Nvidia, Micron Gain as AI and China Tailwinds Build

December 24, 2025 at 19:24 UTC

6 min read

Nvidia and Micron stock surge with AI and China market optimism in semiconductor sector

Key Points

Nvidia plans initial H200 AI chip shipments to China by mid-February, pending Beijing’s approval
Raymond James sees up to $12.5 billion in 2026 China AI revenue upside for Nvidia
Micron’s AI-driven memory boom has tripled its stock and sparked a sector ‘supercycle’ view
Bank of America projects global chip sales to top $1 trillion in 2026, led by AI demand

Nvidia moves to restart high-end AI chip sales into China

Nvidia is testing a path back into the Chinese market for advanced AI processors following a shift in U.S. export policy. Multiple reports say the company has told Chinese clients it aims to begin shipping its H200 AI chips before the Lunar New Year holiday in mid-February, using existing inventory. Initial shipments are expected to total 5,000 to 10,000 modules, equivalent to roughly 40,000 to 80,000 individual H200 chips, if Chinese authorities approve the purchases. The plan remains contingent on Beijing’s sign-off, with one source noting that nothing is certain until official approval is granted. The renewed export opportunity follows a U.S. policy change that allows H200 shipments to China subject to a 25% fee and an inter-agency license review process. Chinese cloud and internet companies including Alibaba and ByteDance have expressed interest in the H200, which is described as markedly outperforming the downgraded H20 chips currently available in China. Nvidia has said it manages supply continuously and that any licensed China sales would not affect U.S. customers, but analysts caution that regulatory and geopolitical risks could still alter delivery plans.

Analysts quantify potential China upside for Nvidia and AMD

Raymond James analysts see significant incremental revenue for U.S. chipmakers if AI chip exports to China resume more broadly. The firm estimates Nvidia could generate $7 billion to $12.5 billion in additional revenue in 2026 from limited GPU shipments to Chinese cloud and internet companies, assuming U.S. authorities allow renewed sales. Advanced Micro Devices is seen benefiting to a lesser extent, with a best-case scenario of $500 million to $800 million in extra sales tied to Chinese demand, equating to about $0.10 to $0.20 in non-GAAP earnings per share. Earlier commentary from Raymond James characterized the China upside for both companies as potentially meaningful but difficult to forecast given evolving trade policies and export limits. The analysts noted that if restrictions are eased or rules clarified, Chinese buyers may be able to purchase more of the AI hardware they want. Even with new approvals, Nvidia’s China exposure is described as modest relative to its U.S. data center business, which continues to drive most of its growth. Both Nvidia and AMD currently carry Strong Buy ratings from Wall Street, with analysts cited as seeing slightly higher upside for Nvidia shares.

AI boom underpins bullish forecasts for Nvidia’s core business

Beyond China, Nvidia remains central to the global build-out of AI infrastructure. Its graphics processing units are widely used to train and deploy AI models, and its data center platform generated $51.2 billion of the company’s $57.0 billion in fiscal third-quarter revenue. At Nvidia’s GPU Technology Conference in Washington, D.C., CEO Jensen Huang said management has visibility into $500 billion of demand for its key data center technology over the next five quarters, or about $100 billion per quarter on average. Zacks Investment Research highlights that roughly half a trillion dollars is expected to flow into AI-related capital spending next year, with Nvidia’s chips estimated to account for about 20% of that investment. Consensus forecasts cited by Zacks call for Nvidia’s sales to rise roughly 63% next year and another 43% in 2027, with earnings expected to climb 55.5% and 53% over those years. Despite this growth profile, Nvidia is reported to trade at about 39.5 times forward earnings, with a PEG ratio near 0.85, and has recently broken out technically above its 50-day moving average, earning high rankings from Zacks.

Micron’s high-bandwidth memory surge signals broader chip supercycle

Micron Technology has emerged as one of the strongest beneficiaries of AI-driven demand, particularly for high-bandwidth memory (HBM) used alongside GPUs in data centers. The company reported fiscal first-quarter 2026 revenue of $13.6 billion, up 56% to 57% year over year, marking its third consecutive quarterly revenue record. Gross margin expanded to 56.8%, and adjusted earnings per share reached $4.78, rising 167% from the prior-year period. Micron’s cloud memory segment, which includes data center HBM sales, doubled to $5.3 billion in the quarter. The company expects fiscal second-quarter revenue between $18.3 billion and $19.1 billion, with a midpoint of $18.7 billion implying roughly 132% year-over-year growth, and projects EPS of $8.22 to $8.62. Micron says its entire calendar 2026 HBM supply, including next-generation HBM4 products, is already under pricing and volume agreements. It now forecasts the HBM total addressable market to grow at a 40% annual rate from $35 billion in 2025 to about $100 billion by 2028. Bank of America and other firms describe Micron’s performance as part of a memory “supercycle” tied to AI, with supply constraints so severe that Morgan Stanley estimates about 30% of end demand is currently going unmet.

Sector outlook: toward $1 trillion in chip sales and multi-year AI build-out

The strength at Nvidia and Micron is feeding into increasingly bullish views on the broader semiconductor industry. Bank of America analyst Vivek Arya forecasts that global semiconductor sales will surge 30% year over year in 2026, pushing the sector past $1 trillion in annual revenue for the first time. The bank estimates the total addressable market for AI data center systems will exceed $1.2 trillion by 2030, growing at a 38% compound annual rate, with AI accelerators alone representing a $900 billion opportunity. Arya characterizes Nvidia as operating in a “different galaxy,” noting that while the average chip sells for about $2.40, an Nvidia GPU is priced around $30,000, and Bank of America projects the company’s free cash flow could reach roughly half a trillion dollars over the next three years. At the same time, analysts at Morgan Stanley and others see a multi-year data center expansion cycle driven by AI workloads, with memory and storage suppliers planning capacity ramps over the next two to three years. These firms expect current supply dislocations to normalize as new fabs and nodes come online, while still supporting elevated operating profits across the AI hardware ecosystem.

Key Takeaways

Nvidia’s planned H200 shipments and potential policy easing could reopen China as a meaningful, though still secondary, AI revenue stream.
Analysts see Nvidia’s core AI data center demand and earnings trajectory as strong enough that China-related gains would be incremental rather than foundational.
Micron’s record results and fully booked HBM capacity suggest AI is reshaping memory from a cyclical commodity to a strategic growth driver.
Forecasts for a $1 trillion chip market and a multi-year AI data center build-out indicate that current strength at Nvidia and Micron reflects a broader industry upcycle.
Tight supply and large planned capacity additions point to continued volatility, but also to sustained confidence among chipmakers in long-term AI infrastructure demand.