AMD and Arista Beat on AI Hardware — and Both Jumped: the Halal Investor's Read (2026)
Tuesday night broke this earnings season's most consistent pattern. For weeks, beating the quarter and disappointing the guide meant falling: Apple, Palantir, Alphabet, Amazon, Intel, and Reddit all beat and all fell. Advanced Micro Devices and Arista Networks did the opposite. Both beat the quarter, both guided the next quarter above expectations, and both jumped in after-hours trading — AMD about 5 percent on record revenue of 13.38 billion dollars and a fourth-quarter guide of about 17 billion dollars, Arista about 9.5 percent on revenue up 58.5 percent to 3.66 billion dollars, its fastest growth since its IPO. This article separates the two quarters from the two jumps, then runs both companies through the Shariah screen, because the screen is the one part of the morning that did not move.
Updated November 2026 · Reviewed by the Canadian Halal Investor team
What actually happened on Tuesday night
Two of the companies that build the physical backbone of artificial intelligence reported after the closing bell on Tuesday 3 November 2026, and both delivered the kind of quarter this earnings season has repeatedly punished with a sell-off — and both jumped anyway. Advanced Micro Devices, which makes the processors and AI accelerators that go inside data centres, reported record revenue and raised its outlook so sharply that management now expects 2027 revenue to grow by more than 80 percent. Arista Networks, which makes the high-speed switches that connect those accelerators to each other, reported its fastest revenue growth since going public. AMD shares traded about 5 percent higher after hours, to about 508.20 dollars against a 483.90 dollar close, and Arista shares traded about 9.5 percent higher after hours, to about 186 dollars. Wednesday morning's tape opened with those two moves as its centrepiece.
The backdrop made the jumps more striking, not less. Americans voted in midterm elections on Tuesday, and by Wednesday morning Democrats had secured the Senate and were on course to take the House as well, according to the morning briefings. Index futures first fell on the early results, then reversed: by about 5:20 a.m. Eastern, S&P 500 futures were up about 0.9 percent, Nasdaq-100 futures about 0.8 percent, and Dow futures about 1.3 percent, while gold futures rose about 3 percent and the dollar slipped about 0.5 percent. Asian markets had fallen overnight on the same political news, with chip-heavy markets hit hardest. In other words, the macro tape was confused and the overnight tape was negative — and two AI hardware reports cut straight through both. That is worth noting because it tells you what the market is actually trading right now: not the election, which it had hours to digest, but the demand outlook for AI infrastructure, which these two reports just repriced upward.
The rest of the overnight tape shows why these two reports stood out. Amdocs fell about 43 percent after forecasting a revenue decline of 14.3 to 17.9 percent in its new fiscal year and citing a dispute with T-Mobile. Skyworks Solutions fell about 17 percent on its December-quarter guidance. Duolingo rose about 12 percent after daily active users grew 29 percent. None of those moves is a sector call; each is a company-specific verdict, exactly like AMD's and Arista's. The pattern of this season — Apple and Palantir beat on Monday night and both fell, Alphabet fell about 10 percent on a cloud surge, Amazon fell about 7 percent on record margins — has trained investors to expect punishment even for good quarters. Tuesday night broke that pattern for the two companies whose results were about the future, not just the quarter, as the next two sections show.
Wednesday itself is one of the busiest reporting days of the season, with 689 companies scheduled according to one earnings calendar, including Uber, CVS Health, and McDonald's before the open and Qualcomm, Arm, and AppLovin after the close. Those reports will decide whether Tuesday night's AI hardware strength broadens or stays contained. But the morning run's job is the overnight story, and the overnight story is unambiguous: the two companies that sell the picks and the wiring of the AI build-out both beat, both raised the bar for the next quarter, and both were rewarded for it. The rest of this article separates what they actually reported from what the market paid for, then runs both through the Shariah screen.
AMD: the quarter, line by line
AMD reported third-quarter 2026 revenue of 13.377 billion dollars, up 45 percent from 9.25 billion dollars a year earlier and up 16 percent from 11.54 billion dollars in the second quarter — a new single-quarter record, above the top of the company's own guidance of about 13 billion dollars and above compiled estimates of about 13.09 billion dollars. Non-GAAP diluted earnings per share were 2.09 dollars, up 74 percent from a year earlier and above estimates of about 1.94 to 1.95 dollars. GAAP diluted earnings per share were 1.46 dollars, which included a 1.1 billion dollar non-cash charge from remeasuring the fair value of warrants — the kind of accounting line that depresses the GAAP figure without touching cash flow, and the reason the non-GAAP figure is the one analysts quoted. Non-GAAP gross margin was 56 percent, non-GAAP operating income was 3.92 billion dollars, up 75 percent, and GAAP net income was 2.43 billion dollars.
The engine of the quarter was the Data Center segment, and the scale of it deserves a slow read. Data centre revenue was a record 8.546 billion dollars, up 97 percent from 4.34 billion dollars a year earlier and up from 6.72 billion dollars in the second quarter. That segment now accounts for about 64 percent of AMD's total revenue, up from 47 percent a year earlier, and it earned 2.88 billion dollars of operating income, up from 1.07 billion dollars a year earlier, a margin of about 34 percent. The company credited record server processor revenue from its fifth-generation EPYC chips and the first shipments of its Instinct MI450-series AI accelerators, with its first Helios rack-scale systems now running at Anthropic, Meta, Microsoft, OpenAI, and Oracle Cloud Infrastructure. In plain terms, AMD is no longer mostly a PC and gaming chip company with a data centre division. It is a data centre company that also sells PC and gaming chips, and the quarter's mix proves it.
The guidance is what turned a strong quarter into a jumping stock. For the fourth quarter, AMD expects revenue of about 17 billion dollars, plus or minus 3 percent — about 66 percent growth year over year and about 27 percent sequentially, above compiled estimates of about 16.3 billion dollars — with a non-GAAP gross margin of about 55 percent. Inside that guide, data centre revenue is expected to rise about 50 percent sequentially, while client and gaming revenue falls more than 20 percent sequentially, which tells you exactly where management sees the growth coming from. Then came the longer outlook, delivered on the earnings call by chief executive Lisa Su: based on current visibility, AMD expects 2027 revenue to grow by more than 80 percent, more than double the growth rate in the long-term model presented at last November's analyst day, with server revenue roughly doubling and data centre AI revenue growing significantly more than twofold, so that data centre becomes more than 80 percent of total revenue. The chief financial officer added that data centre AI revenue should pass server processor revenue for the first time in the fourth quarter. A company guiding its next quarter 27 percent above its record quarter, and its next year more than 80 percent above this one, is not reporting a quarter. It is redrawing its own size.
Two cautions belong in the AMD section before the screen section, because both will matter for the halal read and for the investment read. First, the 2027 outlook is a call, not a contract: it rests on current visibility, on Helios systems ramping on schedule, and on supply — Su addressed supply constraints directly on the call, and a ramp this steep has many places to slip. Second, the GAAP-to-non-GAAP gap is unusually wide this quarter because of the 1.1 billion dollar warrant remeasurement. That charge is non-cash and does not change the business, but it is a reminder to read both earnings figures rather than whichever one flatters the headline. Neither caution changes the quarter's facts. Both change how much weight a careful reader puts on the forecast — which is the right instinct, because the stock's jump is a payment for the forecast, not for the quarter.
Arista: the quarter, line by line
Arista Networks reported third-quarter 2026 revenue of 3.6576 billion dollars for the quarter ended 30 September 2026, up 58.5 percent from 2.3083 billion dollars a year earlier and up 20.5 percent from the second quarter — the fastest year-over-year growth since the company's IPO, and more than 350 million dollars above the roughly 3.3 billion dollars Arista itself had guided in August. GAAP diluted earnings per share rose to 1.22 dollars from 0.67 dollars a year earlier, and non-GAAP diluted earnings per share were 1.20 dollars against 0.77 dollars a year earlier, above the company's guided range of 1.06 to 1.08 dollars. GAAP operating margin was 44.9 percent and non-GAAP operating margin 49.1 percent, against 42.4 percent and 48.6 percent respectively a year earlier. Arista is a networking company, and networking companies do not normally grow at 58.5 percent. The reason this one did is the same reason AMD did: its switches are the connective tissue of AI data centres, and AI data centres are being built at a pace that has overwhelmed normal growth arithmetic.
The fourth-quarter guide kept the shape intact rather than promising a new acceleration. Arista expects fourth-quarter revenue of 3.9 to 4.0 billion dollars, a non-GAAP operating margin of 48 to 49 percent, and non-GAAP diluted earnings per share of 1.26 to 1.30 dollars. That is sequential revenue growth of about 6.6 to 9.4 percent on a base that just grew 20.5 percent, with margin essentially held — a guide that reads as conservative after a beat of more than 10 percent against the company's own forecast, which is usually how Arista guides. The company had already raised its full-year 2026 revenue outlook to about 12.6 billion dollars after the second quarter, implying about 40 percent annual growth, and it counts more than 100 cumulative Etherlink customers for its AI networking fabrics, with a target of at least 3.5 billion dollars of AI fabrics revenue in 2026. The customer concentration note from before the report still applies: Microsoft and Meta are expected to anchor one or two customers accounting for at least 10 percent of revenue each, and management has warned that industry component shortages could persist until 2028.
Read together, the AMD and Arista quarters describe one build-out from two ends. AMD sells the accelerators and processors that do the computing. Arista sells the switches that move data between them at speeds the older generation of networks cannot handle. When both beat by wide margins in the same night, the market is not getting two independent signals. It is getting the same signal twice: the AI data centre build-out is running ahead of what even the companies building it forecast a quarter ago. That is also why the two stocks moved together after hours while other technology names fell on their own news. Amdocs and Skyworks were judged on their own guidance. AMD and Arista were judged, in part, on each other's — and both confirmed the same story.
Why both jumped when other beats fell
This season has taught one lesson so consistently that it looked like a law: beat the quarter, disappoint the guide, fall. Apple beat and fell on a December-quarter margin guide of 45 to 46 percent against 46.3 percent expected. Palantir beat and fell because total contract value grew only 6 percent after 49 percent in the prior quarter. Alphabet fell about 10 percent while Google Cloud grew 103 percent, because capital spending hit 51.4 billion dollars and free cash flow went negative. Amazon fell about 7 percent on record margins because quarterly capital spending reached 77.3 billion dollars. Intel plunged about 25 percent despite beating, because its fourth-quarter guide of 15 to 16 billion dollars trailed the roughly 17.07 billion dollars expected. Reddit crashed 29 percent on a beat, because guidance and user numbers disappointed. Shopify fell about 19 percent on its first revenue miss since 2022. In every case the market traded the future, not the quarter.
Tuesday night is the exception that proves the mechanism rather than breaking it. AMD and Arista did not just beat the quarter. They beat the future that the market had priced. AMD's fourth-quarter guide of about 17 billion dollars sits above the roughly 16.3 billion dollars analysts expected, and its 2027 outlook of more than 80 percent growth replaces a long-term model that had assumed less than half that pace. Arista's third-quarter revenue beat its own guide by more than 10 percent, and its fourth-quarter guide holds margins near 49 percent while growing revenue again. The market's rule never changed: it trades the guide. On Tuesday night, for the first time in weeks, the guide was better than expected for both companies at once, and the stocks did exactly what the rule predicts. The lesson is not that the market has become forgiving. It is that the market was never punishing beats. It was punishing guides, and these two guides were genuinely good.
There is a second, structural reason the two jumps reinforce each other. AMD's accelerators are useless without Arista's switches, and Arista's switches are bought because customers are installing AMD's and Nvidia's accelerators. A halal investor watching only one of the two names would see half a story. Watching both, the story is a single capital-spending cycle with two suppliers reporting on the same night, both confirming that the cycle is accelerating rather than cooling. That is also the risk, stated plainly: a single cycle means correlated exposure. If the data centre build-out slows — because of power constraints, component shortages, or customers pausing to digest what they have bought — both names face the same headwind at the same time. Diversification across two tickers in the same cycle is not diversification across two different businesses. The screen section that follows treats them as two separate passes, because the screen judges each company on its own facts. The portfolio section treats them as one theme, because the market will price them that way.
What the Shariah screen says: two passes, two caveats
Both companies still pass this site's Shariah screen, as of the October 2026 screening, and nothing in Tuesday night's reports changes either result. But the two passes are not equally comfortable, and a halal investor should know exactly where each one is tight. The screen works in two gates. The first gate is business activity: is the company's core business permissible? AMD makes semiconductors and Arista makes networking hardware and software. Neither lends at interest, sells alcohol or pork, runs gambling, produces tobacco, or manufactures weapons, and neither appears on the BDS movement's published boycott, divestment, or pressure lists as of the site's October 2026 check — reported as separate informational context, not as a screen change. Both clear the business gate cleanly, and an earnings report does not touch that gate at all: both companies sold the same products on Wednesday that they sold on Tuesday.
The second gate is financial ratios, and here the two companies diverge. Arista's balance sheet carries no debt — 0.00 percent of its roughly 258.5 billion dollar market capitalisation at screening — against a ceiling of roughly 33 percent, and some providers use a 30 percent ceiling, which Arista clears either way. Its interest income of 383.4 million dollars against full-year 2025 revenue of 9,005.7 million dollars is about 4.26 percent, under the 5 percent non-compliant income ceiling but elevated — the highest ratio among the recent United States expansion passes on this site, and the number to watch. Zoya's screener independently rates Arista compliant, which is a useful cross-check, but the cross-check does not move the ratio: at 4.26 percent, Arista has less than one percentage point of headroom before the ceiling, and a year in which interest income grows faster than revenue could close that gap. AMD's ratios look more comfortable on the balance sheet — debt of 0.32 percent of market capitalisation and cash of 1.31 percent — but its pass carries an asterisk the site does not hide: AMD does not separately disclose interest income, bundling it inside other income, net, so the 5 percent income screen cannot be fully verified from the filing, and the pass stays provisional rather than adopting another vendor's estimate.
Two further points keep the screen honest. First, the screening figures are October 2026 figures. Tuesday night's earnings reports do not restate them: market capitalisations moved, which moves the denominators, but the screen is refreshed on its own cycle, not on every price move. The correct statement is that both passed as of October 2026 and that nothing in the earnings reports changes either business gate or suggests a ratio breach; the next full refresh, not Wednesday's price, is what could change a result. Second, the provider disagreement about ceilings — roughly 30 percent versus roughly 33 percent for debt — does not affect either company, since both sit at or near zero debt, and both clear the 5 percent income ceiling on the figures available. The disagreements that matter here are narrower: for Arista, how much headroom 4.26 percent really leaves, and for AMD, whether an undisclosed income figure should keep a pass provisional. The site reports both openly and does not pick a position beyond its own methodology, and readers who follow a scholar or provider with a stricter view should apply that view.
One practical screen fact applies to both names: neither company currently pays a dividend, so there is currently nothing to purify — purification applies to dividends received, and a company that pays none generates no purification obligation. If either company ever starts paying a dividend, the purification ratio would come from the screening refresh in force at the time, and the site's purification calculator runs that arithmetic holding by holding. Capital gains are a separate question on which scholars differ: the widely practised retail approach purifies dividends, while some scholars describe purifying capital gains as more precautious. The site reports that disagreement and does not pick a position. For the full company-by-company detail, see the AMD screener page and the Arista screener page, and for the method behind every number, see the screening methodology.
How a halal investor should approach a morning like this
An after-hours jump triggers the same three instincts as a premarket fall, in reverse: buy immediately before the move gets away, add to an existing holding because the thesis feels confirmed, or do nothing and regret it by noon. None of the three is a framework. The framework this site's methodology implies is a set of questions, answered in order, before any order is placed — and on a morning like this, the order of the questions matters more than usual, because the screen gives a green light that is easy to misread as a recommendation.
First, run the screen before you run the trade, and read the caveats, not just the verdict. Both companies pass, but Arista's pass has 0.74 percentage points of income headroom and AMD's pass is provisional on an undisclosed figure. A pass with a caveat is still a pass — the screen is a filter, not a recommendation — but the caveat tells you what to monitor, and monitoring is the whole job after the purchase. Second, separate the price from the business, in writing. Write down what changed about the business: AMD's data centre revenue nearly doubled, its 2027 outlook more than doubled its old model, Arista's revenue grew at its fastest pace since its IPO, and both guided the next quarter above expectations. Then write down what changed about the price: AMD up about 5 percent after hours, Arista up about 9.5 percent. If the business paragraph still describes companies you would own at the new prices, the jump is a repricing you can live with. If it describes companies whose prices now assume the 2027 outlook arrives on schedule, the honest question is whether you are paying for a forecast rather than a business.
Third, check the guide, not just the quarter — and check it against what you already believed. AMD's 2027 outlook of more than 80 percent growth is the single most important sentence in either report, and it is also the least verifiable: it rests on current visibility and on a Helios ramp that has barely begun. If your thesis already assumed AI infrastructure growth continuing at this pace, the outlook confirms it. If your thesis assumed a normalisation, the outlook challenges it, and the honest response is to revise the thesis, not to chase the stock. Fourth, treat the two names as one theme for sizing purposes. AMD and Arista are different companies, different tickers, and different screen pages, but they share one customer cycle, and a pause in that cycle would hit both. A position sized as if they were independent bets is a position sized wrong. Fifth, recheck the tight ratio at the next refresh: Arista's 4.26 percent interest income first, AMD's disclosure second, because those are the slow-moving numbers that a daily price watch misses and a quarterly refresh catches.
Finally, remember what this article is not. It is not a fatwa, and it is not a buy or sell call on AMD or Arista. The screen results are facts about October 2026 figures. The earnings numbers are facts about one quarter and two guides. Whether either company belongs in your portfolio, at what size, and in which account, depends on your circumstances, your scholars' guidance, and advice suited to your situation. The framework above is designed to make that decision slower, written down, and checkable — the opposite of an after-hours impulse. A halal portfolio is built on gates and discipline, not on reacting to the loudest morning of the quarter, even when the morning is genuinely good news.
What to watch next
Wednesday's own tape will test whether Tuesday night's strength broadens. Uber, CVS Health, and McDonald's report before the open, and Qualcomm, Arm, and AppLovin report after the close — the last two are the closest read-across to the AMD and Arista story, because Qualcomm and Arm both touch the same AI and mobile silicon demand, and their guides will either confirm the hardware cycle or complicate it. Beyond Wednesday, the jobs report due Friday 6 November is the week's macro anchor, and the midterm results will keep working through policy expectations for the rest of the quarter. None of those events changes either company's screen. All of them change the market the two stocks trade in, which is why the discipline is to follow the companies for the business, the screen for compliance, and the rest of the tape as background unless it changes one of those two things.
For the two companies themselves, watch three things. First, delivery: AMD's fourth-quarter guide of about 17 billion dollars requires data centre revenue to rise about 50 percent sequentially, and its 2027 outlook requires the Helios ramp to hold — any slip there reprices the forecast that Wednesday's jump just paid for. Second, Arista's fourth-quarter delivery against its 3.9 to 4.0 billion dollar guide at a 48 to 49 percent non-GAAP operating margin, and whether component shortages, which management has warned could persist until 2028, start to bite. Third, the next screening refresh, which should recheck Arista's interest income ratio first — 4.26 percent against a 5 percent ceiling, with a growing cash pile in a high-rate environment — and AMD's income disclosure second. For readers who hold either name, the practical next step is unglamorous and that is the point: re-read your own thesis, check it against the numbers in this article, and confirm the position size still matches the volatility you just watched. Run it through the screening methodology, check the companies on the AMD screener page and the Arista screener page, and build the rest of the decision from there. Earnings season rewards the investor who reads the guide behind the headline. Tuesday night was the latest proof.
FAQ
What did AMD actually report on Tuesday 3 November 2026?
AMD reported third-quarter 2026 revenue of 13.377 billion dollars, up 45 percent from 9.25 billion dollars a year earlier and above compiled estimates of about 13.09 billion dollars. Non-GAAP diluted earnings per share were 2.09 dollars, up 74 percent and above estimates of about 1.94 to 1.95 dollars, while GAAP diluted earnings per share were 1.46 dollars including a 1.1 billion dollar non-cash warrant remeasurement charge. Data centre revenue was a record 8.546 billion dollars, up 97 percent, now about 64 percent of total revenue. For the fourth quarter AMD guided revenue to about 17 billion dollars, plus or minus 3 percent, above estimates of about 16.3 billion dollars, and on the call management said 2027 revenue should grow by more than 80 percent.
What did Arista Networks actually report?
Arista reported third-quarter 2026 revenue of 3.6576 billion dollars for the quarter ended 30 September 2026, up 58.5 percent from 2.3083 billion dollars a year earlier and up 20.5 percent from the second quarter, its fastest growth since its IPO and more than 350 million dollars above its own guide of about 3.3 billion dollars. GAAP diluted earnings per share rose to 1.22 dollars from 0.67 dollars a year earlier, and non-GAAP diluted earnings per share were 1.20 dollars against 0.77 dollars a year earlier. GAAP operating margin was 44.9 percent and non-GAAP operating margin 49.1 percent. For the fourth quarter Arista guided revenue to 3.9 to 4.0 billion dollars, a non-GAAP operating margin of 48 to 49 percent, and non-GAAP diluted earnings per share of 1.26 to 1.30 dollars.
Why did both stocks jump when other recent beats fell?
Because the market trades the guide, not the quarter, and these two guides were better than expected. Recent beats fell because their guides disappointed: Apple on margin, Palantir on contract growth, Alphabet and Amazon on capital spending, Intel on a weak fourth-quarter guide, and Reddit on guidance and user numbers. AMD beat the quarter and guided the fourth quarter to about 17 billion dollars against about 16.3 billion dollars expected, plus a 2027 outlook of more than 80 percent growth. Arista beat its own guide by more than 10 percent and guided the fourth quarter to 3.9 to 4.0 billion dollars at near-49 percent non-GAAP operating margin. The rule never changed. These two guides were genuinely good.
Do AMD and Arista still pass the Shariah screen after these reports?
Yes, as of this site's October 2026 screening, which these earnings reports do not change. Both clear the business-activity gate: AMD makes semiconductors and Arista makes networking hardware and software, neither with a prohibited core business, and neither appears on the BDS movement's published boycott, divestment, or pressure lists as of the October 2026 check. AMD passes provisionally with debt of 0.32 percent of market capitalisation and cash of 1.31 percent. Arista passes with zero debt and interest income of 4.26 percent against a 5 percent ceiling. The next full screening refresh, not Wednesday's price moves, is what could change either result.
What is the number to watch on each company's screen?
For Arista, the interest income ratio of 4.26 percent against the 5 percent non-compliant income ceiling, leaving less than one percentage point of headroom, which is the highest ratio among the recent United States expansion passes on this site. For AMD, the missing figure itself: AMD does not separately disclose interest income, bundling it inside other income, net, so the 5 percent income screen cannot be fully verified from the filing and the pass stays provisional. Debt is not the tight number for either company, since Arista carries zero debt and AMD carries debt of only 0.32 percent of market capitalisation.
Is anything in this article a fatwa or a buy or sell call?
No. Everything here is facts and screening methodology only, with no fatwas and no personalised buy or sell calls. Rulings on your own situation belong with a qualified scholar, and investment decisions belong with advice suited to your circumstances. Not financial advice. Facts and screening methodology only, no fatwas.