GE announces Predix Cloud for the Internet of industrial things, available to customers in 2016
Barb Darrow / Fortune :
Context & Ripple Effects
When GE announced Predix Cloud in August 2015, it was betting that an industrial manufacturer could own the full stack — machines, analytics, and the cloud they run on — rather than rent it from Amazon or Microsoft. Two months later GE was claiming the platform was on pace for $6B in revenue that year, and through 2016 it kept buying into the thesis, partnering with HP Enterprise on IoT offerings across aerospace, oil, gas, and manufacturing and acquiring machine-learning firm Wise.io to deepen the analytics layer.
The rest of the arc is the correction: by mid-2017 Reuters reported Predix facing delays, cost cuts, and a retreat from building its own data centers toward expanded partnerships, and by December 2018 GE had spun the whole business out as a separate company with just $1.2B in revenue. The gap between the 2015 claim and the 2018 outcome is what makes this announcement worth rereading.
First-order effects
- GE's industrial customers get a vendor-controlled destination for machine data in 2016, making GE both their equipment supplier and their infrastructure provider — a bundling no hyperscaler could offer.
- Bosch's parallel move to build an internal IoT cloud, later offered externally, shows GE's announcement immediately framed the category as one every large industrial must enter rather than rent into.
Second-order effects
- GE's retreat from owning data centers forces it into exactly the partnership structure it announced against — HPE becoming preferred Predix service vendor is hyperscaler-style economics winning inside an anti-hyperscaler strategy.
- Competing industrial giants face a pricing benchmark problem: GE's public $6B revenue framing sets expectations for industrial-IoT platforms that Bosch's slower, internal-first launch deliberately declines to match.
Third-order effects
- If the pattern holds, vertical software stacks built by manufacturers consolidate away from owner-operated clouds toward platform companies funded and governed separately — GE's 2018 spinout of a $1.2B-revenue business is the template.
- The durable lesson for industrial buyers is that analytics capability (the Wise.io acquisition) survives the infrastructure bet: the value concentrates in the machine-learning layer, not the cloud beneath it.
The trend: Industrial manufacturers that tried to own their own clouds are converging on a split structure — hardware and analytics kept close, commodity infrastructure rented — with spinouts and partnerships marking each step of the reversal.