The post Nvidia builds new China chip to skirt US limits appeared on BitcoinEthereumNews.com. Nvidia is building a new chip for the Chinese market that will be more powerful than the H20, and it’s doing it using its most advanced chip architecture; Blackwell. This comes at a time when the company is trying to keep its foothold in China without crossing lines drawn by U.S. regulators. The chip is unofficially called the B30A, and while it won’t match the raw power of Nvidia’s flagship B300, it’s designed to perform significantly better than the limited H20 China is currently stuck with. The B30A uses a single-die design, unlike the dual-die setup used in the full-powered B300. A single-die means all the chip’s core components sit on one continuous piece of silicon. This makes the chip easier to produce under U.S. restrictions, while still packing enough power to make it useful for Chinese developers. Nvidia is planning to ship test units to clients in China as early as next month, though no final specs have been released yet. Trump opens door for scaled-down chip sales Donald Trump, now back in the White House, recently said he might allow Nvidia to sell toned-down versions of its newer chips to Chinese buyers. He made the comment after announcing a deal that would force Nvidia and AMD to hand over 15% of revenue from certain chip sales in China to the U.S. government. Trump also said the new chips could come with “30% to 50% off” in computing power, calling the current H20 “obsolete.” Despite Trump’s remarks, U.S. approval for new AI chip sales isn’t guaranteed. Lawmakers on both sides of Congress have been openly worried that even weaker versions of Nvidia’s top chips could still give China too much of a leg up in AI development. They’ve been pushing back on any move that might undermine U.S. control… The post Nvidia builds new China chip to skirt US limits appeared on BitcoinEthereumNews.com. Nvidia is building a new chip for the Chinese market that will be more powerful than the H20, and it’s doing it using its most advanced chip architecture; Blackwell. This comes at a time when the company is trying to keep its foothold in China without crossing lines drawn by U.S. regulators. The chip is unofficially called the B30A, and while it won’t match the raw power of Nvidia’s flagship B300, it’s designed to perform significantly better than the limited H20 China is currently stuck with. The B30A uses a single-die design, unlike the dual-die setup used in the full-powered B300. A single-die means all the chip’s core components sit on one continuous piece of silicon. This makes the chip easier to produce under U.S. restrictions, while still packing enough power to make it useful for Chinese developers. Nvidia is planning to ship test units to clients in China as early as next month, though no final specs have been released yet. Trump opens door for scaled-down chip sales Donald Trump, now back in the White House, recently said he might allow Nvidia to sell toned-down versions of its newer chips to Chinese buyers. He made the comment after announcing a deal that would force Nvidia and AMD to hand over 15% of revenue from certain chip sales in China to the U.S. government. Trump also said the new chips could come with “30% to 50% off” in computing power, calling the current H20 “obsolete.” Despite Trump’s remarks, U.S. approval for new AI chip sales isn’t guaranteed. Lawmakers on both sides of Congress have been openly worried that even weaker versions of Nvidia’s top chips could still give China too much of a leg up in AI development. They’ve been pushing back on any move that might undermine U.S. control…

Nvidia builds new China chip to skirt US limits

Nvidia is building a new chip for the Chinese market that will be more powerful than the H20, and it’s doing it using its most advanced chip architecture; Blackwell.

This comes at a time when the company is trying to keep its foothold in China without crossing lines drawn by U.S. regulators.

The chip is unofficially called the B30A, and while it won’t match the raw power of Nvidia’s flagship B300, it’s designed to perform significantly better than the limited H20 China is currently stuck with.

The B30A uses a single-die design, unlike the dual-die setup used in the full-powered B300. A single-die means all the chip’s core components sit on one continuous piece of silicon.

This makes the chip easier to produce under U.S. restrictions, while still packing enough power to make it useful for Chinese developers. Nvidia is planning to ship test units to clients in China as early as next month, though no final specs have been released yet.

Trump opens door for scaled-down chip sales

Donald Trump, now back in the White House, recently said he might allow Nvidia to sell toned-down versions of its newer chips to Chinese buyers. He made the comment after announcing a deal that would force Nvidia and AMD to hand over 15% of revenue from certain chip sales in China to the U.S. government. Trump also said the new chips could come with “30% to 50% off” in computing power, calling the current H20 “obsolete.”

Despite Trump’s remarks, U.S. approval for new AI chip sales isn’t guaranteed. Lawmakers on both sides of Congress have been openly worried that even weaker versions of Nvidia’s top chips could still give China too much of a leg up in AI development. They’ve been pushing back on any move that might undermine U.S. control over next-gen chip tech.

For Nvidia, China is a revenue stream it can’t afford to lose. In the last financial year, 13% of its total income came from China. The H20 chip, which was created specifically to get around export restrictions, was banned in April before being cleared for sale again in July. But with ongoing regulatory uncertainty, Nvidia is now hedging its bets with the B30A and a second chip built for a different task.

New RTX6000D will launch in September

Alongside the B30A, Nvidia is preparing to deliver another chip to China. This one is built for AI inference tasks and is called the RTX6000D. It’s also based on the Blackwell design, but it’s cheaper to produce and weaker in specs.

This chip uses standard GDDR memory and is being deliberately built to stay under Washington’s red line. Its memory bandwidth hits 1,398 gigabytes per second, just short of the 1.4 terabyte threshold that triggered the H20 ban.

Nvidia is planning to send small batches of RTX6000D to Chinese clients in September, according to Reuters. The goal is to maintain interest from Chinese developers without breaching export limits.

But that’s proving harder as local chipmakers like Huawei begin catching up. Huawei’s latest AI chips reportedly match Nvidia’s in compute strength, though analysts still say they trail in software support and memory bandwidth.

Nvidia has argued it’s better to give China limited access to its hardware than to risk companies switching entirely to Huawei. That strategy is now under pressure. Over the past few weeks, Chinese state media has accused Nvidia chips of being security threats, and government officials have warned companies not to buy the H20. Nvidia has responded by saying its chips have no backdoors.

As for Nvidia’s public stance, the company said in a statement: “We evaluate a variety of products for our roadmap, so that we can be prepared to compete to the extent that governments allow.” They added: “Everything we offer is with the full approval of the applicable authorities and designed solely for beneficial commercial use.”

Despite the noise, Nvidia’s only real option is to keep building whatever the U.S. government will let them sell, and hope China keeps buying.

KEY Difference Wire helps crypto brands break through and dominate headlines fast

Source: https://www.cryptopolitan.com/nvidia-building-new-china-ai-chip-blackwell/

Market Opportunity
Threshold Logo
Threshold Price(T)
$0.010224
$0.010224$0.010224
+2.44%
USD
Threshold (T) Live Price Chart
Disclaimer: The articles reposted on this site are sourced from public platforms and are provided for informational purposes only. They do not necessarily reflect the views of MEXC. All rights remain with the original authors. If you believe any content infringes on third-party rights, please contact service@support.mexc.com for removal. MEXC makes no guarantees regarding the accuracy, completeness, or timeliness of the content and is not responsible for any actions taken based on the information provided. The content does not constitute financial, legal, or other professional advice, nor should it be considered a recommendation or endorsement by MEXC.

You May Also Like

WLFI Bank Charter Faces Urgent Halt as Warren Exposes Trump’s Alarming Conflict of Interest

WLFI Bank Charter Faces Urgent Halt as Warren Exposes Trump’s Alarming Conflict of Interest

BitcoinWorld WLFI Bank Charter Faces Urgent Halt as Warren Exposes Trump’s Alarming Conflict of Interest WASHINGTON, D.C. – March 15, 2025 – In a dramatic escalation
Share
bitcoinworld2026/01/14 06:40
UNI Price Prediction: Targets $5.85-$6.29 by Late January 2026

UNI Price Prediction: Targets $5.85-$6.29 by Late January 2026

The post UNI Price Prediction: Targets $5.85-$6.29 by Late January 2026 appeared on BitcoinEthereumNews.com. Rebeca Moen Jan 13, 2026 13:37 UNI Price Prediction
Share
BitcoinEthereumNews2026/01/14 05:50
Taiko Makes Chainlink Data Streams Its Official Oracle

Taiko Makes Chainlink Data Streams Its Official Oracle

The post Taiko Makes Chainlink Data Streams Its Official Oracle appeared on BitcoinEthereumNews.com. Key Notes Taiko has officially integrated Chainlink Data Streams for its Layer 2 network. The integration provides developers with high-speed market data to build advanced DeFi applications. The move aims to improve security and attract institutional adoption by using Chainlink’s established infrastructure. Taiko, an Ethereum-based ETH $4 514 24h volatility: 0.4% Market cap: $545.57 B Vol. 24h: $28.23 B Layer 2 rollup, has announced the integration of Chainlink LINK $23.26 24h volatility: 1.7% Market cap: $15.75 B Vol. 24h: $787.15 M Data Streams. The development comes as the underlying Ethereum network continues to see significant on-chain activity, including large sales from ETH whales. The partnership establishes Chainlink as the official oracle infrastructure for the network. It is designed to provide developers on the Taiko platform with reliable and high-speed market data, essential for building a wide range of decentralized finance (DeFi) applications, from complex derivatives platforms to more niche projects involving unique token governance models. According to the project’s official announcement on Sept. 17, the integration enables the creation of more advanced on-chain products that require high-quality, tamper-proof data to function securely. Taiko operates as a “based rollup,” which means it leverages Ethereum validators for transaction sequencing for strong decentralization. Boosting DeFi and Institutional Interest Oracles are fundamental services in the blockchain industry. They act as secure bridges that feed external, off-chain information to on-chain smart contracts. DeFi protocols, in particular, rely on oracles for accurate, real-time price feeds. Taiko leadership stated that using Chainlink’s infrastructure aligns with its goals. The team hopes the partnership will help attract institutional crypto investment and support the development of real-world applications, a goal that aligns with Chainlink’s broader mission to bring global data on-chain. Integrating real-world economic information is part of a broader industry trend. Just last week, Chainlink partnered with the Sei…
Share
BitcoinEthereumNews2025/09/18 03:34