I have delayed my purchase for the next gen 300 series, but its only arriving in a smaller 12" laptops. Can that be instaled in a 16" or is 16" laptops geting their own amd 300 series mainborad?
After a quick search i havent found any info regarding this info.
If I had to guess, probably at the same time with new GPU modules, I believe AMD is announcing something sometime next month, so we’ll probably find out around that time or shortly after.
You got that one mixed up. The Ryzen AI 300 series is only available for the Framework Laptop 13. The 12" one is only going to have intel APUs as per announcement.
What framework did was a great lack of respect for the customer, the 16 laptop needed the 9955hx3d to at least announce it for April or June, the 7945hx, because the upgradeable philosophy makes a lot of sense in these computers with this range of processors
strix point with a maximum of 12 cores, which were already burned because it came out last year, they released it as the latest, that was a joke, and the people applauding it was a joke having an alliance with AMD and they did not release the 9955hx3d that would skyrocket sales, and very soon a hard graphics, and not what they did
About Strix Halo only on desktop PC hahahahahahahaha, is this a company or is it a circus?
I am also interested in this, as the discrete GPU, larger screen and greater number of expansion modules is very appealing.
I love my power-sipping Laptop 13 with the Intel Tiger Lake Core-i7, though am debating whether I should spend the money on a (recently-announced) AMD mainboard upgrade for it, versus a larger, newer chassis.
I wonder if they’re holding off on it until they can announce new GPUs for the laptop, I was just looking at benchmark, the Radeon 8050S in used in AMD Ryzen AI Max 385 seems to be identical to 7700S (though not sure about the TDP used for this).
If AMD Ryzen AI Max 385 was available, I’d have ordered it right away, even if it was capped at 45-55w TDP, it’s just about perfect for the usecase I have with FW 16, 7840HS + 780M works great, but having that huge jump in gaming performance without needing to grab 7700S would be nice.
I think there are other reasons they don’t want to bring the AI Max series to the FW16. AMD is currently requiring soldered, LPDDR5X memory with the AI Max APUs. Also, it only has 16 PCIe lanes, and only 5 high-speed USB ports natively. They may be able to work around this, but I don’t think it’s quite the perfect fit for the FW 16 that Framework would have liked it to be. They mentioned at the recent press event that they really liked the AI Max 395+, but going with soldered RAM in the FW 16 is just something they didn’t want to do. But they really wanted to utilize that chip in something, which is why they came up with the Framework desktop.
Even with the downsides, I would jump on an AI Max 395+ mainboard real fast.
XMG is on-track to have both a 14-inch and 15-inch Strix Point laptop with DDR5 SODIMM slots (upgradeable to 96GB) ready by this Summer 2025.
Upcoming XMG NEO 14 photos:
GizmotechSlip: Photos of the XMG EVO 14 E25! - Album on Imgur
Upcoming XMG NEO 15 photos:
GizmotechSlip: Photos of the brand new XMG EVO 15 E25! - Album on Imgur
General Specifications for both:
GizmotechSlip: Specs of the Brand New XMG EVO 14 E25 & EVO 15 E25! - Album on Imgur
And Moore’s Law is Dead recently leaked that Medusa Halo will sport 24-48CUs and that all upcoming Medusa APUs will have MALL Cache (kinda like how ReBAR allows CPU direct access to load into VRAM when running GPU workloads, this allows iGPU/NPU direct access to Infinity Cache) which combined could boost iGPU performance across the board by 35%-45%
Right now…
AMD:
~> 680M = RDNA 2.0 12CUs = HD 7970
~> 780M = RDNA 3.0 12CUs = R9 280X
~> 880M = RDNA 3.5 12CUs = RX 6400
~> 890M = RDNA 3.5 16CUs (32 ROPs) = RX 470
~> 8040S = RDNA 3.5 16CUs (40 ROPs) = RX 480/580
~> 8050S = RDNA 3.5 32CUs = RX 5700XT
~> 8060S = RDNA 3.5 40CUs = RX 7600/7600XT
Intel:
~> Arc 7-core (Meteor Lake) = Xe 112EUs = HD 7970
~> Arc 8-core (Meteor Lake) = Xe 128EUs = R9 280X
~> 130V (Lunar Lake) = Xe2 112EUs = R9 280X
~> 140T (Arrow Lake) = Xe+ 128EUs = RX 6400
~> 140V (Lunar Lake) = Xe2 128EUs = RX 570
Medusa Point (Q4 2026, hopefully DDR6 SODIMM support)…
~> “980M” = RDNA 4.0 12CUs = RX 480/580
~> “990M” = RDNA 4.0 16CUs = GTX 1660TI/1070/3050
Medusa Halo (Q4 2027, likely LPDDR6X soldered or LPCAMM2 via a 384-bit bus)…
~> “9040S” = RDNA 5.0 24CUs = RX 6600XT
~> “9050S” = RDNA 5.0 32CUs = RTX 2080TI/4060TI
~> “9060S” = RDNA 5.0 48CUs = RX 7900GRE
IMO RTX 5080 laptops are going depreciate like cheap wine 2 years from now if laptops sporting 7900GRE-class iGPUs with up to 144GB of LPDDR6X-15000 accessible as VRAM become widely available by Q4 2027 – assuming AMD ROCm/HIP catches up to CUDA as fast as they have with FSR4 against DLSS 4.0 
P.S.A. if you’re in the budget to upgrade your laptop to something powered by an APU and iGPU within ballpark of the RX 470-480 (Ryzen 370/375/380 or Core Ultra 7/9 200V) or RX 5700XT-6700 (Ryzen 385/390/395), I would do it soon! Tariffs will start affecting prices more heavily towards the end of this year and Zen6 APUs will not land on shelves for another 18-36 months.
HX/HS mobile workstation CPUs (minimal to no onboard graphics, higher TDP and core count, more connectivity for dGPUs and similar) tend to refresh around June/July. The existing stuff announced and out is all competing with Intel’s U and P series.
If the FW16 went to the Medusa options they would only have 16x total pcie lanes to play around with vs the current gen’s 20x (that’s a lot less to feed USB4 TB ports, dual nvme, etc).