Shortages of memory and other components are starting to impact server availability. This is already having tangible consequences for organizations that can no longer afford to refresh their infrastructure. The situation is also creating new challenges for software vendors.
The hardware market is in rough waters, and this is not an isolated problem. Today, almost every company is a digital company, and software still needs to run on a server somewhere. When server manufacturers can no longer find RAM, SSDs, or CPUs to build their servers, the foundation of the digital business world begins to show cracks.
Luc Costers, Regional Leader Nutanix Belux, CIS, and Eastern Europe at Nutanix, is seeing this from the front row. While he may be the regional head of a major software company, the hardware crisis is affecting him and, above all, his customers.
New norm
“Anyone looking to order servers or storage components today is faced with prices that are barely predictable and lead times that stretch to five months or more,” says Costers. “This is not a temporary inconvenience; the situation has become the norm.” For both SSDs and RAM, major players have bought up production capacity for more than a year. Costers therefore calls the expectation that the situation will normalize within a few months unrealistic.
Anyone looking to order servers or storage components today is faced with prices that are barely predictable and lead times that stretch to five months or more.
Luc Costers, Nutanix
What started as a theoretical problem is increasingly translating into concrete limitations for businesses. Supply chain issues in hardware begin as a budget matter but grow into an operational risk that is difficult to manage.
There are two major problems: first and foremost, hardware is rising in price, which impacts the infrastructure organizations can afford to buy. However, deep pockets are not enough: even those with large budgets cannot simply obtain the necessary servers because there are simply insufficient components available.
Unpredictable prices
Regarding price, there is a lot of uncertainty. Hardware prices are currently rising very rapidly. “Companies that wait three months to order sometimes find themselves forty percent short of their originally planned budget,” Costers observes. “By the time a budget increase is approved, prices have risen again.”
Furthermore, some suppliers are communicating to customers that the price at the time of delivery may be higher than the ordered price. “The order price is no longer guaranteed to be the delivery price,” Costers emphasizes.
The most direct consequence of the price increases is that companies are ordering less hardware. Customers are playing it safe financially and ordering only half of what they actually need operationally. When infrastructure is budgeted too tightly, follow-up problems arise. Costers: “The pressure on operational infrastructure increases, with a higher risk of problems as a direct result.”
Limited availability
Even when budgets are high enough, hardware proves unavailable in the short term. Manufacturers have taken precautions and stocked up on older equipment, although this is also sold at a premium. Purchasing new infrastructure has become a complex, pricey, uncertain, and slow process today.
I don’t think we’ve ever experienced a problem on this scale before.
Luc Costers, Nutanix
Costers, who has been active in the IT sector since the 1980s, calls the situation unprecedented. “I don’t think we’ve ever experienced a problem on this scale before,” he states. “If the major AI players cancel some orders soon, some oxygen might return to the system, but as long as they buy everything up, this painful situation will persist. I don’t actually see the hardware problem being resolved before the end of 2027.”
Solutions from the software side
Nutanix is not standing on the sidelines. On one hand, the company is feeling the impact. Costers explains: “Software licenses are calculated based on the underlying hardware. The ratio between hardware and software licenses in an average deal has already shifted. While license costs accounted for 50 to 55 percent of the total deal size last year, that share has dropped to 35 to 40 percent. Hardware has become proportionally more expensive.”
On the other hand, Nutanix is positioned to offer a solution in these difficult circumstances. The Nutanix platform sits between the hardware and the applications and databases that organizations truly need. This offers possibilities.
Combining and retaining
“The platform allows nodes of different generations to be combined into a single cluster,” says Costers. “A customer who started with ninth-generation servers can supplement them with generation eleven hardware without replacing the entire cluster.”
Expanding instead of replacing is a key argument. This allows organizations to limit their capital expenditure to some extent. Nutanix actively tries to support this expandability.
“We are working hard on the certification of older servers for the platform,” Costers illustrates. “Servers that might normally be nearing the end of their depreciation but could actually last a few more years, we try to support so they can continue to function in hardware clusters.”
Servers that might normally be nearing the end of their depreciation but could actually last a few more years, we try to support.
Luc Costers, Nutanix
This shows how much hardware shortages have changed the narrative. Whereas the power of HCI and, by extension, Nutanix a few years ago was that you could easily save energy and space with new and efficient hardware, the focus is now on integrating what is available. No server is more interesting today than the one you already have.
The cloud as an interim solution
The role of the cloud is also changing. Companies that urgently need computing power now but cannot obtain or afford hardware can look to the cloud as an interim solution. Running stable workloads in the cloud is generally not the most economically viable long-term solution, but the availability of the cloud is a major asset right now.
Costers: “Via NC2, the Nutanix platform can be deployed on bare metal servers at Azure, Google, or Microsoft, or at a European alternative like OVH or a local provider. This offers a way out as long as on-premises hardware is difficult to obtain or too expensive.”
The cloud also has its limits. Costers describes it vividly: “If the people at Azure rent out all the boxes they have, then there are no more. Cloud capacity is also finite, and the prices of rented servers can also rise as demand grows.”
Pragmatic approach
Costers suggests a pragmatic strategy: put workloads in the cloud for now, and reconsider the situation when on-premises hardware becomes affordable again. “Because Nutanix licenses are portable, a return to your own infrastructure is possible at any time,” he says of the company’s own platform.
That idea of flexibility, independent of hypervisor, hardware, or location, is exactly what Nutanix wants to convey right now. The same platform runs on-premises, in the public cloud, or on bare metal. In a market where hardware availability is uncertain and prices fluctuate, you can see that independence as a practical necessity. Whether it is enough to offset the consequences of the hardware crisis remains to be seen, but every little bit helps for those looking for a solution today.