All articles tagged with #semiconductors
Meta is pursuing a massive AI-chip supply deal with AMD that targets about 6 gigawatts of compute, potentially exceeding $100 billion in value, with a pathway for AMD to hold up to a 10% stake in the venture.
A New York Times investigation reveals a July 2023 classified CIA briefing warned China could attack Taiwan by 2027, a session attended by Tim Cook and other tech CEOs. The report highlights Taiwan's central role in global chip production via TSMC and the potential economic fallout if access is disrupted. In response, Washington has pushed for domestic chip manufacturing through CHIPS Act subsidies, with Apple pledging substantial U.S. investment and TSMC committing large U.S. investments; however, the so‑called silicon shield remains uncertain as geopolitical tensions persist and Taiwan seeks to retain advanced manufacturing on the island.
Meta has struck a multibillion-dollar deal with AMD to buy 6 gigawatts of custom MI450 AI chips to accelerate its AI models, while AMD granted Meta a performance-based warrant to buy up to 160 million AMD shares at $0.01, with the first tranche due in H2 as the initial gigawatt ships; the arrangement could yield about a 10% stake in AMD over time, echoing OpenAI’s deal, and sent AMD stock higher pre-market. Meta plans to spend up to $135bn on AI infra this year and seeks to diversify beyond Nvidia, using the chips mainly for inference workloads that require 6GW; the warrant expires Feb 2031 and final tranche triggers at $600 per share.
US trade deficit widened in December as imports rose 3.6%, led by computer chips and other tech goods and boosted by capital goods for AI data centers. The 2025 goods deficit hit a record, while exports fell slightly; trade with China declined but deficits with Taiwan and Vietnam grew, and tariffs stopped delivering a manufacturing rebound.
Meta has agreed to spend billions in a multiyear deal to buy Nvidia’s next‑generation chips, including Vera Rubin GPUs and standalone CPUs, to expand its AI infrastructure while continuing work on in‑house processors. The arrangement reinforces Nvidia’s lead in AI data‑center hardware even as rivals like AMD and tech giants such as Google and Microsoft push self‑developed chips. Meta’s AI investment is expected to be substantial, with estimates of up to $135 billion in AI infrastructure this year, highlighting a broader industry shift toward inference workloads and greater chip self-reliance, despite ongoing rollout challenges for Meta’s own designs.
Cantor Fitzgerald’s C.J. Muse says Nvidia’s upcoming quarterly results could mark the start of a rally, anchored by a Rubin AI chip ramp and a booming demand for AI compute. With shares trading around 21x 2026 earnings and a backlog-building run into 2027–28, Muse sees upside toward $400 as Nvidia highlights its AI leadership at key events and potentially lifts data-center guidance, though investors remain wary of competition from Google’s TPUs and broader AI-deal headwinds.
The article argues Nvidia, Alphabet, and Taiwan Semiconductor (TSMC) are strong long‑term holdings due to AI-driven data‑center demand, escalating capex by leading tech players, and solid fundamentals with attractive relative valuations, supporting a buy‑and‑hold strategy despite overall market volatility.
A surge in 2026 AI-related spending by hyperscalers is accelerating data-center capex, with Amazon, Alphabet, and Meta guiding over $500 billion in planned spend. Nvidia, Broadcom, and Taiwan Semiconductor stand to benefit the most since chips can account for about half of data-center costs; analysts expect strong revenue growth for all three (Nvidia ~64% in 2027, Broadcom ~51% in 2026, TSM ~34%), and Nvidia even projects global data-center capex topping $3–4 trillion annually by 2030. The buildout is just beginning, making these three stocks attractive picks amid the AI infrastructure boom.
A roughly $700 billion AI infrastructure capex wave from hyperscalers like Amazon, Alphabet, Meta, and Microsoft has Nvidia underperforming relative to the spend, while suppliers such as Broadcom, Micron, Lumentum, and Bloom Energy have posted strong gains. The AI buildout spans chips, memory, optical components, and AI-powered power infrastructure, signaling multiple winners beyond Nvidia as the sector expands.
Taiwan’s top trade official called Washington’s bid to move 40% of the island’s semiconductor supply chain to the U.S. 'impossible,' arguing the mature ecosystem cannot be relocated. The stance comes as Taiwan and its chip giant TSMC continue expanding in the U.S. under CHIPS Act incentives, despite analysts’ view that relocating such a complex supply chain is impractical due to costs, labor constraints, and geopolitical considerations like the Silicon Shield.
Taiwan’s vice premier says moving 40% of Taiwan’s chip output to the U.S. isn’t feasible, signaling that Taiwan will continue expanding at home even as firms grow presence abroad. Separately, TSMC is investing about $165 billion in new Arizona plants, underscoring growth in the U.S. While a tariff deal reduces Taiwan exports to 15% from 20% in exchange for increased U.S. investment, U.S. policy aims to lift domestic chip output gradually, suggesting a slow, costly shift that spreads supply over time rather than a rapid, across-the-board relocation.
The article argues Tesla’s EV-turned-robotics/AV pivot is expensive and slowing its core business, with high costs and a rich valuation, making it less attractive now. In contrast, Micron and Taiwan Semiconductor are benefiting from surging AI infrastructure demand, reporting strong recent earnings growth and trading at much cheaper multiples, positioning them as the stronger AI stock picks today.
Taiwan’s vice premier said moving 40% of the island’s semiconductor capacity to the United States is “impossible,” emphasizing that Taiwan’s chip ecosystem will continue to grow at home while expanding its presence in the U.S.; Washington has pressed for more onshore production, and last month a tariff deal reduced Taiwan’s export tariffs to 15% as Taiwan increases U.S. investment; TSMC is also investing heavily in Arizona.
As AI workloads intensify, memory chips—especially Micron's high-bandwidth memory—are becoming a key growth driver, with demand outpacing GPUs and a TAM expected to grow from about $35B to $100B by 2028. The piece suggests Micron could follow Nvidia's playbook, trading at a modest forward P/E around 12 and offering upside from a fast-expanding memory market, potentially earning the label of 'new Nvidia' in investors' eyes.
MarketWatch identifies 20 of 82 large-cap tech names with the fastest projected 2026 sales growth, led by Sandisk (SNDK) at about 131% and Micron (MU) at 65.5%; Palantir (PLTR), Nvidia (NVDA), Broadcom (AVGO) and others follow, with Palantir guiding to roughly 61% revenue growth in 2026 per its earnings, and Wedbush calling Palantir a top tech pick. The trend skews toward hardware and AI-related winners as Google and Amazon lift capex for chipmakers, while several names trade at lower forward P/Es than the IT sector despite higher earnings forecasts. Overall, the 2026 sales growth for these names far surpasses the S&P 500 information-technology sector’s expected ~17% sales growth for the year.