The semiconductor world buzzes as NVIDIA (NASDAQ: NVDA) plans to ramp up production of its Blackwell chips. While this development presents exciting opportunities, challenges remain. The favorable aspect is the potential for expanded Blackwell output and minimized geopolitical risk. However, Taiwan Semiconductor (NYSE: TSM) is still not fully equipped to manufacture these chips in America—at least, not right now.
NVIDIA’s Goals: Onshoring Blackwell Production
Bringing NVIDIA chip production to the U.S. aligns with national interests and lessens dependence on China, although this shift isn’t feasible currently. Although TSMC’s Arizona plant is cutting-edge, it lacks the capability to create CoWoS (Chip-on-Wafer-on-Substrate) packaging, which is essential for advanced computing and AI applications. As a result, Blackwell chips must be sent to Taiwan for finishing, reintroducing some geopolitical complexities. Still, there is optimism for the future.
As part of its commitment to domestic chip production, TSMC received $6.6 billion from the CHIPS Act amid a broader $65 billion investment strategy. This funding aims to establish new facilities, potentially including a CoWoS production site crucial for onshoring Blackwell chips. In addition, TSMC is eyeing plans for additional factories in Japan and the U.S. to cater to surging demand, with intentions to significantly increase production capacity before 2026.
Moreover, a collaboration with Amkor (NASDAQ: AMKR) may meet NVIDIA’s requirements in the meantime. Amkor offers global microchip packaging services and will handle CoWoS production at its Arizona facilities. Earlier this year, TSMC and Amkor signed a partnership, and plans for the first production site are underway. Key early customers like Advanced Micro Devices (NASDAQ: AMD), NVIDIA’s main competitor in the AI GPU and CPU arena, along with Apple (NASDAQ: AAPL), are poised to benefit. Unfortunately, the initial phase isn’t expected to be completed until 2027 or beyond, with final stages projected for the middle of the next decade.
NVIDIA’s Potential Market Surprise with Rubin in 2025
Market speculation suggests that NVIDIA might surprise investors by launching its next generation of AI technology six months sooner than previously thought. This news is buoying share prices and may lead to sustained upgrades in data centers for the foreseeable future. The upcoming Rubin chips should yield improved performance thanks to AI-specific designs, TSMC’s advanced 3nm manufacturing, and next-gen HBM4 memory. These chips are anticipated to be more efficient, store significantly more data, and deliver superior performance compared to earlier models. NVIDIA is also working with Schneider Electric on novel cooling systems for its cutting-edge GPU and CPU products.
While MarketBeat has not tracked analyst activity specifically tied to TSMC’s Arizona developments, general market trends show positive momentum for 2024. Analyst coverage is increasing, indicating a Strong Buy sentiment, alongside rising consensus price targets. In November, the target elevated by nearly 15%, with further revisions pushing estimates even higher. The consensus suggests about 15% upside potential from a key support level, while the highest estimates suggest an additional 25% gain.
Late 2024 shows a bullish trend for NVIDIA, with stock prices reaching new all-time highs and holding strong at critical levels. This key support aligns with a prior all-time peak and the upper limit of a consolidation area. The next significant upward movement could push stock prices beyond the crucial resistance target of $150, which may lead to additional upward momentum, possibly driving the market towards the $190 to $200 range.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
