The CVS-Aetna merger is all about the analytics: CVS Health has made its bid for Aetna official after months of speculation, saying it will pay $69 billion for the insurer in a deal that on the surface has as much risk as reward. The plan is to leverage CVS's nearly 10,000 retail pharmacy locations—an increasing number of which also contain MinuteClinc care centers—along with Aetna's substantial subscriber base, many of whom are already CVS customers, adding new services while driving down costs along the way. CVS is using about $4 billion of its own cash, with the rest coming from new debt and equity.
In a statement, CVS and Aetna characterized the merger as "a natural evolution for both companies as they seek to put the consumer at the center of healthcare delivery." The companies plan to drive the merger's success through their respective analytics and data platforms. Each brings somewhat differing strengths to the table. Aetna has for years been building out a deep data science and predictive analytics bench (and has had great success using it to combat fraud), while CVS has invested in a analytics platform from Epic, the large EHR (electronic health records) software vendor, among many other ventures.
The potential combinations of CVS's retail purchase and pharmacy data—both of which are tracked through its ExtraCare rewards program—along with Aetna's rich pools of provider-side data are all about creating a personalized health care experience that leads to reduced costs and improved patient health, the companies say. Here's one example provided by CVS and Aetna:
Twenty percent of Medicare patients are readmitted to the hospital soon after being discharged at significant annual costs, much of which is avoidable. Readmission rates can be cut in half if patients have a complete review of their medications after discharge from the hospital to help them manage their care at home.
POV: CVS and Aetna's merger is unprecedented from an industry standpoint, making the cultural adjustments found with any corporate merger perhaps more difficult than most. (One wonders if the belief that Amazon will enter the retail pharmacy business helped make CVS pull the trigger on such an expensive deal sooner rather than later—particular as competitors such as Walgreens are stumbling financially.) It must also pass antitrust muster with U.S. authorities.
CVS and Aetna are giving themselves an ample timeline for clearing that hurdle, saying they expect the deal to close in the second half of next year. The integration work that follows will no doubt play out over a number of years, but CVS and Aetna will have to make serious decisions about their analytics strategy going forward much sooner. The good news is that both companies have strong technical leadership in the persons of CVS CIO Stephen Gold and Aetna EVP Meg McCarthy.
IBM launches new Power Systems geared for AI: While IBM has lagged behind Amazon, Google and Microsoft in the cloud platform market, it's hoping to be the leader in AI workloads with the introduction of new Power Systems Servers that incorporate the Power9 microprocessor.
Built specifically for compute-intensive AI workloads, the new POWER9 systems are capable of improving the training times of deep learning frameworks by nearly 4x[2] allowing enterprises to build more accurate AI applications, faster.
The system was designed to drive demonstrable performance improvements across popular AI frameworks such as Chainer, TensorFlow and Caffe, as well as accelerated databases such as Kinetica.
As a result, data scientists can build applications faster, ranging from deep learning insights in scientific research, real-time fraud detection and credit risk analysis.
IBM has the U.S. Department of Energy as an initial Power9 customer; the agency will use the chips in its Summit and Sierra supercomputers. Google also plans to use POWER9 in its data centers.
POV: IBM has been working on Power9 for four years, and the fact that the likes of Google, which has developed its own AI-oriented chips, is using them speaks to the progress Big Blue has made. The Power9 systems also incorporate GPUs from NVIDIA
Power 9 supports up to 5.6 times more I/O and double the threads than "its x86 contemporaries," IBM said, in an allusion to Intel's x86-based servers that hold a 90-plus percent market share in data centers. IBM is hoping to capture 20 percent of the market through Power9 by 2020, Network World reports. That may be a lofty goal but IBM will certainly try its best. One way it will surely seek attention for Power9 is through deep learning benchmarks and its PowerAI software.
Earlier this year, it announced the results of tests using a 64-server Power system with 256 NVIDIA GPUs on the Caffe deep learning framework, saying it had bested a team from Facebook's AI research arm. The question is whether IBM can succeed in getting customers to move deep learning projects to its cloud—which will surely offer Power9-based instances soon—or make the investment in them for on-premises systems.
New York AG, senators demand net neutrality vote delay: As with all major policy decisions, the Federal Communications Commission held a public comment period concerning its upcoming vote to overhaul so-called net neutrality rules. Underscoring the topic's interest among the public, the period drew more than 22 million comments. The problem, says New York attorney general Eric Schneiderman and a group of 28 senators, is that more than a million comments supporting net neutrality being overturned may have been fake ones posted via bots.
The FCC board has a Republican majority, led by chairman and prominent net neutrality critic Ajit Pai, and it seems a foregone conclusion that the rules will be voted down. Net neutrality, which was passed in 2015 after years of debate, prohibits ISPs from favoring legal Internet traffic based on payments or other considerations. But opponents characterize the rules as an overreach that's been bad for competition and ultimately, consumers.
"A transparent and open process is vitally important to how the FCC functions," the officials said in a letter to Pai. "The FCC must invest its time and resources into obtaining a more accurate picture of the record as understanding that record is essential to reaching a defensible resolution."
POV: Thousands of the fake comments used the actual names and addresses of New York residents, which amounts to identity theft, Schneiderman wrote in a post on Medium. Schneiderman's office asked for relevant records in the course of its investigation nine times, but the FCC provided "no substantive response," he wrote.
This week, the FCC's inspector general agreed to cooperate in Schneiderman's investigation, but it's not clear whether anything will stop the Dec. 14 vote from occurring. For one thing, the public comment period has been over for some time, making it difficult from a procedural perspective to call for a makeover. Still, net neutrality proponents have other options; federal law states that decisions made by agencies such as the FCC can be overturned if they're deemed to have been made in an "arbitrary" or "capricious" manner.