DD
MM
YYYY

PAGES

DD
MM
YYYY

spot_img

PAGES

Home Blog Page 845

White House’s AI and Crypto Chief Defends Nvidia’s China Chip Deal as Strategic Move to Undercut Huawei’s Rise

0

White House AI and crypto adviser David Sacks has come out in full defense of the Biden administration’s decision to allow Nvidia to resume sales of its H20 AI chips to China, arguing that the move is a tactical step to stop China’s tech giant Huawei from gaining unchallenged dominance in the world’s second-largest economy.

In an interview with Bloomberg on Tuesday, Sacks said the government’s shift in position “makes a lot of sense” given Huawei’s fast-rising capabilities.

“There is a compelling argument here that you just don’t want to hand Huawei the entire Chinese market when Nvidia is capable of competing for a big slice of it with a deprecated, less capable chip,” he said.

Nvidia’s H20 chip — designed specifically to comply with earlier U.S. restrictions — had its sales halted in April after the U.S. government said licenses would be required for any such transactions. But in a blog post on July 14, Nvidia confirmed that the U.S. had now given assurances that licenses will be granted and that the company hopes to begin deliveries soon.

The announcement followed a private meeting between Nvidia CEO Jensen Huang and President Donald Trump last week, where Nvidia reaffirmed its support for the administration’s goals of job creation, onshoring, and AI leadership. Commerce Secretary Howard Lutnick later said the green light to resume H20 sales was part of a wider negotiation involving rare-earth deals with China. Lutnick emphasized that the administration wants Chinese developers to remain tied to American technology by giving them “the fourth best” chip, not the most advanced models like the H100 or H200.

“We want to keep having the Chinese use the American technology stack, because they still rely upon it,” Lutnick said. “You want to sell the Chinese enough that their developers get addicted to the American technology stack.”

But Nvidia’s CEO, Jensen Huang, has not hidden his frustration with the export curbs. He has repeatedly warned that Washington’s clampdown is self-defeating and risks backfiring — both on U.S. companies and on America’s long-term tech leadership.

Speaking the Financial Times in March, Huang said that the restrictions hurt Nvidia’s bottom line, cost it significant market share, and undermine America’s chip supremacy. added that Huawei’s “presence in AI is growing every single year” and Nvidia “can’t assume they are not going to be a factor.”

He further called the U.S. strategy “poorly executed,” noting that restricting chip access was emboldening Beijing’s push for semiconductor independence. Huang added that Huawei, in particular, had become “the single most formidable technology company” in China.

“Their presence in AI is growing every single year, and we can’t assume they are not going to be a factor. They have conquered every market they’ve engaged,” Huang said.

Despite sweeping U.S. sanctions intended to hobble China’s access to advanced semiconductors, Huawei has emerged more resilient than expected. With massive support from the Chinese government — including funding, partnerships, and intellectual property protections — the tech giant has defied Western expectations and built cutting-edge chips that power everything from smartphones to AI data centers.

In 2023, Huawei shocked the industry by unveiling a 7-nanometer chip inside its Mate 60 Pro smartphone — a feat once thought to be impossible under U.S. restrictions. The move drew global attention and exposed gaps in the effectiveness of U.S. export bans.

U.S. intelligence officials have since warned that China’s semiconductor ambitions are gaining traction faster than predicted, partly because companies like Huawei are using sanctions as a rallying cry for tech self-reliance.

“China is maybe one and a half to two years behind us in chip design,” Sacks admitted, “but Huawei is moving fast to catch up. Even before they fully catch up, I think you will see them exporting their chips for the global market.”

Nvidia’s ability to sell to China is not just about quarterly earnings. China represents one of the largest markets for AI chips globally, with data centers, cloud computing, surveillance, and military sectors all demanding high-powered GPUs. U.S. companies dominate this segment, but with restrictions in place, they are now at risk of being squeezed out.

Without access to Chinese buyers, American chipmakers risk losing out to domestic rivals like Huawei, SMIC, and new players backed by Beijing’s multibillion-dollar semiconductor funds.

Dan Ives of Wedbush Securities echoed that view in a note following the H20 decision. “Nvidia resuming H20 chip sales in China is a gamechanger in our view. Trump knows there is one chip in the world fueling the AI revolution, and it’s Nvidia,” Ives wrote. “Giving the green light to Jensen/Nvidia is all part of negotiations with China. Nvidia gets $30 billion+ annual biz back.”

Following the news, Nvidia’s stock surged by over 5% on Tuesday, closing at a record high of $170.70. Investors interpreted the decision as a strategic win for Nvidia, ensuring it can reclaim lost ground in China’s rapidly growing AI sector while staying within the bounds of national security.

The move reflects Washington’s delicate balancing act — protecting U.S. innovation and security interests without triggering a full-scale decoupling that could alienate key markets and accelerate China’s tech independence.

“We’re not selling our latest greatest chips to China, but we can deprive Huawei of basically having this giant market share in China that they can then use to scale up and compete with us globally,” Sacks told Ludlow.

Fasmicro Opens Calls for Microprocessors & AI Training for African Governments

0

Our microelectronics business will be running some special workshops sponsored by governments at state and national levels across Africa. Yes, Fasmicro, Africa’s only programmable microprocessor knowledge partner for Intel Corp in Africa, will begin teaching the fusion of microprocessors and artificial intelligence (AI). Fasmicro is here to ensure companies, research institutions and broad customers understand how these systems will help in this fledgling AI era. We have the tools and knowledge capabilities to support Africa in the evolving amazing era of AI-driven opportunities.

While most have focused on the software aspect of AI, it is important to note that until microprocessors have evolved, software will have limitations. With more than a decade of partnership with Intel in programmable microprocessors, we will be introducing new systems into the African markets.

Visit intel.com and contact Fasmicro through the email provided https://www.intel.com/content/www/us/en/support/programmable/support-resources/fpga-training/overview.html . You must be a state or national government in Africa to qualify. We come with microchips, designkits, process systems, etc. We have just a few slots for this batch of training, and we look forward to support governments as they design and architect AI roadmap, pushing them to remember that AI doesn’t end with software; we are the hardware people, and we have provided engineering services across Africa with absolute quality.

Prof Ndubuisi Ekekwe

(PhD, Electrical & Computer Engineering / with focus on Microelectronics & Robotics Engineering, Johns Hopkins University)

Chairman, Fasmicro Microelectronics

Fasmicro recently celebrated 15 solid years of Intel Corp partnership

Server CPU Impact on Forex EA Performance – What Trading Session Data Reveals

0

During a particularly volatile Asian session, a trader running multiple grid EAs watched in horror as his execution times doubled. His MT4 platform showed all the usual metrics – stable internet, normal ping times, and steady broker connection. The culprit? CPU throttling on his supposedly “optimized” VPS.

This scenario highlights why selecting vps for forex traders requires understanding how processor architecture affects automated trading performance. The relationship between CPU cycles and EA execution isn’t linear – it’s exponential under certain market conditions.

CPU Threading Patterns During High-Impact News Events

When analyzing server performance during major economic releases, single-threaded CPU performance becomes critically important. MT4 and MT5 platforms primarily utilize single-core processing for trade execution, regardless of how many cores your VPS has available. During the recent Fed rate decisions, traders using multi-core servers but lower single-thread performance experienced significant delays.

Providers like NewYorkCityServers typically optimize for this with processors featuring higher base clock speeds, but many traders don’t realize why this matters more than total core count. A 4-core CPU running at 3.5GHz often outperforms an 8-core system at 2.8GHz for forex automation.

Memory-to-CPU Pipeline Optimization for Multiple EAs

Running concurrent expert advisors creates unique memory access patterns that can bottleneck even powerful servers. Each EA instance needs rapid access to both market data and its own operational memory space. The speed of this data transfer between RAM and CPU becomes a critical performance factor.

In testing environments, we’ve observed that EAs processing tick data can generate up to 2GB of memory transactions per hour per instance. Without proper memory channel configuration, this creates processing queues that delay trade execution. 

The Hidden Cost of CPU Power Management

Modern server processors include sophisticated power management features that can severely impact trading performance. These systems often take microseconds to ramp up from power-saving states – an eternity in forex trading terms. During recent volatility in EUR/USD pairs, traders experienced delays of up to 200ms due to processor power state transitions.

Proper configuration requires disabling various C-states and ensuring consistent CPU frequency. This maintains immediate response times but increases power consumption – a tradeoff many traders don’t consider when selecting budget VPS options.

Real-World Impact on Grid Trading Systems

Grid trading strategies are particularly sensitive to processor performance due to their concurrent order management requirements. Each grid level needs constant monitoring and rapid execution capability. A system running 20 grid levels across multiple currency pairs can generate over 100 processor interrupts per second during active market periods.

The processing overhead increases exponentially with each additional grid level or currency pair. Traders often discover this limitation only after scaling their strategies, leading to missed opportunities and inconsistent execution.

Network Interface Controller (NIC) CPU Offloading

One frequently overlooked aspect of server configuration is NIC processor offloading. Modern network interfaces can handle packet processing independently, reducing CPU load. However, many VPS providers don’t properly configure these features, forcing the main processor to handle network tasks that could be offloaded.

During high-frequency trading periods, this can consume up to 15% of available CPU cycles – resources that should be dedicated to EA execution and market analysis.

Session-Specific Performance Requirements

Different trading sessions demand varying levels of processor performance. Asian session algorithmic trading typically requires more consistent, sustained performance due to the nature of price movement. European and US sessions often need burst performance capability to handle sudden market shifts.

Configure processor performance profiles based on your primary trading sessions. European session traders might benefit from higher turbo boost frequencies, while Asian session strategies often perform better with steady base clock speeds.

Monitoring and Optimization Strategy

Implement continuous CPU performance monitoring focusing on three key metrics: process time, thread switching, and interrupt handling. These measurements provide early warning of potential execution problems before they impact trading performance.

NewYorkCityServers and similar quality providers typically offer monitoring tools, but traders should implement their own metrics tracking. Record execution times across different market conditions and correlate them with CPU performance data to optimize their setup.

The relationship between processor architecture and trading performance becomes more critical as strategies grow more sophisticated. Understanding these technical requirements helps traders build infrastructure that scales with their strategies rather than becoming a bottleneck to growth.

Remember: CPU performance isn’t just about raw speed – it’s about consistent, reliable execution under all market conditions. Choose and configure your trading infrastructure with this principle in mind.

New Exit Model and IP Dynamics: Scale AI, Windsurf, OpenAI/ Microsoft [podcast]

1

The video podcast discusses a “new exit model” emerging in the startup world, particularly in the context of Artificial Intelligence. Driven by the increasing value of “knowledge” (embodied in human talent and IP) and the desire to avoid regulatory scrutiny, large tech companies are opting to “dis-member” startups rather than acquire them outright.

Examples like Meta’s “reverse acquire-hiring” of Scale AI’s leadership and Google’s acquisition of key R&D teams (e.g., from Windsurf) illustrate this trend. These strategies allow big tech to gain crucial human capital and intellectual property without triggering antitrust concerns associated with full company acquisitions.

The lecture highlights that this approach, while legally permissible currently, may lead to the degradation of the “stripped” companies and could have long-term implications for market competition and innovation. The IP dynamics between OpenAI and Microsoft also underscore how intellectual property can be shared or accessed through strategic alliances, further diversifying the ways in which valuable assets are transferred in the tech landscape.

Lecture summary is available here.

OpenAI’s planned $3 billion acquisition of AI coding startup Windsurf collapsed due to tensions with Microsoft, OpenAI’s largest investor. Microsoft’s existing partnership with OpenAI entitled it to Windsurf’s intellectual property (IP), but OpenAI was reportedly unwilling to grant this access, creating a major sticking point in the deal.

The situation was further complicated by Windsurf’s reluctance to share its IP with Microsoft. The collapse of the deal created an opportunity for other companies:

Google stepped in to acquire Windsurf’s leadership team and licensed the company’s technology for $2.4 billion.

Cognition subsequently acquired the remaining assets of Windsurf, including its product, IP, and the majority of its employees.

The Windsurf deal highlights the intensifying competition in the AI sector for talent and technology.

The failed acquisition also exposed the complexities and potential limitations that can arise in partnerships between large tech companies and smaller startups in the rapidly evolving AI landscape.

In summary, OpenAI did not understand what it signed into. Yes, OpenAI is tethered to Microsoft and anything it gets belongs to Microsoft. Simply, whether the IP was created internally or acquired like the Windsurf failed deal, Microsoft is going to partake in the IP cake.

Of course, we can also learn a new exit model. Largely, Windsurf has been cannibalized without annoying the regulators. Google picked the things it liked, and the remaining parts have been absorbed by Cognition. And just like that, the exit happened and that is it!

The podcast video is at Blucera.com.

How To Listen to Tekedia Daily

At Blucera, home of Blucera WinGPT (AI personal educator and coach), eVault Legal Custodial services (store vital personal, family and business documents securely), business tools to grow enterprises, and global archives of Tekedia courses and libraries, Ndubuisi Ekekwe podcasts every week day. Some Tekedia Institute programs offer bonus access to Tekedia Daily or one can register at Blucera for the podcast.

Rwazi Raises $12M Series A to Revolutionize Global Consumer Insights with Real-Time AI-Powered Data

0

Rwazi, a data intelligence startup, has secured a $12 million Series A funding round led by Bonfire Ventures to expand its AI-powered platform that delivers real-time consumer insights across global markets.

The startup was born out of a stark realization, while regions like the U.S., UK, and parts of Western Europe had an abundance of consumer and market-level data, vast international markets including India, Brazil, Mexico, China, and Turkey lacked usable, real-time insights into consumer behavior. “There was no real picture on consumption, what consumers wanted, or how their behavior was shifting,” Rutakangwa told TechCrunch.

Initial attempts to rely on data from government trade agencies and consumer reports proved unreliable—often outdated, fragmented, or unverifiable. This led the founders to pioneer a new approach: zero-party data—consumption data voluntarily shared by consumers as part of their everyday routines. Using advanced validation and verification tools, Rwazi captures this data across global markets in real time.

The result is a powerful AI-driven intelligence system that allows companies to visualize consumer behavior live, predict market trends, and optimize decision-making. This has helped Rwazi’s clients among them global giants like Coca-Cola, Visa, Pampers, and Nestlé cut customer acquisition costs, boost loyalty, and stay competitive in fast-moving markets.

This new round of funding follows a $4 million seed round in 2023 also led by Bonfire Ventures, and includes participation from Santa Barbara Ventures, Newfund, and Alumni Ventures. Rutakangwa emphasized that the raise was “selective,” focused on investors who fully understood the challenges of building data infrastructure for underrepresented markets.

With presence in 190+ countries and clients primarily in the U.S. and Europe, Rwazi plans to use the funds to scale its AI co-pilot, expand engineering talent, and further refine its data infrastructure.

In a space dominated by legacy players like GFK and Ipsos, Rwazi distinguishes itself by offering direct, real-time, non-modeled insights. “Winning today means anticipating shifts, seeing around corners, and making confident moves before the competition even senses a change,” Rutakangwa said.

Founded in 2021 by Joseph Rutakangwa and Eric Sewankambo, Rwazi is redefining how global brands make decisions turning real-time, voluntarily shared consumer data into actionable insight at a global scale. The system captures how people consume and shop both online and offline, at home and away across 190+ countries.

The platform crowdsources data collection through an app, working with on-the-ground consumers who are paid for sharing details about the products they buy. This information comes in handy for global brands that leverage Rwazi’s customer dashboard to access actionable insights

The AI identifies what’s working, where demand is shifting, and how to respond—so teams move faster, waste less, and grow revenue with precision. Its AI processes millions of data points daily to help businesses fine-tune product-market fit, launch smarter campaigns, and move faster in an increasingly noisy and competitive market landscape.

Rwazi is playing big in the data analytics market, which is projected to reach $745.15 billion by 2033, with a CAGR of 10.3%. The increasing integration of AI and machine learning technologies is enhancing data analytics capabilities, allowing businesses to derive more insights from their data.

The rise of big data analytics is crucial, as organizations are generating massive amounts of data that require sophisticated analysis to drive decision-making. With its latest raise, Rwazi hopes to position itself as a market leader and capture a larger share of its expanding market.