Steam Deck released in 2022 and then in 2023 we
got a Steam Deck refresh. The ROG Ally released in 2023 and I’ve gotta say that in some ways, the
MSI Claw is looking a lot like a 2024 refresh of the ROG Ally, and I think that’s amazing.
But just like with the ROG Ally, we do have to be a little bit cautious with previews. There
are a lot of marketing claims being thrown around and for the most part, these claims
are being looked at uncritically. To be fair, this is not a judgment on the journalists
that are covering the MSI Claw but I feel strongly that it would be helpful to look at
the ways that a tech company can manipulate the coverage especially for a device that is
still a month or so away from release.
Unfortunately for me, there’s always the
chance that this sort of video ruins my odds at getting a review unit but as usual,
I’m OK with that proposition. I’m not owed a review unit and I don’t want a review unit that
comes at the cost of my honesty. Still though, I give MSI the benefit of the doubt and
honestly it seems more reasonable to assume that they’d never see my video in the first
place or even know who I am. All the same, I will purchase the MSI Claw myself
if I don’t get a review unit and I will give you my honest appraisal of the
device once I have it in my hands.
In any case, before I go into the previews,
let’s quickly run down what we learned.
There are 3 models for the MSI Claw that should
be available sometime in February or March; more likely March. . For $699 you get a Core Ultra 5
with 512 GB of storage and 16GB of RAM. For $749, you get a Core Ultra 7 155 H also with 512GB
of storage and 16GB of RAM. Finally for 799, you get the same 155 H with 1TB of storage
and still 16GB of RAM. So there is no model with 32GB of RAM. Also - the drive is a 2230 so it
should be pretty straightforward to replace.
The claw is so called for its ergonomic
benefits - essentially they took the ROG Ally’s basic design and made the grips more
substantial. There is a lot of ventilation, like literally all over, and that’s good because
this appears to be an APU that has a different power envelope to the AMD handhelds. It doesn’t
appear to hit its stride until you feed it more power and according to Dave2D, performance at 30W
compares to the Ally’s performance at 25W.
There’s a 53Whr battery, 7” 1080 LCD screen with
500 nits peak brightness. VRR is not available but they’re still investigating whether or not
it will be included by launch… we’ll talk about that. The I/O seems straight out of the ROG
Ally mold, including an SD card slot right at the top and a dual fan design. This is also
the first handheld with Wi-Fi 7 and Bluetooth 5.4 and it’s the first handheld from a major
distributor that has both hall sensor analog sticks as well as hall sensor triggers.
So with that, let's take a look at the ways that MSI have manipulated circumstances,
just like other tech companies do, in order to make sure the MSI Claw gets
the most positive coverage possible.
So I watched every video I could find and
also read some articles too and there’s one term that showed up in I think every single
post and that was “engineering sample”. All of the journalists wanted to stress that these were
engineering samples and that a lot of the details being discussed were subject to change.
Now, the fact that all the journalists were evaluating engineering samples meant a few things
but one of the most important things that caught my eye were some unavailable features.
The biggest feature that was unavailable but marketed, at least to some degree, was
Variable Refresh Rate. To be fair to MSI, they did not say that VRR will absolutely
be available on the MSI Claw, but they did say they are looking into it. And it should be
noted here that the MSI Claw is set to release in February or March according to MSI and while
it is plausible they are still working on VRR, that is sort of specification is something that
I would expect to be finalized by now.
Let me give you a related example that comes
to mind with the Lenovo Legion Go. During the initial reveal of the Legion Go, some
people noted that the engineering samples there were using Portrait displays. One of the
Lenovo engineers in the Lenovo forums pointed out that although the early samples were
using portrait displays, the final product would be using landscape displays.
This engineer turned out to be wrong. Now I don’t know what happened in the short time frame between
when he made this claim and when the Lenovo Legion Go was ultimately released, but the Legion Go
ended up launching with a Portrait display despite his pre-release protests to the contrary.
These sort of unavailable features present a dangerous proposition. A journalist is forced to
report on the fact that an MSI rep is telling them that VRR may be in the ultimate product but MSI
themselves are hesitant to put a VRR bullet point in their official marketing materials despite
the product being a month or two away.
Another similar feature is Wi-Fi 7. We learned
that the MSI Claw sports Wi-Fi 7, a standard which was, more or less, officially launched at
CES. To an extent, Wi-Fi 7 itself suffers from the same exact marketing problems that we’re talking
about here. Check this quote from an article on ZDNet that covers the launch of the Wi-Fi 7
Certification Project. The author says:
“ our homes and offices will soon see speeds
of -- believe it or not -- almost 5.8 Gigabits per second (Gbps).”
But then he immediately follows that by saying:
“Actually, just between us, I don't believe Intel when it states, "a typical
Wi-Fi 7 laptop [will] have a potential maximum data rate of almost 5.8 Gbps."”
So yeah, these are some lofty goals that currently can’t really be proven. It
is great that the MSI Claw will be Wi-Fi 7 capable. Even if the claims are a bit dubious,
it stands to be a big improvement over Wi-Fi 6E which itself has been impressive. It’s got
increased bandwidth, lower latency, the ability to spread across multiple frequencies, and the
ability to connect to even more devices.
gets to tack on for free without anyone knowing
if this chip is a good Wi-Fi 7 chip or not.
Again, I don’t think there’s anything that should
have been different here, I just think this is something that’s worth being aware of.
One of the biggest marketing points about the Core Ultra 7 155H (and Meteor Lake in general)
are these NPUs, or Neural Processing Units, that are onboard. NPUs are focused on AI
workloads and as of right now, it feels like a solution in search of a problem. But that
hasn’t stopped Intel and MSI from advertising the AI-capabilities of Meteor Lake chips and the
MSI Claw is no exception. Several outlets, including PCWorld have even described this
as the first handheld with an NPU.
Additionally, one of MSI’s bullet points
for the Claw is their “MSI AI Engine adjusts various system settings automatically
that best fit your needs.” and I don’t know a single preview that covers what this
means or what it can actually do.
When an Intel rep was actually asked about
how NPUs could actually enhance the gaming experience on a handheld, he starts off
with some more grounded use cases like facial recognition or gesture recognition but he
eventually goes onto describe a system that takes shitty animation as an input and outputs awesome
animation. No really - have a listen.
That’s right, Epic - you’re welcome. The reason
Tommy here can get away with describing this pie-in-the-sky use case unchecked is because,
as he says, we’re still in the early days of discovering how NPUs will be helpful. So without
actually knowing how NPUs will be helpful, Tommy Boy can just make something up and say,
wouldn’t it be wonderful if we fed something pixel art and out came high poly 3d models? Tom was
asked how NPUs could be helpful in the MSI Claw, and he took that question, and remixed it.
Instead of answering the original question, he answered a question of how AI
could be helpful in gaming one day.
Another tool that is utilized here to receive
the best possible coverage is the concept of multiple models. This gives MSI the best of both
worlds because they’re strongly advertising the lowest price of $699 while strongly focusing
on the benefits of the Core Ultra 7 that you don’t get in that lowest model. You need
to cough up at least an extra $50 to get the performance being showcased on the show
floor. Personally, I don’t find that to be a small difference in price specifically because
$699 is the price of the ROG Ally and the Lenovo Legion Go. The Core Ultra 5 version of the MSI
Claw will not compete with these handhelds and so MSI loses price parity with the Ally and
the Go despite being a direct competitor.
Journalists have done a great job of covering
this discrepancy but it’s unclear if the Core Ultra 5 version of the Claw will even be a viable
model or if it exists simply so that MSI can say starting at 699. This happened with the ROG
Ally as well. They heavily advertised that the ROG Ally started at 599 because you could
either buy the Z1 Extreme model or eventually, you could purchase the ROG Ally with the base Z1
APU. Unfortunately the disparity in performance between the base Z1 and the Z1 Extreme is so
large that it effectively makes the $599 ROG Ally useless to the PC handheld market and I
have some concerns that the Core Ultra 5 MSI Claw will suffer the same fate. What do you
think? Am I being too hard on this APU here?
One more tool in MSI’s toolbox is an obvious
one: they were able to make big claims that went unchecked because the journalists covering
the Claw didn’t have access to things that would help them prove or disprove these claims. Even
the performance overlay didn’t show the FPS. They made a conscious decision to not let journalists
see the framerate and I’m not gonna sit here and say it’s because they were trying to hide bad
performance; that would be a big leap.
I think the basic reasoning for not having
framerate available to measure is that: drivers are not final, firmware is not final,
even the VRR spec is not final. In that way, displaying the framerate would likely cause more
trouble than it’s worth, so I get it. I also have to give MSI a lot of credit for not making really
big claims about the performance of the MSI Claw. They made a lot of claims about battery life and
efficiency but not about the performance.
About those battery life claims. Like I said, I
watched every video I could find and I saw a lot of interesting claims. In the official marketing
materials, they said things like “Best-in-Class Battery Life” and “Power that lasts 50% longer”.
I think these claims were mostly drawn from comparisons to the ROG Ally but it’s intentionally
vague. Like wouldn’t “best in class battery life” imply that the MSI Claw is claiming to have even
more battery life than the Steam Deck OLED?
50% longer battery life is one I can’t wait
to test for myself. Compared to the ROG Ally, the battery in the MSI Claw is about 33% bigger
in terms of capacity. That means that if the Claw was exactly as efficient as the Ally (assuming
that’s the device they’re testing against) then the Claw would have 33% longer battery life,
simply based on the capacity. So for it to have 50% more battery life would mean that it’s
even more efficient than the Ally and that leaves me with so many questions. Like - at what TDP? Is
performance still similar to the Ally at that TDP or are you testing at such a low TDP that the Ally
has the advantage in terms of performance?
Then if you watch the videos, there was one number
that kept coming up: 2 hours minimum battery life. In Linus Tech Tips’ video, Linus said 2 hours
at full bore, 35W. But other videos said 2 hours while playing Forza Horizon 5 at 15W. The latter
seems to line up better with some of the stats I could manage to get a sneak peek of. For example,
in one video, I could see that at a TDP of 20W, the total system power was at 35W. That would give
you about 1.5 hours at 20W so you could see how 2 hours at 15W sounds plausible.
To be honest, if you take all of the above, I have to say that I think MSI managed to keep
their claims somewhat reasonable. They admitted that VRR may not make it in, they didn’t make a
big deal about Wi-Fi 7 or NPUs (although those are pet topics for Intel), and they were clear
about pricing. I couldn’t find any really boastful claims about performance and although I’m quite
dubious of the battery life claims, it does seem like 2 hours at 15W is at least plausible.
That all puts my mind at ease because while I do still worry about how well Intel can compete with
AMD in this field, I’d be lying if I said I wasn’t excited to see more and more handhelds. The fact
that Intel, AMD, ASUS, Lenovo, and MSI all see potential in this market is reassuring.
Nevertheless, I think it’s worthwhile to keep expectations in check. There’s a lot we won’t
know about the Claw until we have it in our hands and can test it without restriction. Let
me know what you think though. Are you more or less excited about the Claw than you were
before the previews? Deck gang out. Goodbye!