Read More : Panther Lake to Feature New Core Ultra X CPUs

Huawei manufactures the Ascend 910C at SMIC, China’s largest foundry, aiming for complete vertical integration. However, SMIC faces challenges with its 7nm process yields, which limits production capabilities. To bypass export controls, Huawei allegedly used a shell company to obtain nearly three million TSMC dies. These dies are essential for assembling the AI accelerators Huawei claims to produce domestically.

TSMC confirmed the dies in the teardown are from a batch analyzed in October 2024, not newer models. The company stated it halted production and sales to Huawei after discovering the export violations. Semiconductor teardown experts verified that the 910C uses these older TSMC dies along with previous-generation HBM chips from Samsung and SK Hynix.

The use of these foreign components highlights ongoing challenges for Huawei. While it pushes to build AI chips domestically, critical elements like advanced dies and HBM memory remain dependent on international suppliers, especially amid strict export restrictions. This reliance poses a risk to Huawei’s ambitions to compete with global leaders like Nvidia.

Production Bottlenecks and Future Outlook for Huawei’s AI Chips

Huawei’s Ascend 910C AI chip faces significant production bottlenecks, especially concerning high-bandwidth memory supply. Although Huawei stocked some Samsung and SK Hynix HBM chips before export controls tightened, these are aging components running low in inventory. To address shortages, Huawei reportedly desolders memory from other products repackaged solely to reclaim HBM chips.

This HBM shortage is the biggest hurdle in scaling production. The 910C delivers roughly half the performance of Nvidia’s popular H100 AI chip. Despite being less powerful, it reportedly costs Huawei nearly as much to produce as the H100 sells for on Amazon.

Additionally, the Ascend 910C suffers from subpar packaging and thermal inefficiencies. Huawei must still manufacture millions of these chips to meet growing domestic AI demand. Analysts estimate that Huawei will produce around one million 910C units in 2026 due to these supply constraints.

Read More : Final Fantasy 7 Remake Part 3 Hints at Wider Launch

However, the Chinese government is investing billions in local AI chip and HBM manufacturing facilities. These loans aim to reduce reliance on foreign components and boost production capacity. As domestic foundries improve, Huawei could increase output and improve its chip designs.

In summary, Huawei’s Ascend 910C AI chip shows China’s semiconductor ambitions but also reveals critical supply vulnerabilities. Overcoming foreign component dependence remains essential for Huawei’s success in AI hardware. With government support, China may accelerate its push toward fully self-reliant AI chip production in the near future.

The post Huawei 910C Uses TSMC Die and Samsung Memory, Replacing Nvidia appeared first on Senior Spectrum Newspaper.

Read More : Silent Hill f Early Review Score Surfaces Ahead of Launch

This speculation originated from the YouTube channel Moore’s Law is Dead. Which suggested that some AMD Zen 6 CPU tiles could be produced using TSMC’s N2X process. AMD’s upcoming Epyc Venice CPU has already taped out on an unspecified N2 node, adding some credibility to these rumors. However, Nvidia appears poised to push ahead by leveraging the newer A16 node first.

Historically, Nvidia tends to adopt TSMC’s latest process nodes a generation later than the bleeding edge. For example, the recent RTX 50 series GPUs use a custom 4NP node instead of the TSMC N3 process. This 4NP node was developed for Nvidia’s upcoming Rubin chips, which have already completed the tape-out stage. Additionally, Nvidia’s RTX 60 series GPUs, launching next year, will likely continue using the 3N or 3NP processes.

TSMC’s A16 node is notable for being the first to introduce backside power delivery technology. A feature Intel already uses in its 18A process. This technique moves power delivery lines to the back of the wafer, improving power efficiency and thermal performance. TSMC also claims the A16 node will deliver an 8-10% speed boost, a 15-20% reduction in power consumption at equal speeds, and up to a 1.10x increase in chip density compared to previous nodes.

What Nvidia’s Adoption of TSMC A16 Means for Future Performance and Launch Timeline

According to an earlier product roadmap highlighted by Tom’s Hardware, TSMC expects the A16 node to enter mass production in late 2026. This timeline suggests that any Nvidia product built on A16 technology will likely debut in mid-to-late 2027 or later. The Taiwanese report aligns with Nvidia’s plans to use the node for its 2028 Feynman chips, which were previewed earlier this year.

Adopting the A16 node could help Nvidia achieve greater power efficiency and higher clock speeds in its future GPUs and processors. The backside power delivery technology also offers better thermal management, which is crucial for high-performance computing. This technical edge might allow Nvidia to maintain its competitiveness against AMD, which is pursuing a different advanced node path with TSMC’s 2nm process.

Read More : Spotify Gets Free Song Selection and New Features

Nvidia’s strategy to leverage the A16 node earlier than usual marks a shift in its approach to semiconductor manufacturing. It indicates a desire to secure a node advantage and improve performance-per-watt metrics in the long term. However, the actual impact on gaming and professional workloads will depend on how Nvidia integrates these manufacturing benefits into its chip architecture.

As competition intensifies in the semiconductor industry, early adoption of advanced nodes like A16 could be a key differentiator. Nvidia’s move may pressure other chipmakers to accelerate their own manufacturing roadmaps. The coming years will reveal how well Nvidia’s gamble on TSMC’s A16 process pays off in real-world products and performance.

The post Nvidia Set to Adopt TSMC’s Latest A16 Chip Technology appeared first on Senior Spectrum Newspaper.

Read More : Z2 Mini G1a Launches as HP’s First Strix Halo Mini PC

The company’s gross margin expanded to 58.6 percent, with an operating margin of 49.6 percent and net margin reaching 42.7 percent—levels rarely seen among global semiconductor firms. Chief Financial Officer Wendell Huang credited strong results to “continued robust AI and high-performance computing demand.” Advanced manufacturing nodes once again played a leading role in the quarter’s success.

Processes at or below 7 nanometers accounted for 74 percent of total wafer revenue. Specifically, 3-nanometer chips made up 24 percent of sales, 5-nanometer chips contributed 36 percent, and 7-nanometer processes added 14 percent. Chief Executive C.C. Wei called Nvidia’s newly reauthorized H20 shipments to China “very positive news,” highlighting their shared position in the AI ecosystem.

TSMC’s quarterly revenue increased 17.8 percent over Q1, despite a stronger Taiwan dollar that typically weighs on export earnings. Management now expects third-quarter revenue to range between $31.8 billion and $33.0 billion, which would mark up to 40 percent year-over-year growth.

Outlook Raised Despite Margin Pressures and Global Uncertainty

While Q2 performance exceeded expectations, TSMC anticipates some pressure on margins going forward. For the third quarter, the company guides gross margins between 55.5 and 57.5 percent and operating margins between 45.5 and 47.5 percent. These declines are expected due to rising operating costs and the strengthening Taiwan dollar.

The full-year revenue forecast has been raised to around 30 percent growth in U.S. dollar terms, a notable upgrade from earlier projections. However, several risks could weigh on future performance. These include potential semiconductor-specific tariffs in the United States, continued currency appreciation, and weaker-than-expected iPhone sales.

Additionally, TSMC is investing heavily in global expansion. It plans to maintain capital expenditures between $38 billion and $42 billion, with major new fabs underway in Arizona and Japan. While these projects will increase long-term capacity, the near-term costs will impact profitability. Despite these challenges, customer demand remains strong, with no signs of a slowdown in orders.

Read More : North Korea Closes New Seaside Resort to Foreign Visitors

So far in 2025, TSMC shares have gained just 5 percent after surging roughly 80 percent last year. Investor concerns about policy shifts and foreign exchange volatility have tempered enthusiasm, even as demand for generative AI and advanced chips continues to rise.

As the center of the AI chip supply chain, TSMC remains well-positioned to meet the growing needs of the tech industry. Whether the company can sustain its industry-leading margins amid geopolitical and economic headwinds will become clearer in the quarters ahead.

The post TSMC Q2 Revenue Soars to $31.8B on AI Chip Demand appeared first on Senior Spectrum Newspaper.