By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel is divided on AMD's stock outlook, with concerns about margin compression, intense competition, and supply chain risks, but also acknowledging the potential of AI inference demand and the 2nm Venice ramp.
Risk: TSMC capacity allocation and customer validation of Venice in 2026
Opportunity: AI inference-driven demand and the 2nm Venice ramp
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 →
Wells Fargo analyst Aaron Rakers raised his price target on Advanced Micro Devices (AMD) stock to $615 on Tuesday, June 30, maintaining an "Overweight" rating on the chipmaker.
The new target implies roughly 14% upside from recent trading levels, making it one of the more bullish positions among Wall Street analysts covering the stock.
AMD shares responded positively, gaining nearly 8% today, and are now up an overall 170% year-to-date.
The core thesis behind the price target increase centers on AMD's growing role in AI infrastructure, particularly as the industry transitions from AI model training toward AI inference workloads.
Rakers expects this shift toward running AI applications in real-world environments to materially increase demand for AMD's high-core-count EPYC server CPUs alongside its Instinct AI-focused GPUs, allowing the company to compete more aggressively with Nvidia.
The analyst sees AMD's server CPU business generating some $16 billion in revenue this year, representing a 68% year-over-year increase, followed by growth to $20.5 billion in 2027 and $25 billion in 2028.
On the GPU side, Rakers forecasts data center GPU revenue of $15.6 billion in 2026, growing to $40.6 billion next year and nearly $63 billion in 2028, reflecting accelerating adoption of artificial intelligence computing hardware across cloud providers and enterprises.
These projections underpin his earnings estimates of $13.40 per share in 2027, roughly 3% above Wall Street consensus, and $18.75 per share in 2028, about 8% above consensus.
He further believes AMD has the potential to approach $20 per share in peak earnings, a level that could be reached earlier than the company itself has guided.
A key product cycle catalyst supporting the bull case is AMD's next-generation 2nm EPYC Venice chips, which entered production in late May 2026 with volume shipments expected in the second half of the year.
Customer validation and ramp activity for Venice is reportedly running ahead of any prior EPYC generation, signaling strong market pull.
Four leading AI models discuss this article
"AMD’s transition to inference-heavy workloads provides a durable path to $18+ EPS by 2028, provided they maintain gross margins above 50% amidst competitive pricing pressure."
Rakers’ $615 target is aggressive, effectively pricing in a flawless execution of the MI300/MI400 roadmap. The shift from training to inference is the correct structural tailwind, as AMD’s EPYC CPUs provide a crucial architectural advantage in power-efficient inference clusters. However, the market is ignoring the margin compression risk inherent in the transition to custom silicon and the intense competition from Nvidia’s Blackwell platform. While the 2nm Venice ramp is a strong catalyst, AMD’s ability to maintain a 50%+ gross margin while scaling these massive revenue targets is the real hurdle. If AMD hits these EPS targets of $18.75 by 2028, the stock is currently trading at a reasonable forward multiple, but there is zero room for supply chain or yield errors.
The bull case assumes AMD can capture massive market share from Nvidia without triggering a price war that would permanently erode the high-margin profile of their data center segment.
"The bull case hinges entirely on AMD capturing 30-40% of incremental AI inference GPUs by 2028—a market share swing that has never happened in AMD's history against Nvidia in data center."
Rakers' $615 target implies 14% upside, but the real story is the GPU forecast: $15.6B→$40.6B→$63B (2026-2028) requires AMD to capture ~40% of incremental AI inference spend while Nvidia defends 80%+ market share. The Venice ramp is real and ahead of schedule, but the earnings bridge—$13.40 (2027) to $18.75 (2028) to $20 peak—assumes flawless execution, zero competitive pressure from Nvidia's next gen, and sustained cloud capex. AMD is 170% YTD; much of this upside is priced in. The call option volume spike today may reflect profit-taking into strength, not fresh conviction.
If Nvidia's next-gen H200/Blackwell derivatives maintain 85%+ share of inference workloads (plausible given software lock-in and customer inertia), AMD's GPU revenue tops out ~$25-30B by 2028, not $63B, crushing the 2028 EPS estimate and justifying a 12-15x multiple, not 30x+.
"AMD faces steeper competitive and execution risks in capturing AI inference share than the price-target raise implies."
Wells Fargo's upgrade to a $615 target on AMD bets heavily on inference-driven demand lifting both EPYC server CPUs to $25 billion by 2028 and data-center GPUs to $63 billion. The 2nm Venice ramp starting late May 2026 looks promising on paper, yet the article downplays Nvidia's software lock-in and AMD's track record of missing aggressive share targets. Current valuation already prices in substantial success; any delay in customer validation or slower-than-expected inference adoption would compress multiples quickly. The 170% YTD run leaves little margin for execution slips.
If AMD's Instinct GPUs achieve even modest share gains in inference workloads and Venice volumes exceed prior cycles, the $20 peak EPS scenario could materialize faster than consensus expects, validating the aggressive revenue forecasts.
"The core upside hinges on a flawless Venice 2nm ramp and sustained AI data-center capex; if either falters, the bullish price target is at significant risk."
AMD's Wells Fargo bull case relies on AI inference driving data-center spend, with Venice 2nm ramp as a near-term catalyst. The article's numbers imply a multi-year EPS path above consensus (e.g., $13.40 in 2027, $18.75 in 2028) and GPU revenue swelling to $63B by 2028. But upside is contingent on a rapid, smooth 2nm Venice ramp and sustained AI capex, which is far from assured. Nvidia's entrenched position, potential demand softness in cloud spend, and ramp risks could derail the bull case even if AI stays secular.
Strongest counter: AI hardware demand remains cyclical and Nvidia's dominance in inference GPUs could cap AMD's share gains; a delayed or weaker Venice ramp would undercut the entire thesis.
"AMD's aggressive revenue targets are physically constrained by TSMC 2nm capacity allocation, regardless of product demand."
Claude and Grok focus on the $63B GPU target, but you are all ignoring the critical bottleneck: TSMC capacity allocation. AMD is competing for 2nm wafers against Apple and Nvidia's next-gen Blackwell Ultra. Even if Venice is technically superior, AMD lacks the foundry priority to guarantee the volume required for these revenue targets. If supply constraints persist, AMD’s gross margins will be sacrificed to secure wafer priority, invalidating the EPS expansion models discussed.
"TSMC capacity is a 2027+ risk; 2026 validation risk is the binding constraint."
Gemini's TSMC capacity argument is real, but it conflates two separate risks. Yes, wafer scarcity could force margin trade-offs. But AMD's actual 2nm volume needs for the $63B GPU target don't kick in until 2027-2028—by then, TSMC's 2nm capacity expands materially. The nearer bottleneck is customer validation of Venice in 2026. If cloud providers don't adopt at scale, wafer allocation becomes moot. Sequencing matters.
"TSMC pre-allocates 2nm wafers years ahead, so 2026 validation cannot unlock enough capacity for AMD's 2027-2028 targets."
Claude correctly sequences validation before capacity constraints, yet misses that TSMC pre-allocates 2nm wafers years ahead via multi-year commitments from Nvidia and Apple. AMD cannot simply wait until 2027 for expanded capacity; priority slots are claimed now. Any validation success in 2026 arrives too late if those wafers are already spoken for, directly capping the $63B GPU trajectory regardless of cloud adoption.
"Cloud-hyperscaler adoption risk could negate the upside even if Venice ramps and wafer capacity improves."
Your TSMC capacity concern is real but incomplete. Even with priority wafers, the bigger gate is cloud-hyperscaler validation in 2026-27. If Venice proves non-competitive on TCO or cloud buyers stagger rollouts, the $63B GPU target collapses regardless of wafer slots. The analysis should split supply risk from demand timing risk; a failure of early adoption could kill upside before capacity expands, and Nvidia's software moat amplifies that sensitivity.
The panel is divided on AMD's stock outlook, with concerns about margin compression, intense competition, and supply chain risks, but also acknowledging the potential of AI inference demand and the 2nm Venice ramp.
AI inference-driven demand and the 2nm Venice ramp
TSMC capacity allocation and customer validation of Venice in 2026