Taiwan's TSMC Considers Investing in Korean Fabless 'Furiosa AI'

Taiwan's TSMC, the world's No. 1 foundry (semiconductor consignment manufacturing) company, is considering investing in domestic artificial intelligence (AI) semiconductor startup Furiosa AI.



Furiosa AI is also currently in acquisition negotiations with Meta Platform.



TSMC logo [Photo = TSMC]



According to the IT industry on the 27th, TSMC Global, an investment subsidiary of TSMC, has been considering a plan to make a strategic investment in Furiosa AI since the fourth quarter of last year.



In relation to this, Furiosa AI stated, "Nothing has been confirmed."



If Meta's acquisition of Furiosa AI and TSMC's investment are successful, there is an outlook that there will be changes to NVIDIA's monopoly in the AI ​​semiconductor market.



Taiwanese IT media TrendForce explained, "If Meta acquires Furiosa AI, it could be an opportunity for TSMC," adding, "Meta is aiming to strengthen its fabless capabilities and reduce its dependence on NVIDIA, which could potentially expand TSMC's customer base."



If TSMC's 'de-Nvidia' gains momentum, it is assessed that it could affect the domestic semiconductor industry such as Samsung Electronics and SK Hynix.



Furiosa AI is a fabless (semiconductor design company) startup that develops AI inference specialized semiconductors for data center servers. It has developed the cost-effective neural network processing unit (NPU) Warboy and Renegade compared to Nvidia's AI semiconductors.



Renegade, equipped with SK Hynix's high-bandwidth memory HBM3, is produced through TSMC's 5nm process, while the first-generation Warboy used Samsung's 14nm process. Renegade shows similar performance to L40S, which is considered the next-level AI chip for inference after Nvidia's H100, but its power consumption is 150W, which is more than twice as efficient as L40S's 350W.



TSMC's strategic investment review is analyzed to be the result of considering the technological prowess and potential of Furiosa AI, which produced the Furiosa AI chip.



Meta is currently negotiating to acquire FuriosaAI while reducing its dependence on expensive Nvidia chips and developing semiconductors optimized for its own large language model (LLM) performance. The results are expected to come out within the next month.



Meta has announced a goal of investing up to $65 billion (approximately 93 trillion won) in AI infrastructure this year and deploying 1.3 million graphic processing units (GPUs), and it seems that the acquisition of FuriosaAI is intended to reduce chip purchase and data center operating costs and secure AI semiconductor design personnel.









https://www.inews24.com/view/blogger/1818591

Comments

Post a Comment

Popular posts from this blog

Livestock Manure Methane Is Soaring, But ‘Resource Recovery’ Isn’t Working [Now is a Climate Crisis]

Image
It has been pointed out that the methane emissions from the livestock waste treatment process were actually reported to be 2.5 times smaller than the previous calculation standards. It is expected that livestock waste will increase rapidly in the future, but an analysis report has been released that 'biogasification', which turns it into a resource, is only at the 10% level. It is urgent to come up with countermeasures as methane, one of the greenhouse gases, is expected to increase rapidly. According to the analysis of the entire livestock waste treatment process for the first time in Korea, the largest source of methane emissions was pigs (73%), and only 14% of pig waste was used for biogas in the public sector. It has been pointed out that it is urgent to come up with countermeasures as methane from livestock waste is increasing rapidly but 'resource conversion' is insufficient. [Photo = Climate Solution] About half of domestic methane...
Read more

KT-MS, 'Super Cooperation' to Improve Korea's AI Capabilities... "Providing AI Education to All Citizens" (Comprehensive)

Image
KT and Microsoft (MS) are going beyond corporate partnerships to engage in hyper-collaboration to enhance Korea's artificial intelligence (A)I capabilities. The plan is to provide AI education to all citizens interested in AI and contribute to improving the level of "national AI literacy" (the ability to understand and utilize AI technology). KT CEO Kim Young-seop (left) and Microsoft CEO and Chairman of the Board Satya Nadella pose for a commemorative photo after meeting at KT's Gwanghwamun office on the 25th. [Photo = KT] First official meeting of the two companies' CEOs in Korea···Discussion of plans to promote AX in Korea On the 26th, KT announced that KT CEO Kim Young-seop and Microsoft CEO and Chairman of the Board Satya Nadella met and exchanged opinions on accelerating the industry's AX (AI transition). The explanation is that they discussed agenda items that will help strengthen the country's AI competitivenes...
Read more

"TSMC to have 2nm line with monthly production capacity of 50,000 sheets by year-end"

Image
There is a report that TSMC, the world's largest foundry, is expected to have a 2nm process line capable of producing around 50,000 sheets per month by the end of this year. TSMC [Photo = TSMC] Taiwanese economic media Commercial Times reported on the 24th (local time) that TSMC's 2nm process production capacity could reach 50,000 sheets per month by the end of the year, and could increase to 80,000 sheets once the Baoshan and Kaohsiung factories are fully operational. TSMC's Baoshan factory is scheduled to receive its first order for the second half of this year from April 1, and its Kaohsiung factory in southern Taiwan is also reported to hold an expansion ceremony on the 31st. Kuo Ming-chi, an analyst at TF International, claimed via X (formerly Twitter) that "TSMC's 2nm trial production yield had already reached 60-70% three months ago, and is now well above that." TSMC has already secured an overwhelming share of th...
Read more