The co-editor of a computing magazine has said it is “not at all obvious” how HP Enterprise will cope with the common technical and economic issues that affect datacentres across the world.
Timothy Prickett Morgan of Platform said that splitting HP into two separate companies, “one focused on the datacentre and the other on PCs and printers […] doesn’t make any sense than it did bring all of these different businesses under the same roof in the first place”.
The writer mentioned CEO Meg Whitman’s previous comments that she would not consider breaking up the company, and he said that what changed her mind was “that the infrastructure in datacentres is being more precisely tuned for applications”, while increasing numbers of customers are opting for “minimalist server designs” that HP believed would eventually be “dominated by high-margin blade systems with converged sectors, storage and networking”.
He said that despite HP’s efforts to adjust to these changes, its financial results show the pressure it is under, and that its server partnership with Foxconn, established last year, and its new partnership with Tsinhua University, make these pressures “even more obvious”. Morgan continued: “If HP is pitching hyperscale machines made by Foxconn […] and now is ceding the infrastructure market in China to the H3C partnership that it is now a minority stakeholder in, something is clearly up”.
It was “absolutely right” according to the commentator to keep HP a “monolithic IT giant” as it was for IBM, but in the latter case it “broke itself up” and has had to sell off its PC, disk drive, high-end printer, X86 server and chip manufacturing enterprises in the past ten years. HP, meanwhile, “seems to be backing further and further away from high cost, low margin manufacturing and focusing more on being a master distributor of gear that is either made by partners or itself”.
Morgan said it “will not be at all surprising” should HP and IBM stop making datacentre hardware as a desire for bigger profit margins “push them in that direction”, which will still leave the companies with “lots of engineering expertise that can be brought to bear in systems”. Furthermore, the writer said that HP still has a profitable printing business, which will buy HP Inc. some time, although Lenovo and Dell are taking away market share and HP’s “worst enemy” of storage is becoming cheaper each year.
In the IBM case, selling off the PC business meant losing leverage with chip suppliers, leading to higher prices on server processors, while the computing manufacturer was unable to “meet the volumes of HP or Dell in the high-volume PC market”. IBM had its lower volumes and higher component prices undercut when hyperscalers and cloud builders came in, showing that “volume relationships with Intel are a big factor in determining how profitable a system, storage, and increasingly networking device is going to be”.
The technology writer said one can garner an idea of what HP Enterprise will be like by looking at the financial results of four parts of the current HP: Enterprise group, Enterprise services, Software and Financial services, as the majority of these units’ revenues focus on the datacentre. Financial services and Software are “pretty steady”, with revenue of $1 billion (€910,000), while Enterprise group, which includes servers, storage, switches sales, and tech support for these systems, has risen since the recession but has fallen over the past five years.
Morgan said that HP is facing competition “on the X86 front” while its Itaniu-based Integrity enterprise has been in decline for several years. He continued: “HP bought networking and storage businesses, but more open approaches based largely on sophisticated software and cheaper hardware have made it tough for HP to reap the revenues and profits it no doubt had planned”. In this respect, the OEM is in a similar position to other IT giants IBM, Dell, Oracle/Sun and Fujitsu.
Enterprise services have similarly struggled to grow and HP has said it will cut an additional $2 billion (€1.8 billion) in costs in its services business to synch up revenue and profits. Morgan’s view is that while these “are still very large businesses” HP is re-evaluating how to make them “more profitable”. One way to do this it through manufacturing and design partnerships, but these will mean “HP loses a skillset and leverage with parts suppliers”.
HP Enterprise is the preferable half of HP to own, according to the writer, and he reckoned Whitman must be tempted to cut back on outsourcing and hosting businesses that are not bringing in substantial profits as she considers a team for running the Enterprise side. Operating earning are a lot lower compared with six years ago, but Morgan says this is not attributable solely to hardware, as Enterprise services experienced similar revenue drops to the Enterprise group despite several years of restructuring, although he added that the recently strengthened US dollar and the weakening of the euro and the yen have affected “all US-based multinationals who sell a lot of stuff in Europe and Japan”.
Morgan said that now automation has come to the datacenter, the squeeze on it will never end, leading the commentator to question why Whitman and CFO Cathie Lesjak, who is set to join the PC and printer side “are not worried about an additional $400 million (€365 million) to $450 million (€411 million) in annual costs that the separated companies will incur after the breakup”. Other costs will come as the supply chain is affected, although the executives “think they can eliminate around $1 billion (€914,000) in costs to more than cover the extra overhead”. The writer said that “it seems that the risks of breaking up are the same as the risks of staying together” and that a “different set of decisions” may have been made if the executives were not largely paid in stock.
The writer concluded that what is of most import is “how HP competes as a supplier of hardware and services”. He said that if it had been an “innovator” in the public cloud, like Amazon Web Services, “it might have a much larger services business that consumed a lot of its own iron”, but would probably not have Microsoft for a customer for the Azure cloud. Morgan also said that an OS for the X86 machinery would have been “useful”, as would a server-storage hybrid with a similar impact to Nunatix, while it “would have been best” for the OEM not to have passed by the smartphone and tablet revolutions.
“It is hard to do all things and ride all trends to profits. In fact, it is impossible, which is why no company – not even diverse ones built up over time like HP – have ever done it”, he said.