Micron (MU), SK Hynix, NVIDIA (NVDA), AMD, Alibaba (BABA), and Xiaomi form a clear AI-driven ecosystem: memory suppliers power the compute leaders, while Chinese tech firms leverage AI for cloud, devices, and EVs.
As of late August 2026, AI demand remains the dominant theme, with memory shortages driving exceptional profitability for suppliers and sustained growth for GPU leaders. Chinese names show mixed results as AI investments offset core-business pressures.
Micron Technology (MU)
Recent Q3 FY2026 results (ended May 2026) were exceptional: revenue ~$41.5B (+346% YoY), adjusted EPS ~$25.11 (well above estimates), and gross margins ~84.6%. Guidance for Q4 points to ~$50B revenue and ~$31 EPS, with margins near 86%. Long-term supply agreements (including take-or-pay contracts and deposits) lock in multi-year demand; HBM4 volume shipments are ramping. Capex is rising sharply ($27B in FY2026, higher in 2027) to expand capacity.
Shares have surged hundreds of percent over the past year (market cap >$1T) but pulled back from peaks near $1,255 to around $960–$970 recently amid broader AI volatility and some NAND cooling signs. Valuation remains attractive (forward P/E in the low-to-mid teens/single digits on elevated earnings), with analyst targets implying 50%+ upside on average.
Future potential: Strong. Structural AI memory demand (HBM and high-bandwidth needs) should support elevated pricing and margins for several years. Risks include eventual supply response and cyclical pricing. Positive for AI-exposed portfolios seeking memory leverage.
SK Hynix
Q2 2026 delivered record results: revenue ~₩79.3T (~$57B, +257% YoY), operating profit ~₩60.5T (76% margin, +557% YoY). HBM strength, AI DRAM, and enterprise SSDs drove performance; long-term customer contracts are expanding. The company announced a massive ~₩40T ($28–29B) share buyback/cancellation (largest in Korea) and raised shareholder returns to >50% of cumulative FCF (2025–2027). Capex is elevated (high ₩40T range for 2026) for new fabs.
Shares rallied sharply on the buyback but have been volatile (down significantly from peaks amid sector rotation and earnings that slightly missed lofty expectations). Valuation looks inexpensive on peak earnings.
Future potential: Attractive. HBM leadership and AI demand provide multi-year visibility; aggressive capital returns add support. Risks mirror Micron’s (supply ramp, pricing). Suitable for clients seeking Korean AI memory exposure with shareholder-friendly policies.
NVIDIA (NVDA)
Q1 FY2027 revenue hit $81.6B (+85% YoY), with Data Center ~$75B (+92%). Q2 guidance is ~$91B (±2%) at ~75% gross margins (China data-center compute excluded). Demand visibility remains strong (multi-year customer commitments; hyperscaler capex forecasts continue rising). Partnerships and platform expansions (including software/models) reinforce the ecosystem.
Shares trade around $215 (market cap ~$5.2T), consolidating after strong gains. Consensus is Strong Buy with average targets near $300–$310 (substantial upside). Valuation has compressed to more reasonable levels relative to growth.
Future potential: Highest conviction in pure AI compute. Leadership in GPUs, full-stack offerings, and infrastructure financing platforms position it well for sustained multi-year growth. Key near-term catalyst: Aug 26 earnings. Risks include competition, export controls, and any slowdown in hyperscaler spend. Core holding for AI portfolios.
AMD
Q2 2026 revenue reached a record $11.5B (+50% YoY), with Data Center more than doubling to ~$6.7B (now ~58% of sales). Non-GAAP EPS $1.66. Q3 guidance ~$13B (+41% YoY). Helios rack-scale systems and MI450 accelerators are ramping, backed by major commitments from Microsoft, Meta, OpenAI, and Anthropic. EPYC server CPUs continue strong.
Shares pulled back post-earnings (to ~$470 area) on margin pressures from the AI ramp and timing of larger deployments, despite the beat. Analyst targets generally point to further upside.
Future potential: Solid challenger to NVIDIA with accelerating AI share gains and full-stack progress. 2027 should show stronger inflection as Helios scales. Risks include execution on ramps, competition, and higher valuation multiples. Good diversification play within AI semiconductors.
Alibaba (BABA)
June-quarter 2026 revenue rose 9% YoY to ~RMB 269B. Cloud/AI external revenue accelerated to +45%, with AI-related products delivering a 12th consecutive quarter of triple-digit growth. Core e-commerce (customer management revenue) remains soft amid competition and macro pressures; overall profitability was impacted by heavy AI investments and lower investment gains. New models (Qwen series) and agentic tools are advancing.
Shares have recovered from 52-week lows but remain volatile (recently around the $120 area). Valuation is inexpensive relative to AI/cloud growth potential; many analysts maintain Buy ratings with meaningful upside targets.
Future potential: Improving via AI/cloud transformation, which is increasingly offsetting e-commerce headwinds. Success depends on sustained cloud margin expansion and monetization of models. Geopolitical and regulatory risks persist. Attractive for clients seeking Chinese tech exposure with a clearer AI growth path.
Xiaomi
Q2 2026 revenue fell 6% YoY to ~RMB 109B; adjusted net profit dropped ~43% to RMB 6.2B, missing estimates. Smartphone shipments declined sharply (~26%) amid high memory costs and competition (gross margins compressed); ASP rose with a premium mix. EV + AI + new initiatives rose to ~23% of revenue (EV deliveries solid, cumulative SU7 series >500k). Management sees memory cost pressures easing in H2.
Shares have been under pressure from margin compression and China consumer weakness.
Future potential: Diversification into EVs and AI (foundation models, HyperOS, ecosystem) offers longer-term upside, but near-term smartphone/memory cost headwinds and EV profitability remain challenges. Overseas EV expansion (planned 2027) and premiumization are key. More speculative; suitable for higher-risk China growth allocations.
Overall Portfolio Perspective
The AI infrastructure supercycle continues to favor memory (Micron, SK Hynix) and compute (NVIDIA primary, AMD secondary). Chinese names provide differentiated exposure—Alibaba via cloud/AI software, Xiaomi via devices/EVs—but face higher macro and competitive risks. For private-bank clients, a barbell approach (core NVIDIA/Micron + selective AMD/Alibaba) can balance growth and diversification. Monitor hyperscaler capex, memory pricing trends, and China policy closely. Valuations for the US names have moderated after strong runs, creating more balanced entry points in places, while Chinese names trade at discounts reflecting uncertainty.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
