Editorial illustration for ByteDance taps Samsung for custom AI chips and scarce memory supply
ByteDance Develops Custom AI Chips with Samsung
ByteDance taps Samsung for custom AI chips and scarce memory supply
ByteDance is telling everyone it isn't building its own AI chips. It's also signing a major deal with Samsung to build them. The company behind TikTok is doing what every big tech firm does when it's behind: it's trying to buy its way forward.
Reuters reported the partnership, which covers both custom chip production and, crucially, access to Samsung's memory supplies. Memory is the choke point for AI right now, arguably more valuable than the processors themselves. Securing a pipeline for it is a bigger win than the silicon blueprint. The first test chips are expected by the end of this month.
Production is planned to start at 100,000 units this year, potentially scaling to 350,000. The financial commitment is staggering. ByteDance intends to spend over $22 billion on AI procurement in 2026. More than half of that is earmarked for Nvidia's latest H200 chips and for developing its own semiconductor designs.
What makes the deal especially interesting: the talks also cover access to memory chip supplies, which are extremely scarce amid the global AI infrastructure buildout - making the arrangement particularly valuable for Bytedance. The company plans to receive its first sample chips by the end of March and produce at least 100,000 units this year, with a possible ramp-up to 350,000. Bytedance intends to spend more than 160 billion yuan (roughly $22 billion) on AI-related procurement in 2026 - more than half of that going toward Nvidia chips, including H200 models, and development of its own chip.
Bytedance executive Zhao Qi acknowledged during an internal meeting in January that the company's AI models still trail global leaders like OpenAI, but pledged continued support for AI development. Bytedance itself denies the chip project - a spokesperson told Reuters the information was inaccurate without providing further details.
The denial is standard. It means nothing. When a company has specific production targets and a $11 billion budget line for its own chips, the project is real.
The public stance is just noise. The internal admission from executive Zhao Qi is more revealing. He said their models lag behind leaders like OpenAI.
This chip play is an attempt to fix the hardware deficit, because catching up on the software side alone is a brutal, uncertain fight.
They are betting billions that owning the stack, from memory to custom silicon, will give them an edge. It's a long shot. But in the current scramble, securing any part of the supply chain is a win. Their ambition is laid bare in the numbers, not the statements.
Related Reading
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic Google, Samsung unveil Gemini task automation for food delivery and rideshare
AI Tools & Apps
Related Topic Samsung adds Vision AI Companion, an AI Bixby, to TVs for real-time queries
LLMs & Generative AI
From Archives xAI co-founder departures persist as Lambda outlines 2025 AI production shifts
From Archives AI Rivals Launch Joint Accelerator for 20 European Startups per Cohort
From Archives Google's Smart Glasses Lead in Software, Yet Lag on Style, Partner Samsung
From Archives Samsung unveils Bixby-enabled fridge that can open and shut doors
Open Source
Common Questions Answered
How many AI inference chips does ByteDance plan to manufacture in 2026?
ByteDance aims to produce at least 100,000 AI inference chips this year, with plans to gradually scale up production to potentially 350,000 units. The company is working with Samsung Electronics to develop and manufacture these custom chips, with the first sample chips expected by the end of March.
Why is ByteDance negotiating with Samsung beyond chip manufacturing?
ByteDance is not only seeking chip manufacturing capabilities but also looking to secure scarce memory chip supplies amid the global AI infrastructure expansion. The tight demand for memory chips in the AI sector makes this aspect of the negotiations particularly valuable for the company's long-term AI development strategy.
When did ByteDance begin its efforts to develop in-house AI chips?
ByteDance initiated its multi-year effort to build in-house chips for AI workloads in 2022, starting by hiring chip-related talent. In 2024, the company collaborated with Broadcom to develop an advanced AI processor that would be compliant with U.S. export restrictions.
Further Reading
- ByteDance developing AI chip, in manufacturing talks with Samsung, sources say — Reuters via MarketScreener
- ByteDance turns to Samsung for custom AI chip production and scarce memory supplies — The Decoder
- China's ByteDance In Talks With Samsung To Manufacture AI Chips, Secure Scarce Memory Chip Supplies: Report — Finviz
- ByteDance Eyes Homegrown AI Chip Deal With Samsung Amid Global Tech Rivalry — Sri Lanka Guardian
