The electric car market in China has been facing intense competition, with Xiaomi revealed that it suffered a USD 9200 loss per vehicle from April to June. However, the price war is not the only battleground, as the focus now seems to be turned to another front.

According to a report by CNBC, chip-powered tech features, such as the driver assist function, have gradually become the latest trend, while the development of in-house chips emerges as the possible match point for China’s EV makers. The reason behind: the need for customization and a must to reduce reliance on cutting-edge AI chips amid US- China tensions.

Until now, many leading Chinese EV manufacturers have relied on NVIDIA’s chips, with the AI heavyweight’s automotive chip business generating over USD 300 million in quarterly revenue in recent years, CNBC notes.

However, Chinese electric car start-ups Nio and Xpeng both announced progress on their self-developed chips lately, signaling the beginning of a new era in which in-house chips may become the mainstream for the industry.

In late July, Nio announced that it had taped-out its self-developed intelligent driving chip, Shenji NX9031, which is manufactured with 5nm node. The chip is said to be integrated into the company’s ET9 model, which is scheduled for delivery in 2025.

Citing industrial specialists, CNBC states that the move marks the first time that 5nm has been used in the Chinese automotive industry. For now, 3nm node is primarily utilized in smartphones, personal computers, and artificial intelligence-related applications.

On the other hand, another China’s EV start-up, XPeng Motors, announced in late August that its first AI chip, Turing, has been successfully taped-out. It is worth noting that XPeng has a strong relationship with NVIDIA, and Xpeng’s former head of autonomous driving joined Nvidia last year, CNBC reports.

In 2019, Tesla has reportedly moved away from using NVIDIA’s chips to developing its own, with a focus on advanced driver-assist functions. Citing an industrial specialist, CNBC suggests that Tesla and Chinese EV startups are expected to compete by designing their own chips, while traditional automakers will likely continue to depend on NVIDIA and Qualcomm for the foreseeable future.

The report does not anticipate a significant impact on NVIDIA in the short term, as Chinese automakers are expected to test new technology in small batches within the high-end segment of the market.

Anyhow, the reason behind the wave of self-developing chips for Chinese EV makers may be that it would be difficult for a company to differentiate itself if it uses the same silicon to power its infotainment and intelligent driving systems. By designing their own chips, Chinese automakers can customize features and mitigate supply chain risks associated with geopolitical tensions.

According to the report, U.S. restrictions on NVIDIA’s chip sales to China have not directly impacted automakers, as their vehicles have not yet required the most advanced semiconductor technology.

However, with a growing emphasis on driver-assist functions, which depends heavily on artificial intelligence—a key area in the U.S.-China tech rivalry—Chinese automakers are now turning to in-house technology.

Read more

• [News] China Reportedly Demands Automakers Increase Local Chip Procurement to 25%
• [News] Chinese EV Maker XPeng Tapes out Latest In-House Chip, Boasting Triple the Computing Power of its Rivals

(Photo credit: Nio)

As per its official release, SK hynix has announced that it has developed the industry’s first 16Gb DDR5 built using its 1c node, the sixth generation of the 10nm process. Reportedly, it will be ready for mass production of the 1c DDR5 within the year to start volume shipment next year.

To reduce potential errors stemming from the procedure of advancing the process and transfer the advantage of the 1b, the company claims in the release that it extended the platform of the 1b DRAM for development of 1c.

As per the press release, the operating speed of the 1c DDR5, expected to be adopted for high-performance data centers, is improved by 11% from the previous generation, to 8Gbps.

With power efficiency also improved by more than 9%, SK hynix expects adoption of 1c DRAM to help data centers reduce the electricity cost by as much as 30% at a time when advancement of AI era is leading to an increase in power consumption.

Per a report from Businesskorea, the difficulty of advancing the shrinking process for 10nm-range DRAM technology has increased with each generation.

However, with the official release this time, SK hynix has become the first in the industry to overcome these technological limitations by achieving a higher level of design completion.

Per another report from Korea JooAng Daily, this marks a win for SK hynix, as its rival Samsung Electronics had previously outpaced it in the development of the 1b DRAM, which corresponds to nodes in the 12-nanometer range.

Supermicro Computer has reportedly been targeted by short-sellers and questioned over alleged accounting manipulations, leading to a delay in filing its 2024 10-K annual report and causing market unease.

On August 27, Hindenburg Research, a short-selling company known for targeting major entities like India’s Adani, Nikola, Lordstown, and fintech giant Block, released a report accusing Supermicro of accounting violations, inadequate disclosure of related party transactions, and evading sanctions by selling products to Russia.

At the moment following the release of Hindenburg’s report, Supermicro also announced a delay in submitting its 2024 fiscal year 10-K annual report, citing the need for more time to assess the design of internal controls and operational effectiveness. This move has further heightened market concerns about Supermicro.

Moreover, Hindenburg also raised concerns about the quality of Supermicro’s products and services, suggesting that competitors like Dell might capitalize on Supermicro’s lost orders.

According to a report from Barron’s citing Evercore ISI analyst Amit Daryanani’s report, it’s highlighted that Supermicro is facing a risk of customer order withdrawals.

Supermicro, on the other hand, didn’t immediately respond to a request for comment about the delay in filing the annual report.

Read more

• [News] Quick Couplings’ Buying Surge Reportedly Leads to AI Server Component Shortage Brought up by Supermicro
• [News] Supermicro’s Financial Results Reveal Profit Shortfall

(Photo credit: Supermicro)

TSMC is set to offer a new round of its CyberShuttle prototyping service in September. According to sources cited in a report from Commercial Times, it’s revealed that, as per usual practice, there are two opportunities each year, in March and September, for customers to submit their projects. It is indicated that the highlight this time is expected to be the 2nm process, providing leading companies with an opportunity to gain an edge.

TSMC’s 2nm technology is progressing smoothly, with the new Hsinchu Baoshan plant on track for mass production next year. Previously, there were rumors indicating that Apple is considering adopting 2nm chips in 2025, with the iPhone 17 series potentially being among the first devices to use them.

Reportedly, both TSMC’s N2P and A16 technologies are expected to enter mass production in the second half of 2026, offering improvements in power efficiency and chip density.

ASIC companies are eagerly participating in CyberShuttle this time, even though customer intentions for the first 2nm tape-out are still unconfirmed. However, this technology will likely maintain TSMC’s leadership in advanced processes, securing its future technological advantage.

CyberShuttle, also known as MPW (Multi-Project Wafer), refers to the process of placing chips from different customers onto the same test wafer. This approach not only allows for the shared cost of photomasks but also enables rapid chip prototyping and verification, enhancing customers’ cost efficiency and operational effectiveness.

Based on TSMC’s official information, the CyberShuttle prototyping service significantly reduces NRE costs by covering the widest technology range (from 0.5um to 7nm) and the most frequent launch schedule (up to 10 shuttles per month), all through the Foundry segment’s most convenient on-line registration system.)

TSMC’s CyberShuttle prototyping service also validate the sub-circuit functionality and process compatibility of IP, standard cell libraries and I/Os, reducing prototype costs by up to 90%. TSMC states that their current CyberShuttle service covers the broadest range of technologies and can offer up to 10 shuttles per month.

TSMC’s 2nm technology is expected to make its debut in September, offering opportunities for test chips.

Per the report from Commercial Times, IC design companies have pointed out that, unlike the familiar FinFET (Fin Field-Effect Transistor) structure, the industry is transitioning to the Gate-All-Around FET (GAAFET) structure, making it crucial for the market to quickly adapt.

This also allows IC design companies to provide related products to end customers, demonstrating their 2nm design capabilities.

ASIC companies have also revealed that, based on CyberShuttle data, the number of advanced process projects below 7nm is relatively small, with mature processes still dominating.

This suggests that future competition will likely focus on a few leading companies. Those who miss the first wave of 2nm technology may fall behind their competitors by up to six months, making securing a spot on the Shuttle even more critical.

Read more

• [News] TSMC Reportedly to Raise 3nm & 5nm Prices Soon, Looking to Maintain Long-Term Profit Margins 
• [News] Apple Not Adopting TSMC 2nm for iPhone 17? A19 Pro Chip Rumored to Use N3P Process

(Photo credit: TSMC)

AI chip NVIDIA provided financial guidance for the third quarter, estimating its quarterly revenue to reach USD 32.5 billion, with a fluctuation of plus or minus 2%. The figure, though beats market expectations in general, still falls short of the market’s most optimistic forecast of USD 37.9 billion, sparking concerns that its explosive growth is waning.

On the other hand, regarding whether NVIDIA’s next-gen AI chip, Blackwell, faces delay, the tech giant notes that it shipped samples this quarter, and has made an adjustment to the product to make it more efficient to manufacture, according to a report by CNBC.

Citing CFO Colette Kress, the report states that in the fourth quarter, “we expect to ship several billion dollars in Blackwell revenue.”

Q3 Guidance Somehow Fails to Impress the Market

According to NVIDIA’s guidance, the adjusted gross margin in the third quarter is expected to be 75%, with a fluctuation of plus or minus 50 basis points, slightly below the market expectation of 75.5%. The company’s gross margin for the second quarter was 75.7%, with an average expectation of 75.8%.

This quarter’s outlook, though being solid, somehow fails to impress the market. The AI chip manufacturer has been a major beneficiary as companies rush to upgrade their data centers to handle AI software, and its sales forecast has become a barometer of this AI surge.

A report by the Economic Daily raises concerns that though the company has had several consecutive quarters of outstanding performance, most of its growth comes from a small number of customers. For instance, about 40% of its revenue comes from large data center operators, such as Google and Meta.

Although these tech giants have been investing heavily on AI, there are concerns that the scale of the infrastructure being built may exceed current demand, which could lead to a bubble.

Blackwell Update: Obstacles Removed?

There were concerns that there might be design issues with Blackwell, which might cause the shipment to delay. According to the report by the Economic Daily, NVIDIA admitted that it has encountered difficulties in the production process, but stated that they are implementing reforms to improve yield rates.

A report by Wccftech states that NVIDIA has confirmed that it has “implemented a change” to the Blackwell GPU mask produced at TSMC using the 4NP process node, which is expected to further enhance production yield. As this change is not a significant architectural redesign, the product remains on schedule for the ramp to start in the fourth quarter and continuing into FY2026.

According to Wccftech, NVIDIA CEO Jensen Huang reiterates that “Hopper demand remains strong, and the anticipation for Blackwell is incredible.” CFO Colette Kress further states that in the fourth quarter, “we expect to ship several billion dollars in Blackwell revenue.”

Q2 Performance Still Beats Market Expectations

In the second quarter, which ended on July 28, NVIDIA’s revenue was USD 30.04 billion, surpassing the market expectation of USD 28.7 billion. The annual growth rate for this quarter was 122%, marking the third consecutive quarter with a growth rate exceeding 200%. Net income for the quarter more than doubled from USD 6.18 billion, or 25 cents per share, a year ago to USD 16.6 billion, or 67 cents per share. Adjusted earnings per share were 68 cents, beating the market expectation of 64 cents.

It is worth noting that NVIDIA’s data center business revenue in the second quarter increased significantly by 154% YoY to $USD 26.3 billion, accounting for 88% of total sales. Its gaming revenue also grew 16% year over year to USD 2.9 billion, exceeding the expected USD 2.7 billion, which the company attributes to an increase in shipments of PC gaming cards and game console SoCs, reportedly chips for Nintendo’s game consoles.

The company also announced the approval of a USD 50 billion share buyback program.

Read more

• [News] SK hynix, TSMC, and NVIDIA Reportedly Forge Alliance to Develop Next-Generation HBM
• [News] Foxconn Confirms Shipment of NVIDIA’s GB200 to Start in Q4 as Planned

(Photo credit: NVIDIA)