IoT news of the week for Jan. 17, 2020

The climate is beating out cyberattacks and data leaks as a top global risk: Each year the World Economic Forum puts out a report detailing the top global risks. It’s not light reading. This year’s report includes cyberattacks and data fraud or theft as among the top risks, but they have decreased in importance from a year or two ago as climate-related risks have risen. Yay? (ZDnet)

Seeq formally announces its funding: Last week, I linked to a CRN story about analytics startup Seeq raising money. This week, I spoke with Seeq to understand what the company does and to confirm that it had raised the second tranche of a Series B round for a total of $48 million in Series B funding. Turns out it may add a bit more to the overall amount for this raise. Even more interesting, however, is what Seeq is doing. In the last year, it has doubled the number of its staff selling analytics for time series data, from 60 to 120. Companies point Seeq’s software at stored time series data (it can be stored in a historian, in an S3 bucket, or some other locale) and get analytics for the data. The Seeq spokesman likened it to Excel for time series data for plant operators. —Stacey Higginbotham

The Verge discovers Google’s interest in industrial IoT: The Verge has dug into Google’s Coral development boards and history. I didn’t realize that Coral, the machine learning platform for edge AI, is based on the fun AIY kits that Google released for kids and makers. Apparently people were buying them for parts and trying out low-latency machine learning tasks on them, so Google saw a niche. The article also points out that Coral only works with Google’s cloud and infrastructure — which somewhat limits its audience — and correctly assumes that Google’s real aim here isn’t to sell a lot of boards but to get a sense of industrial workloads that might help it build future infrastructure products. (The Verge)

Apple buys XNOR on-device image recognition startup: I’ve liked XNOR since learning about it almost three years ago. The company made software that optimized image recognition on so-called current chips. While many companies are building specialized hardware for image recognition and other ML problems, the XNOR team focused on delivering a software product that could run on existing hardware. The solution was interesting enough that its fellow Seattle startup Wyze put the XNOR software on its cameras to detect people. Unfortunately for Wyze camera users, a few months back XNOR said it would stop providing the software, so Wyze had to shut that feature down. At the time, I speculated that XNOR was in acquisition talks, likely by a company that might have its own cameras that it wanted to optimize. That Apple is the buyer makes that idea so much more interesting. (Geekwire)

Did you migrate your SmartThings app yet? After a year of beta testing, SmartThings users are being moved over to the new SmartThings app (this will require those users to have a Samsung account). The goal is to push its users to the new app and eventually only have one app to support. To that end, users electing to stay with the old SmartThings app will no longer be able to view or edit their routines or add new ones. They will still be able to control their devices using the old app, though. Users can migrate by way of the app migration feature, which theoretically will bring all the existing devices and routines from the old app into the new one. (SmartThings)

Equinix’s Packet buy has an edge computing angle: Despite my former immersion in the world of data centers, I will admit that I didn’t see much of an IoT story in the decision by data center colocation facility Equinix to buy Packet. Packet lets people build their own applications on non-virtualized hardware, or so-called “bare metal” servers. Packet has also provided the underlying hardware for several IoT edge computing networks. IT places the servers inside telco central offices, colocation facilities, and other areas to provide a unified infrastructure layer for companies that want to create IoT networks on top of it. (Data Center Frontier)

NIST has released its framework for privacy: The National Institute of Standards and Technology has released the first version of a framework designed to help companies protect consumer privacy. I have not delved deeply into it yet, but will in the next week. In the meantime, please feel free to take your own look. (NIST)

How retailers plan to build their IT infrastructure for the IoT: In my lead story I talk about how retailers are adopting IoT to build more efficient businesses. If that’s your thing, you may be interested in this report from Nutanix, a hyper-converged infrastructure vendor, which did a deep dive into the infrastructure demands of retailers as they embrace technology. (Nutanix)

Making sense of 1B data points: In this case study for Microsoft Azure services, a glass-making company discloses how it has taken data from 4,000 sensors and tried to pull the 1 billion data points generated from those sensors out of silos so plant operations teams can make sense of it. The company managed to reduce defects by 5% and increase productivity by 25%, but what stuck with me was the fact that to do this, it had to hire a Microsoft partner to build custom software for the glass manufacturing plant that took into consideration custom equipment and processes. Basically, this says that the benefits of IoT and AI are real, but getting those benefits still requires a highly customized solution. (Analytics India Magazine)

Why Wind River purchased Star Lab: I covered the acquisition earlier this week if you want to know more. (StaceyonIoT)

There was a lot of technology for aging at CES: Kevin Tofel writes about a series of startups that we saw at CES focused on providing everything from fall detection capabilities to telemedicine. (StaceyonIoT)