Here's how Microsoft is helping companies build IoT hardware

One of the biggest challenges with building connected hardware is getting from proof-of-concept (PoC) prototypes to devices that are ready for large-scale production rollout. Microsoft is aiming to help through labs that allow companies to come in and work with experts on building internet-connected hardware.

Companies come into one of three Microsoft Internet of Things and Artificial Intelligence (IoT/AI) Insider Labs with the hardware they’ve built so far and a plan for an intense two or three weeks of work. Visitors are paired with mentors who are experts in different areas and given access to machinery that can help them quickly work through different hardware designs.

The goal of the labs is to shave months off companies' development timeline. Microsoft also provides help with configuring its cloud services for use with the hardware developed in the labs, which is how the tech titan benefits from the program.

There’s a need for programs like this because of the challenges of building connected hardware. Creating new IoT devices is often difficult, according to Dima Tokar, the co-founder and CTO of IoT analysis firm MachNation.

“Today, many enterprises are stuck in the PoC phase of their IoT journey because they are finding that it is more challenging than they expected to take that PoC and make it production-grade,” he said in an email. “It’s difficult because all components of an IoT solution need to scale, and figuring out the IoT security posture as well as management and oversight requirements can be a daunting task.”

Squeezing months of work into weeks of lab time

A recurring theme among lab administrators and participants is that the environment Microsoft created lets teams get done in a matter of weeks what might have otherwise taken months. The IoT/AI Labs include machines that can help with rapid prototyping of hardware, including the design and testing of printed circuit boards (PCBs).

Testing PCB designs with a contract manufacturer can be a time-consuming process. Companies have to send their design off, then wait for the manufacturing run and subsequent shipping. In contrast, Microsoft’s labs are capable of cranking out at least two iterations of a circuit board in a single day, according to Cyra Richardson, a general manager of business development for IoT at the company.

Another benefit to visitors are the labs’ dedicated engineers, who are there to help work through problems in person. Richardson pointed out that seeking answers online for engineering problems doesn’t always beget definitive or useful answers.

“Imagine if you had someone who cared about you were doing, and really was invested in your acceleration,” she said. “And imagine if they were right there in front of you.”

Microsoft has also partnered with a set of ecosystem partners, including Cisco, Dassault and Seeed Studio to help lab participants with areas outside of the Redmond company’s expertise.

“When you think about these new intelligent systems, these new products, you have multiple elements that need to come together. Hardware, industrial design, software, cloud services, all the way through to natural user interface for these products,” Kevin Dallas, Microsoft’s corporate vice president for IoT and intelligent cloud business development, said in an interview. “It’s quite a lot of technology, and it’s easy to get lost in that technology in terms of delivering backend products.”

Entering the labs

Getting into the labs requires an online application, which asks for information about what a company is working on, the size of their team, and other details. Microsoft doesn’t charge for a visit to the Insider Labs, but the company does handpick the teams that get to participate.

One such firm was New Sun Road, a company based in Berkeley, California, that builds solar power systems for the developing world.

“We basically needed to fill some holes,” NSR co-founder Jalel Sager said. “Not just in the cloud platform, but also around some of those IoT devices that would take that information up to the cloud.”

In NSR’s case, the staff at the labs helped in a variety of different product areas, including circuit board design and picking out which Microsoft cloud services made the most sense. In the event NSR encountered a snag that the lab engineers couldn’t solve, it was possible for them to get ahold of a member of the Microsoft product team who could answer their question.

The team at Sarcos Robotics came into the IoT/AI labs with the aim of getting their forthcoming set of robots hooked up to the cloud. The company was previously a part of Raytheon, and has plenty of experience building human-controlled robots. Ben Wolff, the company’s chairman and CEO, said that Sarcos went to the Redmond lab for help with getting its Guardian S robot connected to Azure.

“Historically, we have not collected any data from the sensors [on a robot] or used the sensor data in any way other than to just control the real-time operation of the robot,” he said.

In his view, changing technology means Sarcos customers can benefit from using sensor data to monitor a robot’s performance as well as get information about the environment it’s in. Work that the Sarcos team did with Microsoft will allow the robot to be used as a platform for collecting sensor data and bringing it into the cloud for further analysis.

Engagement with the labs doesn’t stop after a company leaves the building. Participants in the program are able to reach back out to Microsoft for more help if they need it, and the program does allow repeat visits.

Into the future

At the moment, Microsoft has three IoT/AI labs. One is on the company’s campus in Redmond, Washington, and the others are in Shenzhen, China and Munich, Germany. Dallas said that he’s happy with the number of locations currently available for the moment, but plans to expand the program in the future.

Microsoft is measuring the success of the labs based on how many companies are able to bring their products to market faster as a result of their engagement.

That’s how the company will see a benefit from the program, after all: The major revenue opportunity for Microsoft Azure comes from having many of these IoT devices deployed in the world and driving consumption of cloud services.

Hardware requires a long lead time, so it could be months if not years before IoT/AI Labs graduates have products available for purchase.

The labs are open for any interested organization to apply, if they think they’re a good fit for the program. Dallas said Microsoft selects companies based on whether the lab can reduce their time to market, and if the company can add unique value to the product in development.

While Azure usage is a clear benefit to Microsoft, entering the lab doesn’t require a commitment to use the company’s cloud platform, according to Richardson.

Microsoft is also using feedback from customers who come through the labs to refine its products going forward.