In the past, whenever NVIDIA spoke about AI with a sense of triumph, praised GPUs for their dominance, or occasionally highlighted Intel’s more provocative stance, Intel remained relatively tolerant and low-key. Facing the failure of Moore's Law and a slow transition in the broader tech landscape, Intel was less vocal in other areas as well. However, at a press conference in September, Yang Xu, President of Intel China, unexpectedly told the media, "The tiger doesn’t speak, thinking it is a sick cat." This statement reflected a shift in Intel’s approach to AI.
With the implementation of its AI investment strategy and the integration of services, Intel has become more vocal in recent times. From the headquarters to top-level executives in China, high-ranking leaders have taken the front line to actively promote Intel’s AI vision and insights. At the same time, Intel has been actively participating in various market activities to drive the adoption of artificial intelligence across industries.
Over the past two months, Intel has made some extraordinary moves that have caught the attention of the industry. In mid-October, Intel CEO Brian Krzanich announced the launch of Nervana’s neural network processor by year-end. Later, Intel revealed a partnership with long-time rival AMD, followed by Intel “poaching†Raja Koduri, AMD’s chief GPU architect, with plans to develop a dedicated GPU.
Wall Street has also turned its focus toward Intel. In mid-November, an investment firm on Wall Street stated that investors should not overlook Intel’s AI capabilities, expecting the company to grow in multiple long-term markets and raising Intel’s target price.
If Intel’s AI strengths finally take off, what impact could this have on the future AI chip market?
Ten years ago, if Intel had acquired ATI instead of AMD, things might have looked very different today. It’s certain that NVIDIA wouldn’t be in the position it is now, and Intel might not be seen as lagging in AI. Back then, ATI and NVIDIA were both GPU vendors, while AMD and Intel competed on the CPU side. In 2006, AMD acquired ATI for $5.4 billion, becoming a chip company with both CPU and GPU capabilities.
AMD’s M&A strategy at the time was forward-thinking, but the acquisition came at a high cost. With only $3 billion in cash before the deal, AMD had to take on significant debt, which made it harder to compete against stronger rivals in both CPU and GPU markets.
NVIDIA’s rise in the GPU space later benefited from this acquisition, giving them more room to grow. As AI emerged in recent years, NVIDIA, already a leader in GPUs and graphics computing, capitalized on the trend. Although GPUs weren’t designed specifically for AI, their strength in graph-related calculations gave them a natural advantage, helping NVIDIA ride the AI wave and soar in stock value.
Intel and AMD clearly don’t want NVIDIA to maintain its dominant position on the GPU. In early November, it was reported that Intel and AMD would collaborate in the next century, with AMD’s independent GPU being integrated into Intel SoC. This move recalls the old rivalry between the two companies, proving once again that there are no eternal enemies—only eternal interests.
However, AMD’s chief GPU architect, Raja Koduri, left shortly after, joining Intel as a senior vice president in charge of the newly established “Core and Visual Computing†group, tasked with developing Intel’s GPU business.
Intel wants to develop its own GPU! This is certainly bad news for NVIDIA and AMD. Over the past decade, AMD and Intel have competed in the CPU space, while AMD and NVIDIA have battled in the GPU arena. Now that Intel is entering the GPU market, it will not only pressure NVIDIA but also put AMD in a more vulnerable position.
Imagine if Intel joins the GPU competition. The market would then have three major players: NVIDIA, AMD, and Intel. How would the future landscape change?
Previously, although Intel, Google, and Apple were developing their own AI chips, they didn’t make much of an impact. NVIDIA’s GPUs became the standard for large-scale AI applications, leading to its popularity on Wall Street and a sharp increase in stock price. Intel, on the other hand, was seen as slower due to its delayed introduction of proprietary AI products, causing its stock to stagnate.
On October 18th, Intel CEO Brian Krzanich announced at the WSJDLive Global Technology Conference that the Intel Nervana Neural Network Processor (NNP) would ship before the end of the year. This marks Intel’s first industry-oriented neural network chip. Though Nervana hasn’t officially launched yet, the announcement caused Intel’s stock to rise sharply, showing strong market expectations.
Nervana originated from the AI startup Nervana Systems, which Intel acquired in 2016. At the time, it was considered the first company to create chips specifically for AI. After the acquisition, founder Raveen Rao became the general manager of Intel’s AI solutions division.
Intel placed high hopes on Nervana, aiming to reclaim market share lost to GPU competition. According to Intel, the Nervana NNP offers higher performance and scalability for AI models, promising a 100x improvement in AI performance by 2020.
Facebook is already involved in the development of Nervana, but the real test will come when the chip hits the market.
In addition to Nervana, Intel also has FPGAs. Intel acquired FPGA manufacturer Altera for $16.7 billion last year, marking the largest acquisition in its history. Altera’s FPGAs are now integrated into Intel’s processors as a GPU-like acceleration technology, creating a CPU+FPGA alternative to GPUs.
For example, the U.S. Group uses Intel’s deep learning platform, including an FPGA cloud host; Microsoft also adopted the Intel Stratix 10 FPGA for its Project Brainwave.
Some experts believe that while GPUs have advantages over CPUs, FPGAs are not inferior. Deep learning algorithms can run faster and more efficiently on FPGAs, with lower power consumption.
At NVIDIA’s GTC 2017, the company announced that the number of participants had tripled in five years, and GPU developers had increased 11-fold to over 500,000.
NVIDIA executives often highlight the growth of GPUs and the ecosystem around them in public settings. Unlike NVIDIA’s heavy reliance on GPUs for AI, Intel emphasizes a full-stack solution, combining CPUs, FPGAs, and Nervana.
Intel has invested heavily in AI, both in hardware and software. Beyond Nervana and Altera, it has acquired companies like Movidius and Itseez. In March, Intel established its AI division and spent $15 billion on Mobileye, along with a $1 billion AI innovation fund. In October, it invested in 15 AI startups, including China’s Horizon Group.
At the 2017 Intel AI Conference, Fiaz Mohamed, general manager of business development at Intel’s AI unit, said that unlike NVIDIA’s GPU-centric AI strategy, Intel’s strategic breadth and depth are unmatched.
Fiaz believes that AI is complex and requires not just powerful computing, but also strong networking, advanced storage, and other capabilities to support AI applications.
Intel’s strategy isn’t just focused on deep learning; it aims to build a full AI ecosystem. From end-to-end investments to hardware and software, Intel’s goal is not just to beat NVIDIA, but to target the entire AI market.
Moreover, Intel’s future investments extend beyond AI, covering autonomous driving, IoT, and 5G. After losing ground in mobile chips, Intel is targeting high-growth markets to avoid missing out on the next big opportunity.
In the past few decades, in the CPU wars, whenever AMD introduced a groundbreaking product, it was eventually overshadowed by Intel. Intel’s financial and technological strength made it difficult for AMD to compete. But in the current AI market, NVIDIA leads with GPUs. However, AI is still in its early stages, and both Intel and other tech giants like Google and Apple are eyeing this growing market.
If Intel’s full AI capabilities kick in with the release of Nervana’s neural network chips, will NVIDIA continue to dominate AI, or will Intel challenge its lead? After all, the chip industry is both tech and capital-intensive. While NVIDIA holds a leading position in GPUs, Intel has greater overall strength and financial resources. With over $20 billion in cash, Intel remains a formidable competitor.
Analysts on Wall Street are concerned that Intel’s entry into AI could threaten NVIDIA’s gross margin sustainability. As a fabless chipmaker, NVIDIA relies on TSMC for manufacturing. In contrast, Intel’s vertical integration allows for more flexible pricing strategies. If Intel competes on price, it could hurt NVIDIA’s margins.
Another concern is the threat from GPU alternatives. Currently, NVIDIA’s dominance in AI is due to the lack of strong alternatives. However, with Intel’s AI chips or Google’s TPUs entering the market, NVIDIA faces potential competition.
Despite Intel’s superior strength, NVIDIA is also a seasoned chip giant with a long history of competing with Intel. It is unlikely to be easily defeated.
Lost Mary OS5000puffs
YIWU JUHE TRADING COMPANY , https://www.nx-vapes.com