Is the AI bubble about to burst? Two years after OpenAI launched ChatGPT, sparkling a surge in generative AI startups, China has now seen a wave of closures on AI companies. According to a report by Commercial Times, citing Chinese media TMTPost, nearly 80,000 AI companies in China have registered and then either closed down or suspended operations within the past 600 days.

The reports note that according to the data from the National Enterprise Credit Information Publicity System of China, between November 30, 2022, the release date of ChatGPT, and July 29, 2024, a total of 78,612 AI-related companies in China, which were newly registered during this period, are now in a deregistered or abnormal business status. This accounts for 8.9% of the 878,000 AI companies registered during the same period.

The current adversity for the AI companies in China, the reports suggest, could be attributed to excessive spendings regarding high computational costs and R&D investments, declines in venture capital leading to a funding freeze, and difficulties in achieving profitability.

According to the reports, over the past three years, more than 200,000 AI-related companies in China have been deregistered or revoked, with a total of 353,000 AI-related companies disappearing within the past decade.

On the other hand, as of August 7th, there are said to be 300,700 new AI companies registered in 2024. Currently, there are 1,804,300 AI-related companies in existence in China, the reports state.

Among them, over 4,500 companies are officially recognized as part of the AI industry system. More than 180 large generative AI models, which have completed registration and be online to provide public services, have been developed, with a registered user base exceeding 564 million.

Wang Xiaochuan, founder of Chinese search engine company Sogou, once stated in 2023 that the Large Language Model (LLM) for AI would undergo an “elimination tournament” in China, with the top tier likely consisting of no more than five companies, according to a report on Soho.

Baidu CEO Robin Li also mentioned in July that China has too many large language models, calling for tech leaders to focus more on building real-world applications driven by AI. A report by South China Morning Post, quoting Li, said that since the launch of ChatGPT in late 2022, China’s generative AI market has become crowded with over 200 large language models.

Read more

• [News] China Reportedly Imposes Mandatory AI Model Reviews on ByteDance, Alibaba, and Others

(Photo credit: )

According to a report by Nikkei, Japanese chip manufacturer Rapidus plans to establish a fully automated production line using robots and AI in northern Japan to produce 2nm chips for advanced AI applications, with mass production anticipated as early as 2027.

Reportedly, Rapidus claims that automated production will significantly accelerate production times, reducing chip delivery time to just one-third of that of its competitors. The company’s fab is expected to complete its external structure by October, with EUV lithography system set to arrive in December.

Compared to other companies already operating fabs, building a fully automated plant could give Rapidus a significant advantage. While the front-end of chip manufacturing are already highly automated, the back-end processes, such as packaging and testing, remain labor-intensive.

Rapidus CEO Atsuyoshi Koike stated that this approach will deliver higher performance and faster turnaround times for the same 2nm products compared to other competing chipmakers.

Per a report from Tom’s Hardware, Rapidus is currently two years behind TSMC and Samsung, both of which are expected to begin 2nm chip production in 2025. If Rapidus can deliver chips faster without compromising on price or quality, it may secure a place in the market.

Despite the optimistic outlook, Rapidus faces operational challenges. The company revealed that it will need JPY 2 trillion (approximately USD 14 billion) to begin operation in 2025, and at least JPY 5 trillion in total for the start of mass production.

Although Rapidus has received JPY 920 billion in subsidies from the Japanese government, private companies remain hesitant to invest due to the company’s lack of track record.

Atsuyoshi Koike added that, given the current situation, it is difficult for Rapidus to secure private financing. The company is discussing ways to make financing easier, such as implementing a government loan guarantee system.

Read more

• [News] Japanese Government Reportedly Plans to Offer Guaranteed Loans to Chip Startup Rapidus
• [News] Rapidus Partners RISC-V Player Esperanto, Targeting Low-Power AI Chip

(Photo credit: Rapidus)

According to a report from Reuters citing sources on August 7th, American chip giant Intel had an opportunity to invest in OpenAI several years ago but ultimately had the investment rejected by company executives, resulting in a missed opportunity.

Reportedly, Intel and OpenAI discussed collaboration several times between 2017 and 2018. At that time, OpenAI was still a nascent nonprofit research organization focused on developing relatively unknown generative AI technologies.

The discussions included Intel potentially purchasing a 15% stake in OpenAI for USD 1 billion in cash and possibly producing hardware for OpenAI at cost in exchange for an additional 15% stake.

Sources cited by the report further reveal that OpenAI was very interested in Intel’s investment, primarily because it would reduce the company’s reliance on NVIDIA chips and enable OpenAI to build its own infrastructure.

However, Intel ultimately rejected the deal. One reason cited by the report was that then-CEO Bob Swan did not believe generative AI could be commercialized in the short term and was concerned that Intel’s investment would not yield returns. Another reason was that Intel’s data center division was unwilling to produce hardware for OpenAI at cost.

After Intel’s refusal, Microsoft began investing in OpenAI in 2019. In 2022, OpenAI launched the chatbot ChatGPT, reportedly sparking a global AI boom and achieving a valuation of USD 80 billion. Per the data from CB Insights, it has made OpenAI the third most valuable tech startup worldwide, behind only ByteDance and SpaceX.

Neither Intel nor OpenAI has commented on these reports.

As per a previous report from The Atlantic, Intel had previously declined to produce processors for Apple’s iPhone, a misstep that caused Intel to miss the opportunity to transition into the mobile area.

The news from Reuters this time further suggests that Intel has made a similar mistake in the AI domain.

Read more

• [News] Intel’s 18A to Start Production in 2025, Featuring Processors for AI PCs and Servers
• [News] Intel’s Earnings Fall Short, Cutting Over 15% of Workforce

(Photo credit: Intel)

Intel not only reported earnings and forecasts that fell short of Wall Street expectations but also announced plans to cut more than 15% of its workforce, halt dividend payments for Q4 2024 (October-December), and reduce its full-year capital expenditure forecast by more than 20%.

According to Intel’s official announcement, its Q2 (April-June) earnings: adjusted earnings per share were $0.02, far below the analyst estimate of $0.10; revenue decreased by 1% year-over-year to USD 12.83 billion, missing the market expectation of USD 12.94 billion; and adjusted gross margin was 35.4%

During Q2, Intel’s Client Computing Group, responsible for producing PC processors, saw its revenue increase by 9% year-over-year to USD 7.41 billion, meeting the market expectation of USD 7.42 billion. However, the revenue from the Data Center and AI Group fell by 3% year-over-year to USD 3.05 billion, missing the market expectation of USD 3.14 billion.

Intel stated that sales of PC chips capable of handling AI tasks exceeded internalAI expectations, with shipments expected to surpass 40 million units in 2024.

Looking ahead to Q3, Intel forecasts revenue between USD 12.5 billion and USD 13.5 billion and an adjusted loss per share of $0.03. According to a report from Reuters citing an LSEG survey, analysts had originally predicted Q3 revenue to reach USD 14.35 billion with an adjusted earnings per share of $0.31. Intel’s adjusted gross margin for the quarter is expected to be 38%.

Intel CEO Pat Gelsinger stated that the latest layoff plan will affect about 15,000 employees. This is the largest single layoff action tracked by tech layoff monitoring site Layoffs.fyi since it began operations in March 2020. Intel currently employs around 110,000 people, meaning over 15% of its workforce will be impacted.

Gelsinger further pointed out that Intel must align its cost structure with the latest operational model and fundamentally change the way the company operates. He indicated that Intel’s revenue growth has not met expectations and has not yet benefited from powerful trends such as AI.

According to Intel’s statement, Intel will suspend dividend payments starting in Q4 until cash flow improves significantly. Since 1992, Intel has consistently paid dividends without interruption.

Intel has also decided to reduce its total capital expenditure budget for new plants and equipment in 2024 by over 20% to between USD 25 billion and USD 27 billion. The estimated total capital expenditure for 2025 will be between USD 20 billion and USD 23 billion.

Read more

• [News] Intel Hires Micron Executive to Lead Its Foundry Business
• [News] Intel Suspended Investment in French and Italian Chip Plants

(Photo credit: Intel)

Japanese NAND flash giant Kioxia announced today (August 1st) that the building construction of Fab2 (K2) of its Kitakami Plant in Iwate Prefecture was completed in July. As demand is recovering, the company will gradually make capital investments while closely monitoring flash memory market trends. Kioxia plans to start operation at K2 in the fall of Calendar Year 2025, according to its press release.

A portion of investment for K2 will be subsidized by the Japanese government according to the plan approved in February 2024, according to Kioxia.

In addition, the company notes that some administration and engineering departments will move into a new administration building located adjacent to K2 beginning in November 2024 to oversee the operation of K2.

According to a report from Nikkei on July 31, Kioxia’s Kitakami Plant started production in 2020,with the construction of K2 began in 2022. Initially, K2 was scheduled to commence production in 2023.

However, due to a downturn in the memory market and weak demand for NAND Flash used in smartphones and PCs, Kioxia started to reduce production in October 2022, with the extent of production cuts exceeding 30%. As part of these production reduction measures, Kioxia postponed the production start of the K2 facility.

Nikkei’s report further indicates that with market conditions recovering, Kioxia ended its production cuts in June 2024, and the current production line utilization rate has returned to 100%.

To mass-produce advanced memory products, Kioxia, in collaboration with Western Digital (WD), plans to invest a total of 729 billion yen in the Yokkaichi and Kitakami plants, with the Japanese government providing up to 243 billion yen in subsidies.

The Kitakami plant will produce the most advanced “8th generation” memory, with a monthly production capacity of 25,000 wafers. These will be used in AI data centers, as well as in smartphones, PCs, and automotive applications.

On June 26, according to industry sources cited in a report from Reuters, Kioxia plans to submit an initial public offering (IPO) application to the Tokyo Stock Exchange in the near future, aiming to go public by the end of October. Sources indicate that Kioxia will submit its official IPO application by the end of August, with a target listing date at the end of October.

Read more

• [News] NAND Flash Giant Kioxia Reportedly Plans IPO by Late October amid Market Recovery
• [News] Kioxia Optimistic about NAND Market Sentiment; Price Hike Expected in Q2

(Photo credit: Kioxia)