华为准备在中国用新的人工智能芯片挑战Nvidia

币界网Published on 2024-08-13Last updated on 2024-08-14

币界网报道:

中国华为技术有限公司正准备推出Ascend 910C,这是一款新的人工智能芯片,将与英伟达在中国人工智能(AI)市场的领先地位相媲美。《华尔街日报》报道称,百度、字节跳动和中国移动等中国主要互联网和电信公司正在测试该芯片,预计将于10月首次亮相。

Ascend 910C的性能据说相当于Nvidia的H100,这是一种由于美国制裁而目前被禁止在中国销售的芯片。去年的制裁以国家安全为由,阻止了英伟达向中国客户出口其最先进的芯片。作为回应,英伟达为中国市场提供了其芯片的修改版本,但与H100相比,这些版本的性能有所降低。华为的这款新芯片有可能填补这一空白,为中国公司提供一个强有力的替代品。

华为用新的人工智能芯片对抗美国制裁

华为推出Ascend 910C之际,中国企业在与美国持续的贸易紧张局势中越来越倾向于国内替代品。如果成功,Ascend 910C可能会打破英伟达在中国使用的人工智能芯片市场的地位。《华尔街日报》称,该芯片的初始订单可能超过7万台,价值约20亿美元。这一需求反映出,面对美国的出口禁令,中国企业越来越渴望采用自己的技术。

该产品一直受到行业专家的密切关注,因为它代表了该公司让中国摆脱外国技术的又一次尝试。通过在中国建立自己作为英伟达的可行竞争对手的地位——两家公司在中国都有很大的业务——华为可以帮助塑造中国人工智能芯片的未来方向。

中国主要公司测试华为的人工智能芯片

字节跳动、百度和中国移动等中国科技巨头将在Ascend 910C上市后立即对其进行测试。这些公司都在人工智能研发方面投入了大量资金,这意味着他们需要强大而高效的芯片来运行他们的业务。试用华为的芯片表明,它可能满足通常与Nvidia的H100相关的高性能要求。

随着10月发布日期的临近,行业观察人士期待更多关于Ascend 910C的功能以及它与Nvidia产品的比较信息。如果这款芯片成功,它将为华为和中国更广泛的人工智能行业带来重大利益,因为它可以加快向本土技术的转变。

在接下来的几个月里,本地和全球市场可能会密切关注Ascend 910C在各种应用中的表现。如果有关其能力的报道属实,那么这可能会改变华为的地位,并在针对中国市场的人工智能芯片制造商之间创造新的竞争动态,其影响将远远超出华为和英伟达。

Trending Cryptos

Related Reads

Semiconductor Stock Rebound: Is the Technical Correction Over or a Trend Reversal?

The core of recent semiconductor stock volatility is not about daily price swings, but rather the market questioning whether AI-driven semiconductor pricing has entered a new phase. Following a sharp sell-off in Korean stocks on June 23rd, led by Samsung and SK Hynix, a subsequent rebound is seen more as a technical positioning adjustment rather than a confirmed trend reversal. The key variable is HBM (High Bandwidth Memory), essential for AI chips. Its supply-demand imbalance granted memory makers significant pricing power. The current market focus is on whether this dynamic remains strong enough to justify elevated valuations. All eyes are on Micron's upcoming earnings report. The critical factor is not whether results meet already high expectations, but whether the company's guidance confirms that AI memory pricing power, order visibility, and future margins are still expanding. Micron's outlook will serve as a crucial test for the broader AI semiconductor chain, including Samsung, SK Hynix, and other infrastructure players. The recent bounce appears to be a pre-earnings positioning repair. For it to evolve into a sustained uptrend, concrete evidence is needed that the AI infrastructure expansion cycle's fundamentals—particularly for high-end memory—remain robust and can continue to surpass elevated market expectations. The risk is that strong demand alone may not be sufficient if future guidance hints at peaking momentum or increasing supply-side pressures.

marsbit14m ago

Semiconductor Stock Rebound: Is the Technical Correction Over or a Trend Reversal?

marsbit14m ago

Global Tech Stocks Plunge: Another Stress Test for the AI Bull Market

Global tech stocks plummeted in a sharp selloff on June 23, with South Korea's KOSPI index crashing nearly 10%, triggering a trading halt. The plunge was led by semiconductor giants Samsung and SK Hynix, dragging down major Asian and US tech indexes and levered ETFs. The immediate trigger was a confluence of three signals within 24 hours: 1) reports that SK Hynix was slowing HBM4 expansion, raising doubts about a key AI chip bottleneck; 2) profit-taking ahead of Micron's earnings report after its massive year-to-date rally; and 3) a warning from Korean regulators about the risks of single-stock leveraged ETFs. The selloff's severity was amplified by Korea's uniquely leveraged market structure: record-high retail margin debt, over $30 billion in volatile single-stock leveraged ETFs requiring daily rebalancing (which creates selling pressure during declines), and the surprising shift of the National Pension Service from a net buyer to a net seller. The event reignited debates about an AI bubble. Analysts differed on timing—some warned of imminent rupture, others saw a temporary liquidity-driven correction—but agreed that sky-high valuations, a shift to retail/leveraged buying, and rising rates created a fragile setup. The parallel steep drop in SpaceX's stock, alongside its massive debt raise for AI infrastructure, underscored a broader market shift from narrative-driven "infinite imagination" to a "return on investment" calculation phase. All eyes are now on Micron's upcoming earnings report, seen as a key test for the AI hardware trade's fundamental logic. A strong report could stabilize markets, while a miss could validate deeper fears. The episode serves as a stark warning: when a bull run becomes dependent on leveraged bets on a single narrative, a violent and rapid correction should not be a surprise. The core question for investors is how much drawdown they are willing to tolerate to stay invested.

marsbit16m ago

Global Tech Stocks Plunge: Another Stress Test for the AI Bull Market

marsbit16m ago

Ethereum Foundation Cuts 20% of Staff, 54 Depart: The Survival Logic Behind the Restructuring

The Ethereum Foundation (EF) has announced a major restructuring, resulting in the departure of 54 staff members, representing approximately 20% of its workforce. This reorganization is not merely a cost-cutting measure but a strategic refocusing. The EF will now concentrate its resources on what it deems critical and unique tasks, structured around five new operational clusters. The new structure comprises clusters dedicated to: the Protocol Layer (ensuring Ethereum's core properties like censorship resistance and security); the Access Layer (enabling trusted, non-intermediated user interactions); the User Layer (grounding decisions in real user needs and constraints); the Community Layer (representing EF's stance and building alliances); and the Institutional Layer (engaging with enterprises, governments, and academia to promote principled adoption). The foundation stated that the layoffs were a difficult but necessary step to align its organization and spending with its long-term mandate, insulating its core work from short-term market fluctuations. Affected employees were offered a severance package and transition support. The restructured EF emerges as a leaner and more focused organization, poised to prioritize the development and preservation of Ethereum's foundational promise of self-sovereignty. Further details on the new operational model are expected in the coming weeks.

marsbit19m ago

Ethereum Foundation Cuts 20% of Staff, 54 Depart: The Survival Logic Behind the Restructuring

marsbit19m ago

Talking About Returns But Not Collection? Goldfinch's Liquidation Sounds the Alarm for RWA Credit

Goldfinch, a crypto lending platform connecting investors with real-world borrowers, has proposed a full wind-down via governance proposal GIP-87. The plan would halt all new development, shut down its flagship Goldfinch Prime product, and allocate 150,000 USDC to manage the collection of outstanding loans. While the proposal is under community vote, it highlights a critical shift for the RWA (Real World Asset) lending sector: the transition from a growth phase focused on yields to a difficult recovery phase focused on collections. The proposal reveals that while the protocol's on-chain TVL is low, it still holds tens of millions in active, non-performing loans off-chain. This gap underscores that tokenizing debt makes tracking exposure transparent but does not simplify the offline, labor-intensive, and legally complex process of loan recovery. The case of the Lend East pool, where only an estimated 42% of a $10.15 million loan may be recovered, exemplifies the potential for significant investor losses. The wind-down plan forces token holders to govern not expansion but the maintenance of a debt collection system, including funding legal trust structures and preserving user access for repayments. This move starkly contrasts with the sector's typical narrative of rapid, AI-powered underwriting and high yields, exposing the often-overlooked necessity for robust borrower vetting, standardized disclosure, and sustainable collection mechanisms. Ultimately, Goldfinch's situation serves as a crucial stress test for the entire RWA lending space. It demonstrates that a platform's true resilience is tested not during capital deployment but during the protracted, costly, and uncertain process of recovering defaulted loans from real-world borrowers, a challenge blockchain transparency alone cannot solve.

Foresight News24m ago

Talking About Returns But Not Collection? Goldfinch's Liquidation Sounds the Alarm for RWA Credit

Foresight News24m ago

Meta Launches Prediction Market, Code-Named "Arena": Not Using Real Money, 3.56 Billion Daily Active Users Are Its Biggest Bargaining Chip

Meta, under the codename "Arena," is developing a standalone prediction market application, according to a June 23 report by The New York Times. Initially, the app will operate on a points-based system rather than real-money betting, though future integration of financial transactions is not ruled out. Meta plans to leverage its vast ecosystem of apps, boasting 3.56 billion daily active users, to distribute the product. The strategy aims to lower user acquisition costs and navigate regulatory complexities associated with real-money prediction markets, which are overseen by bodies like the CFTC. News of Meta's entry caused stock dips for established players like DraftKings and Robinhood, reflecting market concerns over Meta's potential to disrupt incumbents like Polymarket and Kalshi with its massive scale. The prediction market sector has seen explosive growth, with monthly trading volume on major platforms surging to approximately $24 billion as of April 2026. This marks Meta's second foray into the space, having launched and later shut down a similar virtual-points app called Forecast in 2020. While Arena poses a competitive threat to crypto-based platforms like Polymarket, its mainstream reach could also significantly expand the overall user base and awareness for prediction markets. The project remains in development with no public launch timeline, but its announcement has already impacted market sentiment.

marsbit29m ago

Meta Launches Prediction Market, Code-Named "Arena": Not Using Real Money, 3.56 Billion Daily Active Users Are Its Biggest Bargaining Chip

marsbit29m ago

Trading

Spot
Futures

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.

活动图片