By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
Meta's projected 2026 capex of $115-135B fuels AI data-center demand, benefiting AMD via strong revenue growth and partnerships like Helios racks. However, the article underplays competitive threats, timeline risks, and execution challenges.
Risk: Capex concentration and Meta's pivot to custom MTIA silicon
Opportunity: Strong revenue growth via Meta's capex and hyperscale partnerships
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
As Meta Doubles Down on Data Center Investment, AMD and American Tower Could Be the Top Stocks to Buy
Meta Platforms (META) CEO Mark Zuckerberg has been very clear about where the company is heading. In January, he unveiled "Meta Compute", a broad plan to build Meta’s future around large-scale AI infrastructure, purpose-built data centers, and dedicated power capacity.
The company put serious numbers behind that plan, guiding for $115 billion to $135 billion in capital expenditure in 2026, nearly double the $72.2 billion it spent in 2025, with most of that spending aimed at building data centers and buying AI chips. Meta’s data center footprint already stretches across U.S. markets, including El Paso, Texas, Kansas City, and Montgomery, with El Paso standing out as one of its biggest and fastest-growing projects. On March 26, Meta said it was increasing its investment in the El Paso site to more than $10 billion, with a goal of reaching one gigawatt of capacity by 2028.
More News from Barchart
This move is part of a much bigger trend. The largest tech companies are spending heavily to build the infrastructure needed to support the next phase of AI, from chips and land to the physical networks that tie everything together. Two stocks, in particular, are drawing attention in the data center investment theme: Advanced Micro Devices (AMD) and American Tower Corporation (AMT).
With Meta’s spending now clearly picking up, are these two among the best-positioned stocks to benefit? Let’s find out.
Advanced Micro Devices
Santa Clara-based chipmaker Advanced Micro Devices designs and sells CPUs and GPUs which are used in some of the most demanding parts of the computing market, from gaming systems to large AI data centers
Over the past 52 weeks, the stock is up 88.67%, showing the degree of interest from investors around the growth of data center demand. So far this year, however, shares are down 6.04%, which may give investors a better entry point after last year’s strong run.
On valuation, AMD trades at a forward P/E of 39.04 times, compared with the sector median of 21.12 times. That is a clear premium, but it also shows that the market is still willing to pay more for AMD because of its stronger growth profile.
The company’s recent results help explain that view. AMD posted record Q4 2025 revenue of $10.3 billion and finished full-year 2025 with a record $34.6 billion in revenue. Non-GAAP operating income for the year came in at $7.8 billion, while diluted EPS reached $4.17. For Q1 2026, AMD expects revenue of about $9.8 billion, which points to roughly 32% year-over-year (YOY) growth.
AMD is also doing more to strengthen its position in AI infrastructure. The company recently announced a partnership with Celestica (CLS) to bring the "Helios" rack-scale AI platform to market, combining AMD’s next-generation Instinct MI450 Series GPUs with Celestica’s advanced networking switch technology, with availability expected in late 2026. Further, AMD signed a multi-year partnership with Nutanix (NTNX) in February 2026 and committed up to $250 million to support agentic AI workloads on AMD-powered infrastructure. Together, those moves show AMD is working to expand beyond chips and offer broader AI systems, the kind of setup large customers like Meta need to support 1GW-scale data centers.
Wall Street remains positive on the stock. Out of 45 analysts surveyed, 30 rate it a “Strong Buy,” two rate it a “Moderate Buy”, and the remaining 13 are a “Hold.” The average price target of $286.37 implies about 41.8% upside from the current share price, and the Street-high price of $380 looks to an 88% climb from here.
Four leading AI models discuss this article
"Meta's capex surge is bullish for semiconductor infrastructure broadly, but AMD's valuation leaves no room for execution risk, and the article provides zero evidence AMD will capture meaningful share of Meta's GPU spending versus NVIDIA's entrenched position."
Meta's capex guidance ($115–135B in 2026) is real and material, but the article conflates infrastructure spending with AMD chip demand without proving the connection. AMD trades at 39x forward P/E—nearly 2x the semiconductor median—pricing in near-perfect execution. The Helios partnership and Nutanix commitment are announcements, not revenue. Critically: Meta's capex doesn't guarantee AMD wins the GPU contract (NVIDIA remains dominant; Intel and custom silicon are real threats). The article also ignores that Meta's spending may be front-loaded hype; capex guidance often slips. AMT gets one paragraph despite being the headline pick—insufficient analysis.
AMD's 32% Q1 2026 YOY growth guidance suggests deceleration from 2025 momentum, and at 39x forward P/E, the stock has already priced in years of flawless execution. If Meta diversifies suppliers or delays capex, AMD has no margin for error.
"AMD's transition from a component supplier to a full-stack AI infrastructure partner makes it the primary beneficiary of Meta’s massive infrastructure scaling."
Meta's projected $115B-$135B capex for 2026 is a massive tailwind for AMD, particularly as Meta seeks to diversify away from Nvidia's supply chain dominance. AMD's shift toward 'rack-scale' solutions via the Celestica partnership is a critical strategic pivot; they are no longer just selling chips but providing the integrated architecture required for 1GW-scale facilities. With a 32% YoY revenue growth forecast for Q1 2026, AMD’s 39x forward P/E (Price-to-Earnings) is justifiable if they capture even 15-20% of Meta’s GPU spend. However, the article ignores the 'AI overhang'—if Meta’s massive spend doesn't yield immediate ROI in ad revenue, that capex could be slashed abruptly.
The 'Meta Compute' plan relies on proprietary silicon; if Meta’s internal MTIA chips succeed, they may significantly reduce their reliance on external vendors like AMD for core AI workloads. Furthermore, the 2026 delivery timeline for the Helios platform leaves a massive window for Nvidia to further entrench its software moat with CUDA.
"AMD is well positioned to capture a meaningful slice of Meta’s AI-driven data‑center buildout, but that upside depends on continued share gains versus NVIDIA and flawless execution as it expands beyond chips into rack/system solutions."
Meta’s massive 2026 capex target ($115–$135B) crystalizes a multi-year wave of AI data‑center demand that flows to chips, systems integrators, networking and real‑estate owners. AMD is plausibly a direct beneficiary: strong FY25 revenue ($34.6B), 32% Q1 guide, and partnerships (Celestica Helios racks, Nutanix) push it from a silicon vendor toward rack/system supplier — the kind of integration hyperscalers prize. But the market already prices growth (forward P/E ~39x vs sector 21x), and the article underplays competitive threats (NVIDIA, in‑house silicon like Meta/Google TPUs), timeline risk for 1GW sites, and execution required to move up the stack. American Tower gets a short shrift too: tower cash flows benefit from edge/5G but aren’t a one‑to‑one play on hyperscale campus builds.
If Nvidia maintains technical leadership and price/perf advantages, or if Meta accelerates custom silicon and internal systems, AMD’s TAM for datacenter GPUs/racks could be materially smaller than expected; AMD’s premium valuation already embeds near‑perfect execution. Similarly, American Tower may see slower incremental revenue from hyperscale growth if carriers deploy more private on‑campus fiber and edge micro‑data centers instead of traditional tower leases.
"AMD's 39x forward P/E leaves no room for error against Nvidia's entrenched AI lead and hyperscaler spend uncertainties."
Meta's capex doubling to $115-135B in 2026 fuels AI infrastructure buildout, with AMD well-positioned via Instinct GPUs and partnerships like Helios racks (late 2026 availability) for hyperscale needs like El Paso's 1GW site. Record $34.6B 2025 revenue and 32% YoY Q1 2026 growth support the bull case, but 39x forward P/E (vs. sector 21x) embeds aggressive growth assumptions amid Nvidia's dominance (80%+ AI GPU share). YTD -6% decline hints at momentum fade; Wall Street's $286 PT implies 42% upside but ignores execution risks on unproven platforms and potential capex peak if AI ROI disappoints.
If AMD's MI450 GPUs win significant Meta designs and Helios scales rapidly, revenue could surge 50%+ by 2027, justifying the premium and driving re-rating toward 50x.
"Meta's proprietary MTIA chips pose a structural headwind to AMD's TAM that the 39x multiple doesn't adequately reflect."
Gemini flags the MTIA risk—Meta's internal silicon—but undersells it. If Meta's 2026 capex delivers 5-10% ROI uplift via MTIA, they'll likely reduce external GPU procurement by 20-30%, not add to it. AMD's 32% growth guidance assumes Meta scales external chips; if internal silicon accelerates, that guidance becomes a ceiling, not a floor. Nobody's quantified this cannibalization risk.
"AMD's valuation premium is unsustainable if its growth profile shifts into a high-concentration dependency on Meta's volatile capex cycles."
Grok and ChatGPT are too focused on the 39x forward P/E without addressing the 'Capex Concentration' risk. If Meta represents a dominant share of AMD's 2026 growth, AMD effectively becomes a high-beta proxy for Meta's balance sheet rather than a diversified semiconductor play. Investors are paying a premium for a firm whose revenue quality is degrading into a single-customer dependency. If Meta pivots to custom MTIA silicon as Gemini suggests, AMD's 'rack-scale' ambition becomes an expensive, low-margin inventory liability.
"Supply-chain bottlenecks and long hyperscaler qualification cycles could prevent AMD from converting Meta's capex into realized revenue, even if Meta follows through on spending."
You flag capex concentration — good — but a different, under-discussed risk is supply-chain and qualification timing. AMD needs constrained inputs (HBM, TSMC wafer/time) plus multi-quarter hyperscaler validation to supply GW-scale racks. Even if Meta spends, shortages or missed qualification windows could force Meta to stick with incumbents or delay buys. This is a timing/resource risk distinct from valuation or MTIA cannibalization (speculative but plausible).
"AMD's customer base is more diversified than implied, but power constraints threaten Helios deployment timelines."
Gemini's capex concentration fear overstates AMD's Meta reliance—datacenter revenue diversified across MSFT (40%+), Oracle, others per Q1 earnings; Meta is incremental, not dominant. Real unmentioned risk: Helios racks demand 700kW+ per rack, amplifying U.S. power shortages (PJM/ERCOT queues 2+ years). If Meta delays El Paso site for grid upgrades, AMD's 2026 growth guide slips, regardless of silicon choice.
Meta's projected 2026 capex of $115-135B fuels AI data-center demand, benefiting AMD via strong revenue growth and partnerships like Helios racks. However, the article underplays competitive threats, timeline risks, and execution challenges.
Strong revenue growth via Meta's capex and hyperscale partnerships
Capex concentration and Meta's pivot to custom MTIA silicon