慧與科技（Hewlett Packard Enterprise）的CEO安東尼奧·內里稱，中國的科技企業要成為美國公司真正的威脅，或者要成為全球科技領袖企業，還有很長的路要走。

“問題在于，在制造業之外，到底有多少創新？”本周接受《財富》雜志采訪時內里說道，“我認為，中國在科技上總體落后一到兩個檔次。”

慧與科技2015年從惠普拆分出來，主營業務是向企業用戶提供服務器、存儲和軟件等產品和服務。內里1995年加入惠普呼叫中心，他成為慧與科技掌門人已有兩個年頭。他對于競爭對手做了評估，之后作出了此番評論。在云計算領域，慧與面臨亞馬遜和谷歌的競爭，但內里認為這些公司主要是數據存儲提供商，而慧與的切入點是更綜合性的“邊緣計算”服務。

邊緣計算背后的理念其實非常簡單直接。目前，許多分析和人工智能科技，比如語音識別，是通過將數據傳送到中央服務器進行運算的方式運作的。但出生低微卻平步青云的內里，卻預測說越來越多的這種處理將發生在“邊緣”，靠近數據的來源。

由于邊緣節點更接近于用戶終端裝置，可以加快資料的處理與傳送速度，減少延遲。潛在的應用包括基因研究和金融服務。對于這些領域來說，敏感信息如果傳輸到異地云服務器，將減少安全性。另外，內里還指出，邊緣計算可以協助一些比如制造業或自動駕駛的工作，這些領域里傳輸速度是至關重要的。

慧與科技側重于開發基礎設施，以便將邊緣計算與傳統的云服務連接起來。去年，內里在邊緣計算上下了大注，在未來4年里投入40億美元用于研發。

這筆投入要有大產出，還有待時日。最近一個季度，慧與科技的整體營收同比下降了7%。智慧邊緣計算部門，這一專注于移動和無線網絡的部門的營收也下降了3%。

但有許多潛在的邊緣計算應用預示著公司收入的增長。內里預測，在邊緣計算領域的投資，將會在2020年開始見到收益。

內里的擴張性投資計劃，以及數十億美元的股票回購和分紅，讓投資者情緒高漲?；叟c科技的股價在最近兩年營收下降的情況下相對穩定。

內里也能快速地消解與傳統硬件商的競爭，如戴爾和聯想。

對于中國的競爭者，內里認為它們主要是“價格低，把一切都商品化”。在成為CEO前，內里還是慧與總裁時，他曾宣布慧與科技將撤離“商品化”的服務器業務，部分原因就是來自中國低價競爭的壓力。

內里表揚了中國試圖實現科技自給自足的努力，但他又說，“現實是，他們還是需要西方繼續教他們，給他們賦能?！?/p>

這一觀點比較特別，因為很多人害怕中國科技和人工智能會挑戰西方的領導地位。但有些專家確實也與內里觀點一致，即懷疑中國領頭能力，特別是在中國著力發展的人工智能領域。

慧與科技的邊緣計算戰略中，包括了高度專業和高利潤的硬件項目。比如，慧與為美國國家航空航天局的阿耳忒彌斯計劃，定制了一臺超級計算機，用于將人送到月球以及送到火星，慧與甚至還在國際空間站安裝了一臺超級計算機進行本地化實驗數據處理。

由于數據傳輸中的延遲，云技術在太空中特別不適用，這造就了內里所說的“你能想象到的最遠的邊緣運算”。（財富中文網）

譯者：宣峰

Hewlett Packard Enterprise CEO Antonio Neri said Chinese tech companies have a long way to go before they become serious threats to U.S. firms—or become global tech leaders.

“The question is, how much innovation is really happening, as opposed to just manufacturing?” Neri said during an interview with Fortune this week. “I think they’ll always be one or two technologies behind.”

The comments came as Neri, now into his second year at the helm of HPE, assessed the competitors arrayed against him. In cloud computing, HPE faces off against Amazon and Google. But Neri considers those companies as primarily data storage providers, in contrast to HPE’s push into more comprehensive “edge computing” services.

The idea behind edge computing is straightforward enough. Currently, many analytics and artificial intelligence technologies, such as voice recognition, work by sending data to centralized server farms for processing. But Neri, who famously rose through the ranks after joining a Hewlett Packard call center in 1995, predicts that an increasing amount of that processing will take place at the “edge,” nearer where the data is collected.

Potential applications include genetic research and financial services, where sensitive data could be less secure if it's sent to an off-site cloud server. Neri also said that edge computing could help with tasks like manufacturing or autonomous driving, in which speed is paramount.

HPE is focused on developing infrastructure that will connect ‘edge’ cases that are reliant on local data processing to the conventional cloud. Last year, Neri put a big bet on edge computing, pledging $4 billion in R&D spending over four years.

That bet has yet to show big returns. In its most recent quarter, HPE’s overall revenue was down 7% year over year. Revenues for Intelligent Edge, the unit focused on secure mobile and wireless networks, dropped 3% on the year.

But many potential edge computing applications appear poised for growth. Neri has predicted that returns on the edge computing investment will start arriving in 2020.

Neri’s aggressive investment in his plan (along with billions of dollars in stock buybacks and dividends) has already bolstered investor sentiment. HPE's stock price—it closed at $16.41 on Thursday—has been relatively steady over the past two years despite revenue declines.

Neri was also quick to dismiss competition from traditional hardware manufacturers like Dell and Lenovo. Instead, he emphasized that the growing complexity of business technology gave the advantage to companies providing tailored systems bundled with support services.

Then there’s the class of China-based competitors Neri calls “China Inc.,” which he says are “all about pricing, commoditizing everything.” As HPE's president just before becoming CEO, Neri announced that HPE would exit the “commodity” server business – in part under pressure from cheaper Chinese competition.

Neri expressed admiration for China’s push for technological self-sufficiency. But, he said, “the reality is they need the West to continue to teach them and enable them.”

That contradicts worries by some that Chinese tech and A.I. could challenge Western leadership. Some experts do share Neri’s skepticism that China can take the lead, especially when it comes to China's efforts in artificial intelligence.

In contrast with Neri's views about China Inc., HPE's edge strategy includes focusing on highly specialized and high profit-margin hardware projects. For example, HPE built a custom supercomputer for NASA’s Artemis project to return humans to the moon, then send them on to Mars, and even installed a supercomputer on the International Space Station to process experimental data locally.

Because of delays in sending data, cloud technology is particularly impractical in space, making it what Neri called “the farthest edge you can imagine.”