The first AI trade was about building the biggest models. The next one is about running them everywhere.
That means inference. Every chatbot answer, search summary, code suggestion, AI agent, phone feature, and enterprise workflow needs compute after the model is trained. The AI buildout is moving closer to the user, and the market is starting to price that in.
Elite Picks (Sponsored)
This report focuses on a narrow group of stocks identified through a detailed screening process.
Analysts apply a combination of metrics to narrow down potential opportunities.
Past selections have shown strong momentum, but no outcomes are guaranteed.
The newest edition is now open for access.
Theme: AI Inference and Edge Compute
This setup works because AI does not stop at training. Training gets the model ready. Inference is where the model gets used.
That matters because inference demand can become more repeatable than training demand. A company may train a major model occasionally, but users can run AI queries all day. Enterprises can use agents across sales, service, coding, analytics, security, operations, and support. Devices can process more AI locally. Networks can route more AI traffic at the edge.
This is where the AI story starts moving from mega data-center projects into daily usage.
What’s Driving It
The numbers are still strong across the compute stack. Nvidia’s data-center revenue remains massive, and the company continues pushing new AI systems designed for both training and inference. AMD reported Q1 data-center revenue of $5.8 billion, up 57%, helped by EPYC processors and the ramp of Instinct GPU shipments. HPE just delivered a record quarter, with revenue up 40% to $10.68 billion, a $6.3 billion AI backlog, and sharply raised FY26 guidance.
Cloudflare adds another angle. Its Q1 revenue rose 34% to $639.8 million, and current RPO also grew 34%. The company sits at the edge of the internet, where AI traffic, bot traffic, security, and developer infrastructure are becoming more important.
Qualcomm rounds out the basket with on-device AI. If more AI processing happens on phones, PCs, cars, and connected devices, Qualcomm has a clear role through low-power chips and connectivity.
Here is the chain reaction:
AI models get deployed → inference workloads rise
Inference workloads rise → compute moves closer to users
Compute moves closer → edge networks and AI servers matter more
Latency and cost become bottlenecks → efficient chips and infrastructure win
AI adoption broadens → inference becomes the next growth layer
What’s Working
What is working right now is the broadening of AI demand beyond training. The first AI capex wave was about massive clusters. The next question is how all that intelligence gets delivered quickly, cheaply, and reliably.
That favors companies with scale, distribution, and infrastructure control. Nvidia still owns the highest-quality AI compute stack. AMD gives customers an important second source. Qualcomm gives you the edge-device angle. Cloudflare gives you edge network and security exposure. HPE gives you AI server and enterprise infrastructure leverage.
This theme is not about replacing the training trade. It is about adding the usage layer on top of it.
What to Watch
You should watch inference margins, enterprise AI adoption, AI server demand, edge traffic, and whether companies can reduce the cost per AI query.
The biggest risk is that inference gets competitive. If too much capacity comes online too quickly, pricing can get pressured. Another risk is that enterprise AI adoption takes longer than expected. Investors are pricing in a world where AI becomes part of daily work. The numbers need to prove that.
Nvidia (NVDA)
What it does:
Nvidia makes GPUs, networking equipment, AI systems, software, and full-stack accelerated computing platforms used in data centers, enterprises, research labs, and AI applications.
Why it fits:
Nvidia remains the anchor of the AI compute trade. It is not just a training stock. The company is also central to inference because its chips and systems power the models after they are deployed.
What stands out:
This is still the highest-quality name in AI infrastructure. Nvidia benefits from training demand, inference demand, networking demand, and the ecosystem effect around its software stack.
What to watch:
Watch data-center revenue, gross margin, Blackwell and next-generation system demand, and whether inference becomes a bigger part of the growth story.
The Takeaway: Buy this first if you want the highest-quality AI compute stock tied to both training and inference.
The risk is valuation. Nvidia can keep winning and still see the stock pause if expectations get too stretched.
Advanced Micro Devices (AMD)
What it does:
AMD makes CPUs, GPUs, adaptive chips, and AI accelerators for data centers, PCs, gaming, embedded systems, and cloud customers.
Why it fits:
AMD is the second-source AI compute play. Its Q1 data-center revenue rose 57% to $5.8 billion, driven by EPYC processors and Instinct GPU shipments. If customers want alternatives to Nvidia, AMD stays in the conversation.
What stands out:
This is the competitive catch-up name. AMD does not need to beat Nvidia to work. It needs to take enough share in AI accelerators and server CPUs to show that the market is big enough for more than one winner.
What to watch:
Watch Instinct adoption, server CPU growth, gross margin, and whether major cloud customers expand AMD deployments.
The Takeaway: Buy this if you want AI compute upside with more catch-up potential than Nvidia.
The risk is that AMD remains the challenger and struggles to close the software and ecosystem gap fast enough.
Standout Picks Now (Sponsored)
Every market cycle produces a handful of companies that dramatically outperform the rest.
Our latest screening has identified the 5 Stocks Set to Double — companies showing rare early-stage momentum traits.
These picks carry the same indicators that historically precede strong rallies.
Past reports highlighted stocks that surged +175%, +498%, and +673%.
Get the Free 5 Stocks Set to Double Report.
*This free resource is being sent by Zacks. We identify investment resources you may choose to use in making your own decisions. Use of this resource is subject to the Zacks Terms of Service.
*Past performance is no guarantee of future results. Investing involves risk. This material does not constitute investment, legal, accounting, or tax advice. Zacks Investment Research is not a licensed dealer, broker, or investment adviser.
Qualcomm (QCOM)
What it does:
Qualcomm designs chips and connectivity platforms for smartphones, PCs, autos, IoT devices, and edge AI applications.
Why it fits:
Qualcomm gives the basket the on-device AI angle. If more inference happens locally instead of in massive cloud data centers, low-power chips become more important. Phones, AI PCs, cars, and connected devices all need efficient processing.
What stands out:
This is the edge-device name. Qualcomm is not trying to own the giant AI training cluster. It is trying to bring AI closer to the user through chips already connected to the device ecosystem.
What to watch:
Watch AI PC adoption, smartphone upgrade cycles, automotive growth, and whether edge AI becomes a real revenue driver instead of a product-demo story.
The Takeaway: Buy this if you want AI inference exposure outside the data center.
The risk is that device cycles stay uneven and on-device AI takes longer to become a major upgrade driver.
Cloudflare (NET)
Cloudflare provides internet infrastructure, network security, zero trust, developer tools, edge compute, and performance services.
Why it fits:
Cloudflare sits close to where AI traffic moves. Q1 revenue rose 34% to $639.8 million, current RPO grew 34%, and the company continues to benefit from demand for security, speed, and edge infrastructure.
What stands out:
This is the edge network and AI traffic name. As AI agents, bots, apps, and users generate more automated traffic, Cloudflare becomes more relevant to routing, protecting, and managing that traffic.
What to watch:
Watch large-customer growth, RPO, AI traffic commentary, and whether restructuring improves operating leverage without slowing execution.
The Takeaway: Buy this if you want edge infrastructure exposure with cybersecurity and AI-traffic upside.
The risk is valuation and execution. Cloudflare still needs to prove that strong revenue growth can turn into durable profit growth.
Sector Incentive Stocks (Sponsored)
Capital gains taxes can take a bigger bite out of your profits than expected.
Fortunately, some deductions may help reduce the impact — including:
Investment-related expenses
Certain real estate selling costs
Because rules and eligibility vary, many investors turn to fiduciary financial advisors for guidance.
Find an Advisor Match.
Which "old economy" industry do you think is most ripe for tech disruption?
Hewlett Packard Enterprise (HPE)
What it does:
HPE sells servers, networking, hybrid cloud infrastructure, storage, and AI systems for enterprises, governments, and large data-center customers.
Why it fits:
HPE gives you enterprise AI infrastructure exposure. The company reported record Q2 revenue of $10.68 billion, up 40%, raised FY26 revenue growth guidance to 29% to 33%, and reported a $6.3 billion AI backlog.
What stands out:
This is the AI server and enterprise refresh name. HPE benefits as companies modernize traditional infrastructure for AI workloads and start moving from experiments to deployments.
What to watch:
Watch AI backlog conversion, networking revenue growth, server margins, and whether enterprise demand remains strong after the post-earnings surge.
The Takeaway: Buy this if you want AI infrastructure exposure with a stronger enterprise-server angle.
The risk is that expectations have reset higher after the big rally, so any slowdown in backlog conversion could hit the stock hard.
Elite Trade Club Insider
$9 Million Sold Near Strength, $1.1 Million Bought Near The Lows
The CEO of a $30 billion genetic testing stock just sold roughly $9.1 million after a strong one-year run, while a director at a beaten-down alternative asset manager bought more than $1.1 million near the stock’s 52-week low. You’re seeing two very different insider signals. Insider readers are seeing where leadership is cashing in on strength — and where one director is buying weakness with real conviction.
You’re reading the free version. Here’s what we held back.
Every day, insiders and institutions move millions before the market catches on. We surface the data behind those moves before the rest of the market sees it.
A subscription gets you:
The insider buys, options bets, and dark pool moves the free edition can't show you. Unlocked every weekday.
A Sunday Deep Dive that tells you where to look before Monday's bell rings.
The Friday Smart Money Brief: who bought, who sold, where the big options bets landed, and where institutions are hiding volume. Three data layers. One email.
A Monthly Insider Scorecard so you always know whether smart money is buying or selling the market.
Every past Insider edition, unlocked, on elitetrade.club. Go back and see what you missed.
$25/mo or $250/yr. 30-day money back guarantee. Cancel anytime. Founding member pricing: lock in $25/mo before we raise it.
This theme works because AI is moving from the lab to the user. Training still matters, but inference is where the technology becomes part of daily life.
Nvidia is the quality compute anchor. AMD is the second-source challenger. Qualcomm is the edge-device play. Cloudflare is the edge-network and security layer. HPE is the enterprise AI infrastructure name.
Stay bullish on the theme, but focus on proof. The next phase of AI needs more than big promises. It needs usage, margins, backlog conversion, and evidence that inference demand can scale profitably.
Best Regards,
— Adam Garcia
Elite Trade Club
Click here to get our daily newsletter straight to your cell for free.
P.S. Just like this newsletter, it's 100% free*, and you can stop at any time by replying STOP.
