Graphics specialist NVIDIA (NVDA -0.66%) is on a roll. The company reported record financial results in its most recent quarter, with both gaming-related revenue and corporate gross profit margin surging to impressive heights.
Following the results, a number of Street analysts weighed in on the results, with the majority of the remarks coming in positive. What I found particularly interesting, though, was commentary from analyst Craig Ellis with B. Riley & Co.
The analyst believes that the company is on the verge of its "biggest product cycle in a decade," which could potentially drive substantial growth in NVIDIA's gaming business in its current fiscal year -- well ahead of what most are modeling it at.
Does Ellis' analysis make sense? Let's take a closer look.
What Pascal will bring to the table
Typically in the world of gaming graphics processors, performance enhancements are driven by the advancements in chip manufacturing technology, widely dubbed Moore's Law.
Graphics applications are quite unique in that they can be made faster by simply throwing more hardware at the problem. Typically, new generations of chip manufacturing technology allow chip companies like NVIDIA to pack in far more computing power into a given chip area.
NVIDIA has been shipping graphics processors built on Taiwan Semiconductor Manufacturing Co.'s (TSM -1.04%) 28-nanometer chip manufacturing technology since early 2012. The company has been able to deliver substantial performance improvements over the last four years by making targeted improvements to the underlying chip architecture, as well as by simply making the chips larger, but at some point those gains simply become too difficult/costly to achieve.
Fortunately, the company's next-generation Pascal architecture should mark the transition to Taiwan Semiconductor's 16-nanometer FinFET Plus manufacturing technology. The 16-nanometer process delivers a roughly 50% reduction in transistor size (meaning twice as many can be packed into the same area) as well as a substantial boost in underlying transistor performance (Taiwan Semiconductor claims it's on the order of 65% in its marketing materials).
Additionally, NVIDIA is expected to use a new type of memory known as High Bandwidth Memory 2 in higher-end graphics chips built on the Pascal architecture. This new type of memory is expected to deliver a substantial increase in memory bandwidth, which will be required to feed what is almost certainly going to be a much more power.
What this means for NVIDIA
If Pascal is the huge leap forward that I think it will be, then this really could serve to catalyze a massive upgrade cycle. NVIDIA's results over the years have shown that gamers are willing to pay for substantial improvements in performance.
I believe that if NVIDIA is able to roll out a top-to-bottom stack of products based on its Pascal architecture, it should be able to deliver substantial performance increases at just about every price point.
My guess is that the higher-end models will come first, since these command the highest average selling prices and are targeted at the portion of the customer base that's willing to pay the most for excellent performance. Afterwards, I'd bet on the lower-end, lower-cost models rolling out.
As long as NVIDIA delivers with Pascal, and as long as it can get products based on it out solidly in the current fiscal year, the graphics specialist may very well see a monster upgrade cycle -- great news for company and its stockholders.