The price war in the infrastructure-as-a-service cloud market shows no signs of abating this year, as Mike Wheatley pointed out in yesterday’s SiliconANGLE analysis. However, in the longer term both the pace and the variables at work will change as a result of two pivotal hardware trends that are poised to come together in the same narrow time frame, as well as changes in customer demands and expectations.

The end of Moore’s Law?

Fifty years after Intel Corp. co-founder Gordon Moore famously predicted that the density of transistors will double every two years, chip designers are finding themselves pressed against the physical boundaries of the media they work within. A former chief architect of the chip giant believes that the limit could be reached as early as 2020, while theoretical physicist Michio Kaku pins the end of the trail at around 2022.

Advancements in alternative computing paradigms such as the brain-emulating carbon processor that IBM unveiled last August suggest this barrier won’t pose a long-term threat to technological evolution, but many years may pass until breakthrough innovations hit the market. Until then, the pressure is on for conventional chip makers such as Intel to sustain the curve of Moore’s law as long as possible.

It’s already starting to take longer for manufacturers to advance down the scale. Nearly two years ago, Advanced Micro Devices, Inc. (AMD) blamed delays in its transition to the current 28nm process on the increasing difficulty of keeping up with Moore’s Law. That slowdown is creating a proportionate decrease in the amount of additional computational power buyers gain each year for every dollar spent.

As some of the biggest server buyers in the world, the top cloud providers – Amazon.com Inc., Microsoft and Google – will feel that decline the most. This is especially true given that the processor typically accounts for a bigger portion of the price tag on the bare-bone machines powering their data centers than comparable enterprise systems, which have value-added features that inflate costs.

Added up across the tens of thousands of servers that the cloud triumvirate will buy over the next 12 months, these factors could decelerate the rate at which providers can up their power and cut their prices as soon as this year. And since fixed expenses such as personnel and software don’t decline at the same rate, a bottom will eventually be reached.

Storage optimism

While the price of compute instances may not fall as rapidly this year as it has in the past, the opposite trend could emerge on the storage side, particularly in the fast-growing flash market. Wikibon co-founder David Floyer predicts that solid-state memory will become cheaper than traditional mechanical disk across almost every use case by 2016. For the same reason that the big cloud providers will be hit by the slowdown of Moore’s Law, they’ll be able to benefit from the fall in memory prices.

Customers should reasonably expect the cost of premium-priced flash options on the major platforms to come close to that of disk in 2015. Besides reducing how much they charge for storage, providers will be able to offer customers more flexibility in how they consume it. For example, Google is set to launch a new configuration option that allows users to have their virtual machines deployed on servers with flash directly attached, which brings the data closer to the processor and thereby speeds up performance. Microsoft, meanwhile, is rolling out a similar upgrade to its own platform that takes aim at workloads with steeper requirements than what disk can address.

Offering more ways to buy capacity will become increasingly important as the number of applications moving to the cloud continues to grow. After all, customers will move to the platform that can best accommodate the workload.

Cloud providers are also finding ways outside of pure technology to improve price flexibility and reduce costs. For example, Amazon significantly lowered the barrier to entry for its enterprise-oriented Reserved Instances last month with the introduction of two new payment options that offer an alternative to paying for everything in advance. Users now have the choice of putting down a down-payment of their bill or paying in installments over the duration of the lease.

Choice begets transparency

The next 12 months will see more of these kinds of innovations. As simpler pricing makes it easier to move workloads to the cloud, customers will demand more customization flexibility, driving yet more innovation in pricing and provisioning.

The domino effect that characterizes competition among Amazon and its rivals ensures that customers across all three major platforms will benefit from that growing focus on visibility, rounding out what is shaping up to be a very exciting year for cloud computing.