據(jù)貝恩咨詢公司（Bain & Co.）的數(shù)據(jù)顯示，未來四年，很多大公司對視頻監(jiān)控、工廠傳感器等智能聯(lián)網(wǎng)設備的投入將增加一倍以上，總投資額或?qū)⑦_到每年5200億美元。

這5200億美元包括了市場對相關(guān)設備、軟件和服務的需求。2016年，貝恩公司曾做過一項類似調(diào)查，當時貝恩公司曾預估，到2020年，市場對聯(lián)網(wǎng)設備的需求將達到4500億美元。今年的預期已明顯高出了這一數(shù)字。除了企業(yè)界的需求之外，普通消費者對聯(lián)網(wǎng)設備的需求也在上漲，比如智能麥克風、能連接Wi-Fi的燈泡等等。

當前一代的物聯(lián)網(wǎng)設備通常會將收集到的信息（包括大型視頻文件等）發(fā)送到云數(shù)據(jù)中心進行分析，以監(jiān)控安全區(qū)域、分析客流量、監(jiān)測零部件消耗速率等等。但新一代物聯(lián)網(wǎng)產(chǎn)品將自帶更強大的計算能力，有的還內(nèi)置了人工智能應用程序，從而變得更加獨立和高效，理論上這也必將進一步推高它們的銷量。

不過也有一些人認為，物聯(lián)網(wǎng)設備或許不會有大家預期的那樣受歡迎，我們或許有必要給物聯(lián)網(wǎng)的一些熱門領(lǐng)域澆澆涼水。比如，很多公司希望大量收集關(guān)于某些設備的數(shù)據(jù)，以決定如何科學地對其進行維護，進而降低成本。

比如電梯生產(chǎn)商迅達集團（Schindler）就與通用電氣（GE）合作，從6萬余部電梯上收集傳感器數(shù)據(jù)。但由于缺乏歷史數(shù)據(jù)，再加上集成不同數(shù)據(jù)格式的難度，使得迅達很難基于這些傳感器數(shù)據(jù)科學預判維保時間。貝恩公司指出：“見解的獲得要比一開始承諾的更難。”

企業(yè)用戶對物聯(lián)網(wǎng)的最大疑慮，與2016年的調(diào)查并無二致。2%的企業(yè)認為物聯(lián)網(wǎng)設備存在安全風險，可能導致黑客入侵其計算機系統(tǒng)。29%的受訪者認為物聯(lián)網(wǎng)系統(tǒng)很難與其他新舊系統(tǒng)整合，還有28%的受訪企業(yè)擔心物聯(lián)網(wǎng)設備的投資回報率太低。

貝恩公司并未披露受訪企業(yè)的總數(shù)。不過調(diào)查發(fā)現(xiàn)，大企業(yè)客戶一般更喜歡“即插即用”式的物聯(lián)網(wǎng)設備，而不喜歡需要高度定制的系統(tǒng)。目前有60%的受訪企業(yè)都表示需要對物聯(lián)網(wǎng)設備進行定制。

這份調(diào)查還著重介紹了一些成功案例。比如IBM公司利用了與三星（Samsung）合作獲得的見解，幫助電梯制造商通力公司（Kone）優(yōu)化了電梯維保日程。法國鐵路公司SNCF跟蹤了數(shù)千列火車的行車數(shù)據(jù)，提高了鐵路交通安全性。威瑞森公司（Verizon）通過收購了四家公司（包括2016年以34億美元收購車輛軌跡追蹤公司 Fleetmatics Group），建立了一項監(jiān)測汽車軌跡的無線實時服務，等等。（財富中文網(wǎng)）

譯者：樸成奎

Big companies plan to double their annual spending on smart, Internet-connected devices like video surveillance gear and factory sensors over the next four years to an annual total of $520 billion, according to consulting firm Bain & Co.

The estimate, which includes purchases of devices, software, and related services, is up from a projected $450 billion for 2020 made by Bain in its previous survey about the topic in 2016. The higher forecast shows businesses are increasing their appetite for connected devices alongside growing consumer demand for everything from smart speakers to Wi-Fi connected lightbulbs.

Connected products used today usually send the information they collect, including big video files, to cloud data centers for analysis to monitor secure areas, measure crowds, or keep tabs on how quickly parts are wearing out. But newer products will have more of their own computing power and artificial intelligence apps built in, making them more independent and efficient, at least in theory, boosting sales.

But some expected uses of connected devices haven’t caught on as much as expected, according to Bain, providing a reality check for the hot field. For example, many companies hoped that collecting extensive data about their equipment would help predict when to service them, reducing costs.

In one example, Bain pointed to elevator manufacturer Schindler working with GE to collect sensor data from 60,000 elevators. But a lack of historical data and problems integrating different data formats made predicting maintenance needs difficult. “Insights have been harder to glean than first promised,” Bain noted.

Overall, the biggest concerns business customers had about the Internet of Things remained the same now as in the 2016 survey, Bain said. Forth-two percent of companies cited the risk of security weaknesses that could allow hackers to infiltrate their computer systems, while another 29% cited difficulties integrating new and old systems, and another 28% feared poor returns on their investment.

Bain’s survey, based on checking in with an undisclosed number of companies, also found that big customers are looking for connected devices that that they can plug in and use right away, instead of systems that require substantial customization. Currently, 60% of those surveyed said they have to customize the products.

Bain praised IBM for using insights it learned working with Samsung to help elevator and escalator maker Kone improve maintenance schedules and the French rail company SNCF track thousands of trains and improve safety. Using a different strategy, Verizon built a real-time wireless service to monitor vehicle fleets by acquiring four firms, including the $2.4 billion purchase of trucking tracker Fleetmatics Group in 2016.