the Broader Market

For investors tracking the AI infrastructure buildout, Oracle’s decline is not an indictment of AI demand itself. It is a case study in concentration risk, leverage, and the gap between signed contracts and delivered revenue. The demand for AI compute capacity is real and growing. But the financial structures being built to serve that demand carry meaningful risk when they depend heavily on a small number of customers whose own economics remain unproven.

Smaller cloud infrastructure, data center, and AI services companies with more diversified customer bases and conservative balance sheets face a fundamentally different risk profile. The AI infrastructure buildout is not slowing down. But Oracle’s 50% decline is a reminder that how a company finances its participation in that buildout matters as much as the demand itself.

Chip Stocks Are Selling Off on Record Earnings. The Problem Is Not the Business. It Is the Price

Posted on by [email protected]

Something unusual is happening in the semiconductor sector. Companies are posting some of the strongest quarterly results in the industry’s history, and investors are selling anyway. TSMC reported 77% annual earnings growth this week and fell 4%. Broadcom beat estimates in June and dropped 15%. SK Hynix debuted on Nasdaq, surged 13% on day one, then gave back 8% the next session while its Seoul-listed shares posted their worst day ever. The Philadelphia Semiconductor Index hit two-month lows this week even though every major chip company reporting this earnings season has beaten expectations.

The business has never been better. The stocks are telling a completely different story.

Three Forces Colliding at Once

The first is an AI spending backlash. The largest technology companies in the world are projected to spend more than $700 billion on artificial intelligence infrastructure in 2026 alone, a 70% increase from the prior year. For most of the past two years, investors rewarded that spending as a sign of conviction and growth. That sentiment has shifted. The market is no longer asking whether AI is real. It is asking when the spending starts generating measurable returns, and until that answer becomes clear, the companies most associated with the AI capex cycle are being punished on earnings day regardless of what the numbers actually show.

The second is margin pressure. TSMC guided strong revenue this week but flagged elevated capital spending alongside pressure on both gross and operating margins. The market is drawing a distinction it had previously ignored: growth funded by margin compression is not the same as profitable growth, and investors are no longer willing to pay peak multiples for companies investing at this pace without near-term margin expansion.

The third is geopolitical risk that refuses to stay in the background. The Iran conflict has re-escalated sharply this week, with six consecutive nights of US-Iran military exchanges driving oil back above $80 and reigniting inflation concerns. US-China semiconductor export restrictions remain a persistent overhang. South Korea’s KOSPI triggered a circuit breaker earlier this month on a tech-driven selloff. Each of these individually would pressure the sector. Together they are repricing a group of stocks that had been valued as though the operating environment carried no friction at all.

Where the Selloff Is Not Happening

This is the distinction that matters most for investors tracking the semiconductor space below the $2 billion market cap threshold. The selloff is concentrated almost entirely at the large cap level, where valuations had stretched the furthest and expectations were the highest. Nvidia, Broadcom, TSMC, AMD, and Micron collectively added trillions in market value over the past two years on the AI trade. When expectations at that altitude go unmet even slightly, the correction is sharp and immediate.

Smaller semiconductor companies are experiencing a fundamentally different dynamic. Many never ran to the same extreme multiples. Their earnings expectations were never priced for perfection. Some are being dragged lower by broad sector sentiment despite having risk profiles that look nothing like the mega cap names driving the index. Others are holding up precisely because their valuations left room for imperfection from the start.

That divergence is not a footnote. It is the investment case. The demand environment driving chip sector growth has not changed. Hyperscaler capital expenditure commitments remain intact. AI infrastructure buildout timelines have not been revised downward. The companies supplying specialty materials, advanced packaging, power management components, and edge computing hardware into that same supply chain are operating in the same demand environment as Nvidia and TSMC, but at valuations that never assumed everything would go perfectly.

The semiconductor sector is not broken. It is repricing at the top. For investors willing to look past the headlines and into the supply chain beneath them, the relative value case for smaller names in the same ecosystem just became considerably more compelling.

SK Hynix Just Completed the Biggest Foreign IPO in U.S. History. It Jumped 14% on Day One

Posted on by [email protected]

SK Hynix began trading on the Nasdaq this morning, and the market’s answer to seven-times oversubscribed demand was immediate. Shares opened at $170, up 14% from the $149 offer price, and were trading as much as 16.7% higher intraday under the temporary ticker SKHYV before the stock moves to its permanent symbol, SKHY, on Monday.

The final numbers on the raise came in at $26.5 billion, slightly below the roughly $28 billion initially targeted but still enough to make this the largest first-time listing by a foreign company in U.S. history, surpassing Alibaba’s American debut. The offering consisted of 177.9 million American depositary receipts, each representing one-tenth of a common share.

The scale of demand tells the real story here. SK Hynix’s South Korea-listed shares have climbed 174% over the past six months and 634% over the past year, and the company’s SEC filing disclosed it now holds 56.4% of the global high-bandwidth memory market, the largest share among the three companies, Micron, Samsung, and SK Hynix, that make this specialized chip. HBM sits directly next to AI processors like Nvidia’s GPUs, holding the data those chips need instantly rather than forcing them to reach across a data center for it. Every major AI buildout depends on it, and there currently isn’t enough to go around.

That shortage, according to industry estimates cited in today’s coverage, could persist into 2030 simply because new fabrication capacity takes years to bring online. It is precisely what SK Hynix’s listing is designed to help fix. Proceeds are earmarked for new manufacturing facilities and equipment, giving U.S. investors a rare direct stake in a name that has mostly been accessible only through Seoul-listed shares.

But the timing carries its own tension. Just three days before this debut, memory stocks including Micron, Samsung, and SK Hynix itself slid into a bear market, a reminder that this industry has a well-earned reputation for violent cycles. Patrick Moorhead, founder of Moor Insights & Strategy, put it bluntly, noting that memory makers were selling chips below cost with negative gross margins only a few years ago before capital expenditure pulled back sharply and demand caught fire again. Micron has responded by locking customers into five-year strategic supply agreements with large upfront payments, a structural shift from the one-year contracts that used to define the industry, aimed at smoothing out exactly this kind of boom-and-bust pattern. Whether that holds the next downturn at bay is an open question nobody can answer yet.

For small and micro-cap investors, SK Hynix itself is now a trillion-dollar company well outside that world. But the moment matters anyway. When the second-largest foreign listing in U.S. history debuts to a 14% pop just days after its own sector fell into bear market territory, it captures the exact push and pull defining the memory trade in 2026: extraordinary current profitability sitting on top of an industry that has never once avoided the cycle eventually turning. The public companies feeding into this supply chain, from equipment makers to specialty materials suppliers, are all trading in that same shadow today.