Best AI Stocks to Buy Now: 10 Best Performing AI Stocks to Watch in 2026

The best AI stocks of 2026 are not the ones most people would have guessed. The year's greatest gains have come not from chatbot makers or software names but from the infrastructure layer — the memory chips, storage drives, servers, and processors that the AI buildout physically cannot proceed without. A NAND flash specialist spun off barely eighteen months ago is up more than 600% this year; a thirty-year-old PC maker has rallied 345% on AI server orders; and the past week alone saw AMD join the trillion-dollar club as Meta's viral Muse agent rewrote the demand maths for an entire class of chips.

This guide profiles the ten best performing AI stocks of 2026 — ordered from the year's most explosive gainer to its most powerful incumbent — with where each sits in the AI supply chain, what has driven it, and what could go wrong. It closes with the honest risk section this market demands after such a run, and how traders can position across the whole list, long or short, with share CFDs on Markets.com. This is education and analysis, not personal investment advice — every stock here has already run hard, which raises the stakes in both directions.

Key Takeaways

• AI infrastructure stocks have dominated 2026: Sandisk has surged over 600% year to date, Dell about 345%, and Western Digital and Micron roughly 255–260% each.
• The rally broadened beyond GPUs — memory, storage, servers, and CPUs all re-rated as AI demand spread down the supply chain, capped by AMD's $1 trillion milestone on 21 September.
• Nvidia remains the anchor: a $96 billion revenue quarter, $108 billion guided next, and the sector's largest market value by a wide margin.
• The momentum is not risk-free — the memory cohort trades 20–40% below its late-June peaks, and several names now price years of flawless execution.
• Three shared risks hang over the group: stretched valuations, the untested AI capex cycle, and rising bond yields after September's surprise Fed hike.
• All ten stocks are tradable long or short as share CFDs on Markets.com, with a free demo account to practise on before committing capital.

What counts as an AI stock?

The label has stretched to near-meaninglessness — by 2026, almost every listed company claims an AI story — so this guide applies a harder test: material revenue directly tied to the AI buildout, visible in reported numbers rather than press releases. That test admits the companies selling what AI data centres physically consume — accelerators and CPUs (Nvidia, AMD, Intel, Arm's licensed designs), memory and storage (Micron, Sandisk, Western Digital), and the servers and racks that integrate it all (Dell, Super Micro), plus the custom silicon and networking that stitch clusters together (Broadcom). It excludes, for this list, the application layer — software firms using AI, whose 2026 has been far rougher, as the Figma and Adobe sell-off showed — and the private giants (OpenAI, Anthropic) that fuel the demand but cannot be traded. The distinction matters practically: infrastructure names have reported the revenue; much of the software layer is still promising it.

How we selected the best performing AI stocks

The list is built on four filters, applied in order. Performance first: these are the AI names that have actually delivered for shareholders across 2026, ranked broadly from the year's biggest gainer to its established anchor — a deliberate inversion of lists that lead with the household name. Direct AI revenue second: every company here books substantial, disclosed revenue from AI data centre demand — no thematic stretches. Liquidity third: all ten are large-cap US-listed stocks with deep trading volume and tight spreads, which matters for the CFD trader as much as the investor. Supply-chain spread fourth: the ten deliberately cover different layers — memory, storage, servers, CPUs, GPUs, custom silicon — because the layers rotate in and out of leadership, and understanding the map is worth more than any single pick. What the list is *not*: a prediction. Past performance never guarantees future returns, and this year's 600% gainer can be next year's cautionary tale — the risk section below is not decoration.

The AI supply chain: where each stock fits

Picture an AI data centre as a stack, and the ten stocks slot in cleanly. At the top sit the accelerators that run the models: Nvidia's GPUs dominate, AMD's Instinct line challenges, and Broadcom builds custom AI chips for hyperscalers who want alternatives to both. Beside them, CPUs orchestrate everything — AMD's EPYC and Intel's Xeon server chips, with Arm licensing the architecture behind a growing share of custom data centre processors; September's agent-driven rally was precisely the market repricing this layer. Feeding the processors is the memory layer — Micron's high-bandwidth memory sits physically on the AI accelerators, while Sandisk's NAND flash stores the oceans of training data. Below that, mass storage: Western Digital's high-capacity drives hold what the industry cannot delete. And wrapping it all, systems: Dell and Super Micro integrate chips, memory, cooling, and networking into the racks hyperscalers actually buy. The investing insight of 2026 is that scarcity migrated down this stack — GPUs first, then memory, then storage and power — and each migration minted that layer's stocks.

Sandisk (SNDK)

Sandisk is 2026's defining AI trade: a pure-play NAND flash memory maker, spun out of Western Digital in February 2025, whose shares have soared more than 600% year to date — roughly double the gains of any other major memory name. The story is scarcity meeting focus. AI data centres consume staggering quantities of flash storage for training datasets, model checkpoints, and inference caches, and NAND supply — disciplined after years of brutal cycles — could not keep up. As a newly independent company with nothing but NAND exposure, Sandisk became the highest-octane way to own the shortage, and the market treated it accordingly; when Elon Musk publicly flagged memory as AI's next bottleneck in August, Sandisk jumped 8% in a session — the cohort's bellwether reflex.

The risks are the mirror of the reward. Pure-play means no diversification: Sandisk lives and dies with NAND pricing, historically among the most violently cyclical commodities in technology, and the stock already trades 20–40% below its late-June peak alongside the whole memory group — a reminder that supercycle stocks correct like supercycles. New supply eventually answers every shortage, and a 600% year leaves expectations that only a continued shortage can satisfy. Sandisk is the list's highest-beta name: the biggest winner if the memory squeeze runs into 2027, and the most exposed if it breaks.

Dell Technologies (DELL)

Dell is the year's most surprising megacap performer: up roughly 345% year to date after its transformation from steady PC-and-server vendor into the AI buildout's logistics department. The numbers driving it are extraordinary — a record $60.9 billion of AI server orders booked in a single quarter, an AI backlog of $95 billion, quarterly revenue of $47 billion up 58% year on year, and a 16% single-day surge when September's results landed. Dell's edge is unglamorous and real: hyperscalers and enterprises building AI capacity need someone to integrate GPUs, cooling, networking, and support into deployable racks at global scale, and Dell's supply chain does exactly that for both Nvidia- and AMD-based systems.

The bear case centres on margins and concentration. Assembling servers around someone else's $40,000 chips is structurally lower-margin than making the chips, and Dell's first "margin reckoning" headlines have already appeared — the market watches whether AI revenue scale translates into AI profit. The backlog, while enormous, is concentrated among a handful of AI buyers whose ordering patterns can swing quarters violently. And at 345% up, even a valuation still under 22 times earnings — cheap beside its peers on that measure — depends on the capex wave rolling on. Dell is the list's "picks-and-shovels at scale" trade: less explosive than the memory names, more dependent on volume enduring.

Western Digital (WDC)

Western Digital — the storage veteran that sent Sandisk out on its own — has itself rallied roughly 260% year to date, as the AI boom rehabilitated the least fashionable hardware in technology: the hard disk drive. The logic surprised many investors. AI companies hoard data at unprecedented scale — training corpora, model versions, user interactions, none of it deletable — and for bulk "warm" storage, high-capacity hard drives remain several times cheaper per terabyte than flash. Post-spin, Western Digital is a focused HDD company, and nearline drive demand from cloud and AI customers has run ahead of supply for much of the year, with the stock rising 6% on the same August day Musk's bottleneck comment lifted the whole storage complex.

The risks are structural and cyclical at once. Flash keeps closing the cost gap, and the long-term substitution question — will SSDs eventually take even bulk storage? — has never fully gone away; Western Digital is effectively a bet that the crossover stays years out while AI data growth outruns it. Like its spun-off sibling, it trades well below its June peak, and HDD demand is hostage to the same hyperscaler capex line as everything else on this list. It offers the memory-boom exposure with an older-technology discount — and the debates that come with it.

Micron Technology (MU)

Micron is the memory cohort's heavyweight: up roughly 255% year to date, a member of the trillion-dollar chip club alongside Nvidia, Broadcom, and AMD, and the only US maker of the high-bandwidth memory (HBM) that sits physically bonded to AI accelerators. HBM is the supercycle's crown jewel — every advanced AI chip needs stacks of it, capacity is sold out far in advance, and Micron's technology position against Korean rivals SK Hynix and Samsung has never been stronger. Notably, analysts comparing the cohort still call Micron the cheapest of the big three memory names on forward earnings — the rare 2026 AI winner with a valuation argument attached — and the stock has held its uptrend even as the group cooled from June's highs.

The case against is the oldest one in semiconductors: memory is a commodity cycle wearing a growth costume, and every prior supercycle ended with oversupply and a drawdown investors remember painfully. HBM's premium pricing invites exactly the capacity race that breaks premiums, DRAM's spot economics still swing Micron's earnings, and a 20–40% cohort pullback from the June peak shows the market already rehearsing the turn. Micron is the list's "quality cyclical" — the strongest balance of AI exposure, technological moat, and residual valuation sanity, with a cycle that must eventually be respected.

Advanced Micro Devices (AMD)

AMD is the list's newest trillion-dollar member — crossing the line on 21 September 2026 with a near-10% single-day surge to around $611, driven by the Meta Muse agent moment that convinced markets the inference era needs server CPUs as much as GPUs. AMD's uniqueness on this list is owning both sides of that trade: Instinct accelerators (the MI400 generation launches this half, with Helios rack systems already deploying) challenge Nvidia directly, while EPYC server CPUs take share from Intel and stand to benefit from every agentic workload that needs orchestration. Data centre revenue more than doubled year on year to $6.7 billion last quarter, with gigawatt-scale commitments from OpenAI, Meta, and Oracle underwriting the roadmap.

The complications are fresh and specific. After the surge, AMD trades above the average analyst target (~$579) — the Street formally sees limited upside until it re-rates — and its forward multiple now exceeds Nvidia's despite lower margins and a fraction of the AI revenue. Everything therefore rides on execution: MI400 benchmarks, deployment schedules, and data centre growth holding near management's 60% bar.

Arm Holdings (ARM)

Arm is the architecture landlord of the AI era: it designs the CPU blueprints that others license, collecting royalties on chips built by Apple, Nvidia (whose Grace CPUs are Arm-based), Amazon, Microsoft, and Google's custom silicon programmes. The stock was up 167% year over year by May, trading above $300, and after a summer pullback it has become September's designated momentum name — surging 8%, 13%, and 17% on separate sessions this month as the agentic-AI narrative reignited the CPU trade. The bull case is elegant: whichever company wins AI, an increasing share of data centre CPUs carry Arm's architecture, and every agent-driven expansion of CPU demand lifts Arm's royalty base without Arm building a single fab.

The price of elegance is multiple risk. Arm trades on royalty streams at software-like valuations — among the richest multiples in the sector — which makes it the most rate-sensitive and sentiment-sensitive name on this list, capable of double-digit swings in both directions on narrative alone, as this month has demonstrated. Licensing disputes, custom-silicon insourcing by big customers, and RISC-V's slow advance as a free alternative are the structural watch items. Arm is the purest way to own the "CPUs matter again" thesis — and the least forgiving if it wobbles.

Super Micro Computer (SMCI)

Super Micro is the AI server specialist that turns new chips into shippable systems faster than anyone: fiscal 2026 revenue reached $39.06 billion, up 78%, with earnings more than doubling (+109%) — growth rates that would headline any other list. Its edge is speed and thermal engineering: SMCI routinely ships systems built on Nvidia's and AMD's newest accelerators months ahead of larger rivals, and its liquid-cooling expertise matters more with every generation of hotter chips. For traders, it is the sector's established high-beta vehicle — historically amplifying the AI complex's moves in both directions, rallying 7% on good sector days and falling harder on bad ones.

Its risks are equally established. Server integration is the stack's thinnest-margin layer, and SMCI competes directly with Dell's scale on one side and hyperscalers' own designs on the other. The company also carries a governance history — accounting questions and auditor drama in 2024–25 — that resolved without catastrophe but left a discount and a sensitivity to any accounting headline that peers don't share. SMCI is the trader's pick of the list: maximal exposure to AI hardware momentum, minimal margin for anything going wrong.

Intel (INTC)

Intel is 2026's redemption arc — the fallen giant that the AI boom, government backing, and a brutal restructuring have hauled back into relevance. Second-quarter revenue rose 25% year on year to $16.1 billion, with data centre and AI revenue up 59% and the foundry business growing 31%; the stock trades near $111 after September sessions of +10% and +12% as the agent-driven CPU trade lifted every x86 name. Intel's case is optionality stacked on optionality: server CPU demand recovering just as agents expand the market, a foundry business positioned as the West's manufacturing alternative to TSMC — backed by an unprecedented US government equity stake and Nvidia's 2025 investment — and a valuation that, even after the rally, prices scepticism rather than success.

The scepticism exists for reasons. Intel still trails AMD on server CPU performance-per-watt and has no accelerator to answer Instinct or Nvidia's GPUs; the foundry bet consumes billions with customer wins still early; and the turnaround's history is littered with false dawns that punished believers. Intel is the list's value-and-policy play — the cheapest exposure to the CPU renaissance, carrying the most execution risk per dollar of story.

Broadcom (AVGO)

Broadcom is the quiet giant of the AI trade — a trillion-dollar-plus company monetising the two layers everyone else forgets: custom AI accelerators built for hyperscalers who want alternatives to Nvidia (Google's TPUs began this model; OpenAI's 10-gigawatt custom-chip partnership extended it), and the networking silicon that binds thousands of chips into one usable cluster. The scale is staggering: AI bookings exceeded $30 billion in a single quarter against $10.8 billion shipped, and management has reiterated expectations of more than $100 billion of AI revenue in fiscal 2027 — an order book that would rank among the largest in technology history, wrapped in software-division cash flows (VMware) that steady the whole.

Broadcom's September also demonstrated its risk profile: results that merely met elevated expectations sparked a sector-wide sell-off, dragging Intel, AMD, and Micron down with it — the burden of being priced for acceleration. Customer concentration is extreme (a handful of hyperscalers and one AI lab dominate the order book), custom silicon programmes can be cancelled or insourced, and the stock's size means it moves the whole index it lives in. Broadcom is the list's institutional core holding — the widest moat after Nvidia's, with expectations to match.

Nvidia (NVDA)

Nvidia remains the sun the entire list orbits: the dominant AI accelerator maker, the club's founding trillion-dollar member (now worth several times any other name here), and still — remarkably — growing like a startup. The latest quarter delivered $96 billion of revenue, $89 billion of it from data centres, with guidance to roughly $108 billion next quarter; the stock trades near $228, up about 22% year to date with a 16% surge in the past month. That YTD figure — the smallest on this list — is itself the story: Nvidia has become the sector's measured incumbent, compounding at merely spectacular rates while the supply chain beneath it produces the year's fireworks. Its moat remains the industry's deepest: CUDA software lock-in, a relentless product cadence, and an ecosystem every AI company must build on or around.

The risks are those of empire. Every other company on this list is, in part, an attempt to reduce dependence on Nvidia — AMD's challenge, Broadcom's custom chips, hyperscaler silicon — and share loss at the margin is the structural expectation, not the surprise. Its revenue concentrates in the same few AI buyers as everyone else's, making it the ultimate expression of the capex-cycle bet, and its sheer index weight means macro flows move it regardless of execution. Nvidia is the list's anchor: the least likely to double from here, and the name the whole AI trade still cannot function without.

Best performing AI stocks compared

The ten at a glance, in the list's running order:

Two honest footnotes to the table: the memory and storage names (SNDK, WDC, MU) sit 20–40% below their late-June peaks despite the YTD figures, and every number above ages quickly — check live prices before acting on any of it.

Why AI infrastructure stocks have outperformed AI software

The pattern behind this entire list deserves its own explanation, because it inverted most 2024-era predictions. The consensus then held that hardware would commoditise while software captured AI's value; 2026 delivered the opposite — this list's hardware names compounding while the application layer suffered the year's ugliest de-ratings, Figma down more than 80% from its 2025 high among them. Three mechanics explain it. First, infrastructure gets paid now: when Meta or OpenAI budgets tens of billions for capacity, Nvidia, Dell, and Micron book revenue this quarter, while software firms must still convert AI features into subscriptions customer by customer. Second, scarcity prices better than abundance: compute, memory, and power are physically constrained — hence pricing power — while AI makes software easier to produce, threatening the very scarcity software multiples were built on. Third, the spenders are the disruptors: the companies funding the buildout are the same ones whose AI products menace the application layer, so capex flows from software's future margins into hardware's present revenue. The rotation will not run forever — infrastructure booms end when capacity catches demand — but understanding it explains both this list's composition and its central risk.

Risks of investing in AI stocks

After a year like this, the risk section is the most valuable part of the article — three risks in particular apply to every name above.

Valuation and expectations risk

Great companies and great stocks diverge when prices assume perfection. Several list members now trade above or near analyst consensus targets (AMD most visibly), on multiples that require years of uninterrupted hyper-growth; Broadcom's September demonstrated the mechanism, when meeting expectations triggered a sector sell-off. The memory cohort's 20–40% retreat from June peaks is the same lesson in chart form: nothing repriced fundamentally — expectations simply got ahead of even extraordinary results. High valuations don't cause crashes, but they remove the safety net when anything else does.

AI capex slowdown risk

Every revenue stream on this list traces to the same source: a handful of hyperscalers and AI labs spending historically unprecedented sums on infrastructure, funded by investor patience with returns that remain mostly ahead. This cycle has never been through a downturn. If AI application revenue disappoints long enough — or one major spender announces "digestion" — orders, backlogs (even $95 billion ones), and supercycle pricing would reprice together, and the correlation that lifted these ten stocks as a bloc would work in reverse. It is the shared, undiversifiable risk of the whole theme, and the reason position sizing across several AI names is not the diversification it appears.

Interest rates and bond yields

The macro lever is live again: September's surprise Fed hike to 3.75–4.00% reminded markets that the discount rate on 2030 promises can rise. Growth stocks of exactly this kind — long-duration, expectation-heavy — are mechanically the most sensitive to yields, and the highest-multiple names on this list (Arm above all) are the most sensitive of the sensitive. A sustained move higher in bond yields could compress AI valuations across the board without a single order being cancelled; a return to cuts would do the opposite. Trading this list means watching the Fed calendar as closely as the earnings one.

How to Trade AI Stocks With Markets.com

Markets.com offers share CFDs that let traders speculate on price movements without owning the underlying shares. If an AI-related company’s share CFD is available in your region, you can take a Buy position if you expect its price to rise or a Sell position if you expect it to fall.

What you’re actually trading: A share CFD is a contract based on changes in a company’s share price. It does not give you ownership of the shares or voting rights. CFDs use margin, which can magnify losses as well as gains. Check the spread, margin requirement and potential overnight financing charges for each instrument.

Step 1: Open an Account

Create a Markets.com account and review the account terms and CFD risk information.

Step 2: Verify Your Identity

Provide the personal information, risk assessment responses and identity documents requested during verification. You can use a demo account to practise before trading with real funds.

Step 3: Fund Your Account

Once your account is verified, review the funding options available to you. Deposit only after considering the margin needed for your planned position and possible price movements.

Step 4: Find an AI Stock and Place the Trade

Search for the company you want to trade and confirm that you have selected its share CFD. Review the instrument’s trading conditions, choose a position size, then select Buy or Sell and check the order before submitting it.

Step 5: Manage Your Risk

Plan your position size, stop-loss and potential take-profit level before opening the trade. Earnings reports, AI spending forecasts, product announcements and changes in market sentiment can cause sharp price moves. Stop-loss orders can be affected by slippage or market gaps, so continue monitoring open positions.

New to Markets.com? Claim a generous deposit bonus on your first trade. Hurry—this offer is only available for a limited time.

Final words on best AI stocks to buy now

The best AI stocks to buy now, if 2026's tape is the judge, have been the ones building the boom rather than promising its future — memory, storage, servers, and silicon delivering 200–600% gains while the software layer struggled to monetise. The list's shape tells the strategy: Sandisk, Dell, Western Digital, and Micron rode scarcity; AMD, Arm, and Intel caught the agent-era CPU repricing; Super Micro amplified it all; Broadcom and Nvidia anchored it with order books measured in tens of billions. The same tape carries the warnings — June's memory peak, September's expectations sell-off, a Fed hike aimed at exactly these multiples — and the honest conclusion that follows: this is a theme for the informed and the risk-managed, not the fear-of-missing-out. Build the watchlist on a Markets.com demo account, learn how the bloc trades through a real earnings cycle, and position long or short with stops and sane sizing. The AI buildout is the market story of the decade; trading it well means respecting that stories this good are never priced cheap.

Best AI stocks to buy now FAQs

Below are direct answers to the questions traders ask most about the best AI stocks to buy now.

What are the best performing AI stocks of 2026?

By year-to-date share price gains: Sandisk (+600%+), Dell (~+345%), Western Digital (~+260%), and Micron (~+255%) lead, with AMD's $1 trillion milestone, Arm's agent-trade surges, Super Micro's 78% revenue growth, Intel's turnaround, Broadcom's $30B+ AI bookings, and Nvidia's $96B quarters completing the ten.

Why have memory and storage stocks outperformed?

AI data centres consume enormous quantities of high-bandwidth memory, NAND flash, and hard-drive capacity, and supply could not keep pace — a squeeze Elon Musk publicly called AI's next bottleneck. Scarcity gave the memory names pricing power, though the cohort has pulled back 20–40% from its June peaks.

Is Nvidia still the best AI stock?

It remains the dominant company — $96 billion quarterly revenue, $108 billion guided, and the deepest moat in the sector — but its 22% YTD gain trails every other name on this list. The market's 2026 lesson is that the biggest company and the biggest stock gains are different questions.

Are AI stocks overvalued in 2026?

Several trade above analyst targets on multiples that assume years of flawless growth, and September showed merely meeting expectations can trigger sell-offs. That is not a crash prediction — it is a reason the risk section of this article matters as much as the profiles.

What is the biggest risk to AI stocks?

An AI capex slowdown. Every name here ultimately depends on a small group of hyperscalers and AI labs sustaining historic infrastructure spending; any digestion phase would hit the whole correlated bloc at once. Valuations and rising bond yields amplify whatever the cycle does.

How can I trade AI stocks?

All ten are available as share CFDs on Markets.com — long or short, without owning shares. Start on the free demo account, size positions for 10%-day volatility with stops on every trade, watch the earnings and Fed calendar, and treat multiple AI positions as concentrated, not diversified.

Sources

Benzinga, Micron vs. Western Digital vs. SanDisk: which AI memory stock looks best after massive 2026 rallies — https://www.benzinga.com/trading-ideas/technicals/26/06/60095061/micron-vs-western-digital-vs-sandisk-which-ai-memory-stock-looks-best-after-massive-2026-rallies

24/7 Wall St., SanDisk rallies 8%, Western Digital rises 6%, Micron gains 5% as Elon Musk flags a memory bottleneck — https://247wallst.com/investing/2026/08/17/sandisk-rallies-8-western-digital-rises-6-micron-gains-5-as-elon-musk-flags-a-memory-bottleneck/

Via News, Micron and SanDisk rank among 2026's biggest chip winners on AI memory demand — then cool from June highs — https://via.news/markets/micron-and-sandisk-rank-among-2026-s-biggest-chip-winners-on-ai-memory-demand

CNBC, 4 memory stocks are darlings of the AI boom — https://www.cnbc.com/2026/08/21/4-memory-stocks-are-darlings-of-the-ai-boom-how-they-differ-and-why-we-own-micron.html

Crowdfund Insider, AI server demand surges, lifts Dell Technologies shares 350% YTD — https://www.crowdfundinsider.com/2026/09/309925-ai-server-demand-surges-lifts-dell-technologies-shares-350-ytd/

The Motley Fool, Stock market today, Sept. 2: Dell surges 16% on soaring AI backlog — https://www.fool.com/coverage/stock-market-today/2026/09/02/stock-market-today-sept-2-dell-surges-16-on-soaring-ai-backlog/

TradingKey, Dell stock forecast: $95B AI backlog meets rising channel support — https://www.tradingkey.com/analysis/stocks/us-stocks/262178315-dell-stock-forecast-ai-backlog-559-support-tradingkey

NerdWallet, 7 best-performing semiconductor stocks for September 2026 — https://www.nerdwallet.com/investing/learn/best-semiconductor-stocks

KCEX, U.S. semiconductor stocks surge: key drivers behind rally — https://blog.kcex.com/markets/cpu-ai-chip-stocks-rally-intel-amd-arm-qualcomm-nvidia-2026/

Yahoo Finance, Intel, AMD, Micron shares trim losses after Broadcom results spark semiconductor sector sell-off — https://finance.yahoo.com/markets/stocks/article/intel-amd-micron-shares-trim-losses-after-broadcom-results-spark-semiconductor-sector-sell-off-130128361.html

StockAnalysis, Super Micro Computer (SMCI) stock overview — https://stockanalysis.com/stocks/smci/

Semafor, AMD reaches a $1 trillion market cap as chip stocks drive rally — https://www.semafor.com/article/09/21/2026/amd-reaches-a-1-trillion-market-cap-as-chip-stocks-drive-rally