William Blair upgrades AMD on data-center growth and valuation | Magica
William Blair upgrades AMD as data-center growth reshapes its stock case
Editorial Team
••📖7 min read
William Blair raised AMD to Outperform as server-chip prospects and valuation improved. AMD already expects revenue growth above 80% in 2027, but customer incentives, custom accelerators and weaker quarterly cash generation complicate the investment case.
William Blair upgraded AMD to Outperform, citing improved server-chip prospects and valuation.
AMD itself already expects revenue growth above 80% in 2027; that outlook predates the upgrade.
Custom AI chips threaten AMD’s accelerators while opening a separate market for its server CPUs.
William Blair, an employee-owned investment firm, upgraded Advanced Micro Devices to Outperform from Market Perform on November 12, judging that the semiconductor designer’s competitive position and valuation had improved since July. The firm expects more than 80% revenue growth in 2027. AMD had already set out that expectation in its November 3 earnings remarks.
The change is a more favorable stock recommendation for a company whose data-center business is already driving nearly all its revenue expansion. The harder question is how much of that growth will translate into durable cash generation as customers develop competing chips and AMD offers financial incentives alongside deployment agreements.
Company-reported AMD segment revenue for Q3 2025 and Q3 2026. Chart: Magica. Source: AMD.
What changed in William Blair’s stock case
Sebastien Naji, the William Blair analyst who initiated AMD coverage in July, said the four factors behind his earlier cautious rating had shifted in AMD’s favor, according to the account of his upgrade.
AMD designs central processing units, or CPUs, and graphics processing units, or GPUs, for data centers and other markets. Its EPYC server CPUs compete with Intel’s Xeon processors, while its Instinct AI accelerators offer an alternative to Nvidia’s products. This gives AMD two distinct roles in AI systems: supplying host processors and supplying accelerators that perform AI computation.
Naji cited continuing delays and supply shortages in Intel’s Xeon roadmap, a less decisive shift toward Arm-based host processors than expected, and a decline in AMD’s valuation multiple as earnings estimates rose. William Blair also cited scrutiny and potential leadership disruption at Nvidia, AMD’s larger AI-chip rival. These are the brokerage’s reasons for expecting a better competitive opening, rather than evidence of orders transferring to AMD.
Google launched Gemini 4 Hydrogen and automatic model switching in Antigravity, claiming near-Carbon task success with less quota and faster completion. Lower token rates support the efficiency pitch, but reasoning settings, mixed competitive results and inconsistent benchmark labels leave the cost of completed work to be tested.
Editorial Team
AMD’s multiple had returned to approximately its three-year forward median, according to the account of the valuation argument. The change concerns the price investors pay relative to forecast earnings. The same report said the InvestingPro stock-analysis service’s fair-value model considered AMD overvalued; it did not provide enough methodology to reconcile that assessment with William Blair’s multiple-based case. The stock had already gained more than 147% in 2026 at the time of a premarket report of the upgrade, which cited TheFly for the rating change and said no price target was assigned.
William Blair was already positive about AMD’s business before upgrading the shares. Its July initiation forecast sales rising from $52 billion in 2026 to more than $104 billion in 2028, while acknowledging that AMD’s GPU portfolio had lagged its larger rival. Outperform means it expects the stock to beat the broader market over the next 12 months.
The investment firm also has commercial interests relevant to its research. In the July disclosure, William Blair said it or an affiliate was a market maker in AMD shares and expected to receive or intended to seek investment-banking compensation from AMD or an affiliate within three months. That dated disclosure does not establish whether a banking mandate followed, but it is material context for its recommendation.
The growth outlook was already AMD’s
William Blair expects data-center AI revenue to overtake AMD’s server CPU business in the fourth quarter, with Helios systems expanding through 2027. Helios combines Instinct MI450 GPUs and sixth-generation EPYC CPUs in rack-scale systems.
The more-than-80% growth figure is not a new forecast unique to the brokerage. In its November 3 prepared remarks, AMD said it expected company revenue growth above 80% in 2027, more than twice the long-term financial model it had presented the previous November. AMD also forecast a doubling of server revenue and a data-center share above 80% of total sales next year. All are company expectations, not achieved results.
AMD’s third-quarter results show how concentrated the expansion has become. Company revenue rose 45% from a year earlier, while data-center revenue grew 97%. Data centers supplied roughly 64% of sales, compared with about 47% a year earlier, calculated from the reported figures.
Revenue, $ millions
Q3 2025
Q3 2026
Dollar change
Data center
4,341
8,546
+4,205
Other businesses combined
4,905
4,831
−74
Company total
9,246
13,377
+4,131
The changes and combined other-business totals are calculated from AMD’s segment figures for the same quarters. Data centers supplied more than the company’s entire net revenue increase, offsetting the decline elsewhere. Gaming revenue fell 40% to $773 million, while client revenue grew 7% and embedded revenue grew 29%.
Sales growth has not produced equally strong quarterly cash generation. AMD’s non-GAAP free cash flow—operating cash flow from continuing operations less capital expenditure—fell to $554 million from $1.530 billion a year earlier. Its cash-flow statement shows a $2.995 billion quarterly inventory outflow and $648 million of capital expenditure. For the first nine months of 2026, free cash flow rose to $4.678 billion from $3.437 billion in the same period of 2025.
Company-reported announced capacity for OpenAI’s separate AMD and Broadcom programs; these figures do not measure completed deployments, chip performance or spending. Chart: Magica. Source: OpenAI.
OpenAI is both a customer and a source of competition
William Blair flags custom application-specific integrated circuits, or ASICs, particularly at Meta and at OpenAI, the generative AI developer behind ChatGPT, as a risk to Instinct. It also sees these projects creating demand for EPYC host CPUs. AMD’s earnings remarks give a concrete example: the company said OpenAI selected its Turin-generation EPYC processors for the host CPUs in its first custom accelerator platforms.
OpenAI was already collaborating with AMD on the MI300X and MI350X accelerator generations before their October 2025 strategic agreement. That definitive agreement covers six gigawatts of AMD GPU deployments across multiple generations, beginning with an initial gigawatt of MI450 GPUs scheduled to start in the second half of 2026. AMD said the partnership was expected to generate tens of billions of dollars in revenue; the announcement did not give a fixed revenue schedule.
AMD also issued OpenAI a warrant for up to 160 million AMD shares. Vesting depends on deployment milestones, AMD share-price targets, and technical and commercial conditions. The first tranche is tied to the initial gigawatt deployment, with further tranches as purchases scale toward six gigawatts. The customer thus has a potential equity reward linked to AMD’s deployment and stock performance.
OpenAI simultaneously has another hardware path. Its October 2025 collaboration with Broadcom, a supplier of custom AI chips and data-center networking, envisages 10 gigawatts of OpenAI-designed accelerators. The announcement described existing co-development and supply agreements plus a term sheet for deploying racks, with deployment targeted from the second half of 2026 through the end of 2029.
OpenAI said designing its own hardware would incorporate lessons from its models and products; Broadcom would help develop and deploy the systems and supply connectivity. The two gigawatt totals describe different announced programs, not completed capacity, comparable chip performance or disclosed purchase prices. Neither announcement supplies enough pricing detail to establish which route will cost OpenAI less.
Company-reported non-GAAP free cash flow, comparing equal-duration periods across years. Nine-month figures are cumulative. Chart: Magica. Source: AMD.
The next test is delivery, margins and cash
AMD says initial Helios systems are already running models and workloads at Anthropic, Meta, Microsoft, OpenAI and Oracle Cloud Infrastructure. That establishes an initial customer footprint on AMD’s account; it does not establish completion of the multigigawatt agreements.
AMD expects fourth-quarter revenue of about $17 billion, plus or minus 3%, representing roughly 66% year-over-year growth at the midpoint. It forecasts a non-GAAP gross margin of about 55%, compared with 56% in the third quarter. Those figures make both sales conversion and profitability near-term tests of the stock case.
For 2027, AMD says it has secured wafers, substrates, advanced packaging, memory and other components to support higher growth. Its earnings release nevertheless identifies third-party manufacturing capacity and component availability as risks. Subsequent results will need to show that supply commitments become delivered systems and that rising revenue produces sustained cash flow. The custom-chip programs will also test whether EPYC gains can offset any Instinct business lost to customer-designed alternatives.
Japan is considering safety and reporting duties backed by enforceable orders for advanced AI. The proposal would strengthen its nonbinding framework, but capability thresholds, covered businesses and penalties remain unresolved.
Google says its new fast AI model is reaching Search users worldwide, improving complex answers and source discovery. The rollout expands access to Gemini 4, but its effect on answer quality and website traffic remains unmeasured.
Four nonprofits are asking a California court to stop five AI developers from building or deploying more capable systems. The case seeks enforceable limits even as OpenAI has previously reported a training pause and is launching defensive AI tools.
Expo Simulators gives coding agents remote iOS and Android environments with recordings for reviewers. The launch extends Expo's cloud development services into verification, with per-minute charges and shared capacity limits.
Kleiner Perkins and Clearlake Capital led Mirendil’s $1 billion Series A. Its agent-led model-building experiment points to smaller research teams, but company benchmarks, substantial compute costs and a private beta qualify the promise of wider access.
Mount Thor raised $45 million in equity, alongside separate hardware financing, to expand Apple infrastructure for AI agents and model inference. Existing Mac clouds and workload-specific performance limits put its demand and cost claims to the test.
Salesforce Ventures led a Series C that values Q-CTRL above $1 billion. The funding backs GPS-free navigation and quantum computing software, while undisclosed revenue and contested performance benchmarks leave the commercial case open.
Adobe demonstrated 10 experimental creative tools and opened Deep Brush and Hands On for testing. The lineup explores more direct control over AI edits, but existing alternatives and unproven production performance limit claims of novelty.
House Foreign Affairs Committee Chairman Brian Mast says outgoing Nvidia chief Jensen Huang has agreed to questioning, with a date still under negotiation. Speaker Mike Johnson says support for limiting chip exports to China is growing as lawmakers examine allegations involving Huang and a cloud partner.
Leaked correspondence and a reconstructed publishing program expose John Mark Dougan’s use of AI to stock imitation news sites with pro-Russian material. The records illuminate funding and working methods, while his claims of influencing other AI systems remain unproven.
Apple’s $299 first smart display earns praise for fast home controls and useful kitchen features, but reviewers found gaps in how it keeps family members’ messages separate. Its strongest advantages also depend on owning compatible Apple devices and accessories.
Donald Trump proposed Apple executive chairman Tim Cook to succeed Jensen Huang at Nvidia. Cook’s interest and the seriousness of the suggestion remain unclear, while finance chief Colette Kress has taken interim charge.
Norwest’s largest fund is a third bigger than its predecessor, giving the Wells Fargo-backed investor more capital for AI alongside healthcare, consumer and industrial businesses. The firm has not disclosed how much will go to AI.
Nebius raised its 2026 capital spending forecast to about $27 billion while steering revenue to the low end of guidance. Better customer prepayments support expansion, but delayed capacity and a $325 million non-recurring contribution complicate its quarterly revenue beat.
Megaspeed founder Alice Huang is reportedly living in Beijing, but her activities and the espionage allegations remain unresolved. Jensen Huang’s resignation changes Nvidia’s leadership without settling questions about her business connections and access to its chips.
Singapore says it barred former Megaspeed chief Alice Huang after she gave inaccurate information about her role, while finding insufficient evidence against her over server declarations. The wider investigation still depends on tracing shipments beyond Malaysia.
Nvidia’s longtime finance chief takes over from its co-founder as the board promises continuity, searches for a permanent CEO and faces unresolved questions over security and executive conduct.
Chicago Public Schools is seeking a way to preserve Google Search without its generative AI features as parents, educators and students press for stronger district rules. Google’s planned access changes turn an AI safety dispute into a test of who controls classroom technology.
Warby Parker scheduled November 12 preorders and holiday deliveries for its first Android XR glasses. Google’s backing combines development funding with a conditional equity investment, while the retailer expects only a modest initial sales contribution.
