Yesterday, Nintendo
officially unveiled its Switch 2 handheld via a first look video presentation. Featured content did not come as a surprise to many gaming enthusiasts—a steady flow of leaks have already revealed
outer and
inner workings. Earlier today, the Digital Foundry team has offered their collective opinion on Nintendo's formal announcement. Their roundtable discussion first focused on the Switch 2's physical appearance—mainly a showcased physical increase in size, when lined up against the preceding (standard) model. Conversation quickly moved onto technical matters—a topic that Nintendo normally avoids discussing. The video presentation included in-game footage of a next-gen Mario Kart title—Oliver Mackenzie (a contributing DF video producer/writer) was not impressed by this short demo's visual fidelity. He noted an absence of DLSS image enhancement—surprising, given that the rumored NVIDIA Tegra T239 SoC is capable of deploying this graphics technology.
John, Rich and Oliver then moved onto discussing recently
leaked clock speeds and performance figures (in handheld and docked modes)—overall, they reckon that these numbers seem fitting for a hybrid system. They noticed that the handheld GPU clock was lower than expected—based on their judgement of the Switch 2's fairly capable integrated cooling solution. In the past, Digital Foundry theorized that the NVIDIA-designed Tegra T239 will be an 8 nanometer part—rumored to be built on Samsung 8 nm DUV foundry node. Newer gaming community-generated proposals have suggested a shift to Samsung's
5 nm EUV node—mostly based on the chipset's physical footprint. In sharp contrast, the Digital Foundry guys are sticking with their 8 nm theory. Richard Leadbetter (DF's founder) has previously attempted to simulate Switch 2-esque performance on readily available Ampere-based hardware—he could revisit and perform tests on a laptop that sports Team Green's GeForce
RTX 2050 mobile GPU. He believes that the leaked CPU and GPU clocks (across both modes) present plausible evidence of 8 nm-level performance, cross-referenced with his team's past analysis of the system's PCB. Debates will inevitably rage on, but Rich insists that the end result will be an example of "
Occam's razor." The Tegra T239's four (long alleged) Cortex A78 cores appeared to be running at a higher frequency in portable mode than in docked—suggesting some unknown factors; perhaps a switching on or off of cores (situation dependent). Leadbetter and Co. will be looking forward to getting a proper hands-on experience at Nintendo's April to June launch events.