{"id":1280546,"date":"2026-04-27T10:33:49","date_gmt":"2026-04-27T10:33:49","guid":{"rendered":"https:\/\/gpdstore.net\/blog\/gpd-g2%e3%81%a8gpd-box%ef%bc%9amcio%e3%81%aeegpu%e6%a9%9f%e8%83%bd%e3%81%a8gpd%e3%81%ae%e5%88%9d%e4%bb%a3%e3%83%9f%e3%83%8bpc%e3%81%ae%e7%b5%84%e3%81%bf%e5%90%88%e3%82%8f%e3%81%9b\/"},"modified":"2026-05-06T14:21:18","modified_gmt":"2026-05-06T14:21:18","slug":"gpd-g2-and-gpd-box-announced","status":"publish","type":"post","link":"https:\/\/gpdstore.net\/ja\/blog\/gpd-g2-and-gpd-box-announced\/","title":{"rendered":"GPD G2\u3068GPD BOX\uff1aMCIO\u306eeGPU\u6a5f\u80fd\u3068GPD\u306e\u521d\u4ee3\u30df\u30cbPC\u306e\u7d44\u307f\u5408\u308f\u305b"},"content":{"rendered":"\n

What are the GPD G2 and GPD BOX?<\/strong> The GPD G2<\/a> is a newly announced MCIO and USB4 v2.0 eGPU dock, while the GPD BOX<\/a> is a Panther Lake-based compact desktop that GPD is presenting as its first mini PC. Together, they form a new hardware pairing built around higher-bandwidth external graphics expansion and a very small desktop footprint.<\/p>\n\n

With the latest official spec now available, the picture is much clearer than it was at teaser stage. GPD is not only introducing a mini PC and an eGPU dock at the same time, it is also laying out a broader platform idea in which the BOX acts as the compact base system and the GPD G2<\/a> acts as the higher-power graphics and expansion layer.<\/p>\n\n

\u3053\u308c\u304c\u3001\u4eca\u56de\u306e\u767a\u8868\u3092\u5358\u306a\u308b\u88fd\u54c1\u767a\u8868\u4ee5\u4e0a\u306b\u8208\u5473\u6df1\u3044\u3082\u306e\u306b\u3057\u3066\u3044\u308b\u3002GPD\u306f\u3001\u30df\u30cbPC\u3068\u30c9\u30c3\u30af\u3092\u5225\u500b\u306e\u30a2\u30af\u30bb\u30b5\u30ea\u30fc\u3068\u3057\u3066\u6271\u3046\u306e\u3067\u306f\u306a\u304f\u3001\u3088\u308a\u3059\u3063\u304d\u308a\u3068\u3057\u305f\u30c7\u30b9\u30af\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u3001\u3088\u308a\u5c0f\u578b\u306e\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u3001\u5fc5\u8981\u306a\u3068\u304d\u306b\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u30d1\u30ef\u30fc\u3092\u62e1\u5f35\u3067\u304d\u308b\u4f59\u5730\u3092\u6c42\u3081\u308b\u30e6\u30fc\u30b6\u30fc\u306e\u305f\u3081\u306b\u3001\u305d\u308c\u3089\u3092\u30de\u30c3\u30c1\u3057\u305f\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u30b9\u30bf\u30c3\u30af\u3068\u3057\u3066\u63d0\u793a\u3057\u3066\u3044\u308b\u3002\u3059\u3067\u306b\u30b3\u30f3\u30d1\u30af\u30c8\u306aGPD\u30b7\u30b9\u30c6\u30e0\u3001\u30d7\u30ec\u30df\u30a2\u30e0\u30cf\u30f3\u30c9\u30d8\u30eb\u30c9\u96a3\u63a5\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u3001\u307e\u305f\u306f\u67d4\u8edf\u306a\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u4ee3\u66ff\u54c1\u3092\u95b2\u89a7\u3057\u3066\u3044\u308b\u4eba\u306b\u3068\u3063\u3066\u3001\u3053\u306e\u7d44\u307f\u5408\u308f\u305b\u306f\u660e\u3089\u304b\u306b\u3001\u30df\u30cb\u30ce\u30fc\u30c8PC\u3068<\/a>\u5e45\u5e83\u3044\u30cf\u30f3\u30c9\u30d8\u30eb\u30c9\u30b2\u30fc\u30df\u30f3\u30b0PC\u306e<\/a>\u9593\u306b\u65b0\u3057\u3044\u30ec\u30fc\u30f3\u3092\u958b\u304f\u3053\u3068\u3092\u610f\u5473\u3057\u3066\u3044\u308b\u3002 <\/p>\n\n

\u307e\u3060\u91cd\u8981\u306a\u672a\u77e5\u6570\u304c\u3042\u308b\u3002GPD\u306f\u307e\u3060\u4fa1\u683c\u3001\u767a\u58f2\u6642\u671f\u3001\u5730\u57df\u5c55\u958b\u306e\u8a73\u7d30\u3092\u78ba\u5b9a\u3057\u3066\u304a\u3089\u305a\u3001\u898b\u51fa\u3057\u306e\u30b2\u30fc\u30e0\u3068\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u306e\u8b33\u3044\u6587\u53e5\u306e\u3044\u304f\u3064\u304b\u306f\u3001\u307e\u3060\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306e\u6587\u8108\u3092\u5fc5\u8981\u3068\u3057\u3066\u3044\u308b\u3002\u3057\u304b\u3057\u3001\u6700\u521d\u306e\u30c6\u30a3\u30fc\u30b6\u30fc\u3068\u6bd4\u8f03\u3059\u308b\u3068\u3001\u4e21\u88fd\u54c1\u306f\u306f\u308b\u304b\u306b\u5b9f\u8cea\u7684\u306a\u88cf\u4ed8\u3051\u3092\u6301\u3064\u3088\u3046\u306b\u306a\u3063\u305f\u3002 <\/p>\n\n

\u88fd\u54c1<\/th>\u65b0\u305f\u306b\u6c7a\u5b9a\u3057\u305f\u30cf\u30a4\u30e9\u30a4\u30c8<\/th>\u306a\u305c\u76ee\u7acb\u3064\u306e\u304b<\/th><\/tr><\/thead>
GPD G2<\/td>PCIe Gen 5 x16\u30b9\u30ed\u30c3\u30c8\u3001PCIe Gen 4 x8\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30c7\u30d0\u30a4\u30b9\u7528MCIO 8i\u30b5\u30dd\u30fc\u30c8\u3001USB4 v2.0\u3001USB-A x2\u3001RJ45\u30a4\u30fc\u30b5\u30cd\u30c3\u30c8\u3001M.2 2280\u30b9\u30ed\u30c3\u30c8\u3001\u6700\u5927100W PD\u3001\u30b4\u30fc\u30eb\u30c9\u5b9a\u683cATX 3.1\u8a2d\u8a08\u3001800W\u9023\u7d9a12V\u51fa\u529b<\/td>\u5358\u306a\u308b\u30c9\u30c3\u30af\u3067\u306f\u306a\u304f\u3001\u672c\u683c\u7684\u306a\u5916\u90e8\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u3068\u62e1\u5f35\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u306e\u3088\u3046\u3060<\/td><\/tr>
GPD BOX<\/td>GPD\u521d\u306e\u30df\u30cbPC\u3001\u30d1\u30f3\u30b5\u30fc\u30ec\u30a4\u30afCore Ultra X7 358H\u307e\u305f\u306fCore Ultra 7 356H\u3001Xe3\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u3001\u6700\u592764GB LPDDR5x\u3001\u30c7\u30e5\u30a2\u30ebM.2\u30b9\u30c8\u30ec\u30fc\u30b8\u3001MCIO 8i\u3001\u30c7\u30e5\u30a2\u30ebUSB4 v2.0\u3001DP 2.1\u3001HDMI 2.1\u3001\u30c7\u30e5\u30a2\u30eb2.5G LAN<\/td>\u6700\u521d\u306e\u767a\u8868\u3067\u793a\u5506\u3055\u308c\u305f\u3088\u308a\u3082\u306f\u308b\u304b\u306b\u7a4d\u6975\u7684\u306a\u30b9\u30da\u30c3\u30af\u30b7\u30fc\u30c8\u3067\u3001GPD\u3092\u30b3\u30f3\u30d1\u30af\u30c8\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u30b9\u30da\u30fc\u30b9\u306b\u62bc\u3057\u4e0a\u3052\u308b\u3002<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

GPD\u306eG2\u306f\u3001\u30c6\u30a3\u30fc\u30b6\u30fc\u304c\u793a\u5506\u3057\u3066\u3044\u305f\u3088\u308a\u3082\u306f\u308b\u304b\u306b\u91ce\u5fc3\u7684\u306b\u898b\u3048\u308b\u3002<\/h2>\n\n

The biggest upgrade in clarity comes from the new GPD G2<\/a> tech specs page. GPD had already described the unit as the world’s first MCIO dual-port eGPU dock, but the full page makes it look much more like a complete graphics and expansion chassis than a basic external GPU accessory.<\/p>\n\n

\"GPD<\/a>
GPD G2\u3068eGPU\u3092\u642d\u8f09\u3057\u305fGPD BOX<\/figcaption><\/figure>\n\n

According to GPD’s published specifications, the dock includes an internal PCIe Gen 5 x16 slot and an MCIO 8i connection that supports PCIe devices, including graphics cards, through a PCIe Gen 4 x8 link. GPD also lists one USB4 v2.0 port with 80 Gbps symmetric or 120\/40 Gbps asymmetric bandwidth support, along with PD 3.0 charging up to 100W. On top of that, the GPD G2<\/a> includes two USB 3.2 Gen 2 Type-A ports, one RJ45 Gigabit Ethernet port, and one M.2 2280 slot that supports external storage booting.<\/p>\n\n

That matters because it gives the GPD G2<\/a> a much broader identity. This is not being marketed only as a way to attach a desktop GPU to one handheld or one laptop. GPD is presenting it as a flexible external expansion platform for home PCs, high-end handhelds with Thunderbolt or USB4 class connectivity, workstations, servers, and even some HPC-style scenarios. It also explicitly lists both Windows 11 and Linux support.<\/p>\n\n

\"GPD<\/a>
GPD G2 \u30b5\u30a4\u30c9\u30d3\u30e5\u30fc<\/figcaption><\/figure>\n\n

Power delivery is another major part of the updated story. GPD says the GPD G2<\/a> uses a Gold-rated ATX 3.1 design, complies with ATX 3.1 GPU requirements, supports 12V-2×6 GPU power, and can provide 800W of continuous total 12V output. That is a much more serious power profile than most readers would assume from the original teaser graphics.<\/p>\n\n

GPD\u306f\u307e\u305f\u3001GPU\u306e\u4e92\u63db\u6027\u306b\u95a2\u3057\u3066\u3082\u3001\u6700\u521d\u306e\u30de\u30fc\u30b1\u30c6\u30a3\u30f3\u30b0\u30fb\u30a4\u30e1\u30fc\u30b8\u304b\u3089\u60f3\u50cf\u3055\u308c\u308b\u3088\u308a\u3082\u306f\u308b\u304b\u306b\u5e83\u7bc4\u56f2\u3092\u30ab\u30d0\u30fc\u3057\u3066\u3044\u308b\u3002\u516c\u5f0f\u30da\u30fc\u30b8\u306b\u306f\u3001NVIDIA GeForce GTX 10\u3001GTX 16\u3001RTX 20\u3001RTX 30\u3001RTX 40\u3001\u304a\u3088\u3073RTX 50\u30b3\u30f3\u30b7\u30e5\u30fc\u30de\u30fc\u30ab\u30fc\u30c9\u3001AMD Radeon RX 400\u304b\u3089RX 9000\u30b7\u30ea\u30fc\u30ba\u30b3\u30f3\u30b7\u30e5\u30fc\u30de\u30fc\u30ab\u30fc\u30c9\u3001\u304a\u3088\u3073Intel Arc A\u304a\u3088\u3073B\u30b7\u30ea\u30fc\u30ba\u30b3\u30f3\u30b7\u30e5\u30fc\u30de\u30fcGPU\u306e\u30b5\u30dd\u30fc\u30c8\u30ea\u30b9\u30c8\u304c\u63b2\u8f09\u3055\u308c\u3066\u3044\u308b\u3002\u307e\u305f\u3001NVIDIA Quadro\u304a\u3088\u3073RTX\u30ef\u30fc\u30af\u30b9\u30c6\u30fc\u30b7\u30e7\u30f3\u30fb\u30ab\u30fc\u30c9\u3001AMD Radeon Pro\u30e9\u30a4\u30f3\u3001Intel Arc Pro GPU\u3001\u3055\u3089\u306b\u306fNVIDIA\u3001AMD\u3001Intel\u306e\u5404\u7a2e\u30c7\u30fc\u30bf\u30bb\u30f3\u30bf\u30fc\u30fb\u30a2\u30af\u30bb\u30e9\u30ec\u30fc\u30bf\u306e\u30b5\u30dd\u30fc\u30c8\u30fb\u30ea\u30b9\u30c8\u3082\u516c\u958b\u3055\u308c\u3066\u304a\u308a\u3001\u30b3\u30f3\u30b7\u30e5\u30fc\u30de\u30fb\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306e\u67a0\u3092\u8d85\u3048\u3066\u3044\u307e\u3059\u3002 <\/p>\n\n

\"eGPU搭載GPD<\/a>
eGPU\u642d\u8f09GPD G2<\/figcaption><\/figure>\n\n

The headline claim is still the same one that grabbed attention first: GPD says an RTX 4090 can run on the GPD G2<\/a> with only a 2% performance loss. The updated material also broadens the message by framing the dock around 4090 or 5090-class use cases. Those claims are still best treated as GPD positioning rather than settled fact, because the company has not published the test methodology, games, resolutions, or benchmark conditions needed to verify them independently.<\/p>\n\n

Even with that caveat, the GPD G2<\/a> is no longer just an intriguing idea. On the current evidence, it looks like one of GPD’s most aggressive attempts yet to bridge compact devices and desktop-class graphics expansion.<\/p>\n\n

GPD G2\u306e\u6280\u8853\u4ed5\u69d8<\/h2>\n\n
\u30ab\u30c6\u30b4\u30ea\u30fc<\/strong><\/td>\u7279\u5fb4<\/strong><\/td>\u4ed5\u69d8<\/strong><\/td><\/tr><\/thead>
\u57fa\u672c\u60c5\u5831<\/strong><\/td>\u30d6\u30e9\u30f3\u30c9<\/strong><\/td>\u30b8\u30fc\u30d4\u30fc\u30c7\u30a3\u30fc<\/td><\/tr>
<\/td>\u5bfe\u5fdc\u30c7\u30d0\u30a4\u30b9<\/strong><\/td>\u30db\u30fc\u30e0PC\u304a\u3088\u3073\u30cf\u30a4\u30a8\u30f3\u30c9\u30b2\u30fc\u30e0\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\uff08PCIe 3.0 x16\u30b9\u30ed\u30c3\u30c8\u304c1\u3064\u4ee5\u4e0a\u7a7a\u3044\u3066\u3044\u308b\uff09\u3001\u30cf\u30a4\u30a8\u30f3\u30c9\u30b2\u30fc\u30e0\u30cf\u30f3\u30c9\u30d8\u30eb\u30c9\uff08Thunderbolt 3\/4\/5\u307e\u305f\u306fUSB4\/USB4 v2.0\u30dd\u30fc\u30c8\u3092\u30b5\u30dd\u30fc\u30c8\uff09\u3001\u30ef\u30fc\u30af\u30b9\u30c6\u30fc\u30b7\u30e7\u30f3\u3001\u30b5\u30fc\u30d0\u30fc\u3001\u9ad8\u6027\u80fd\u30b3\u30f3\u30d4\u30e5\u30fc\u30c6\u30a3\u30f3\u30b0\uff08HPC\uff09\u30b7\u30ca\u30ea\u30aa\u3002<\/td><\/tr>
<\/td>\u30b7\u30b9\u30c6\u30e0\u30b5\u30dd\u30fc\u30c8<\/strong><\/td>Windows 11\u304a\u3088\u3073Linux\u30c7\u30a3\u30b9\u30c8\u30ea\u30d3\u30e5\u30fc\u30b7\u30e7\u30f3\u306b\u5bfe\u5fdc\u3002<\/td><\/tr>
\u5bfe\u5fdcGPU<\/strong><\/td>\u30b3\u30f3\u30b7\u30e5\u30fc\u30de\u30fc\u5411\u3051GPU<\/strong><\/td>–NVIDIA<\/strong>GeForce GTX 10 \u30b7\u30ea\u30fc\u30ba (Pascal)\u3001RTX 20 \u30b7\u30ea\u30fc\u30ba (Turing)\u3001GTX 16 \u30b7\u30ea\u30fc\u30ba (Turing)\u3001RTX 30 \u30b7\u30ea\u30fc\u30ba (Ampere)\u3001RTX 40 \u30b7\u30ea\u30fc\u30ba (Ada Lovelace)\u3001RTX 50 \u30b7\u30ea\u30fc\u30ba (Blackwell) PCIe \u30d9\u30fc\u30b9\u306e\u30ab\u30fc\u30c9\u3002
–AMD\uff1a<\/strong>Radeon RX 400\/500\u30b7\u30ea\u30fc\u30ba\uff08GCN 4.0\uff09\u3001RX Vega\u30b7\u30ea\u30fc\u30ba\uff08GCN 5.0\uff09\u3001Radeon VII\uff08GCN 5.0\uff09\u3001RX 5000\u30b7\u30ea\u30fc\u30ba\uff08RDNA 1.0\uff09\u3001RX 6000\u30b7\u30ea\u30fc\u30ba\uff08RDNA 2.0\uff09\u3001RX 7000\u30b7\u30ea\u30fc\u30ba\uff08RDNA 3.0\uff09\u3001RX 9000\u30b7\u30ea\u30fc\u30ba\uff08RDNA 4.0\uff09 PCIe\u30d9\u30fc\u30b9\u30ab\u30fc\u30c9\u3002
–Intel\uff1a<\/strong>Arc A \u30b7\u30ea\u30fc\u30ba (Xe HPG) \u304a\u3088\u3073 Arc B \u30b7\u30ea\u30fc\u30ba (Xe2 HPG) PCIe \u30d9\u30fc\u30b9\u30ab\u30fc\u30c9\u3002<\/td><\/tr>
<\/td>\u30ef\u30fc\u30af\u30b9\u30c6\u30fc\u30b7\u30e7\u30f3GPU<\/strong><\/td>–NVIDIA\uff1a<\/strong>Quadro K \u30b7\u30ea\u30fc\u30ba (Kepler)\u3001M \u30b7\u30ea\u30fc\u30ba (Maxwell)\u3001P \u30b7\u30ea\u30fc\u30ba (Pascal)\u3001A \u30b7\u30ea\u30fc\u30ba (Ampere)\u3001\u304a\u3088\u3073 RTX \u30b7\u30ea\u30fc\u30ba (Blackwell) PCIe \u30d9\u30fc\u30b9\u306e\u30ab\u30fc\u30c9\u3002
–AMD\uff1a<\/strong>Radeon Pro WX \u30b7\u30ea\u30fc\u30ba (GCN)\u3001Pro W \u30b7\u30ea\u30fc\u30ba (GCN)\u3001Pro W5700 (RDNA)\u3001Pro W6000 \u30b7\u30ea\u30fc\u30ba (RDNA 2)\u3001Pro W7000 \u30b7\u30ea\u30fc\u30ba (RDNA 3)\u3001\u304a\u3088\u3073 AI PRO R9000 \u30b7\u30ea\u30fc\u30ba (RDNA 4) PCIe \u30d9\u30fc\u30b9\u306e\u30ab\u30fc\u30c9\u3002
–Intel\uff1a<\/strong>Arc Pro A \u30b7\u30ea\u30fc\u30ba (Xe) \u304a\u3088\u3073 Arc Pro B \u30b7\u30ea\u30fc\u30ba (Xe2) PCIe \u30d9\u30fc\u30b9\u30ab\u30fc\u30c9\u3002<\/td><\/tr>
<\/td>\u30c7\u30fc\u30bf\u30bb\u30f3\u30bf\u30fc\u5411\u3051GPU<\/strong><\/td>– NVIDIA\uff1a<\/strong>Tesla\u30b7\u30ea\u30fc\u30ba\uff08Kepler\u3001Maxwell\u3001Pascal\u3001Volta\uff09\u3001T4\uff08Turing\uff09\u3001A100\uff08Ampere\uff09\u3001H100\/H200\uff08Hopper\uff09\u3001\u304a\u3088\u3073\u305d\u308c\u4ee5\u964d\u306e\u88fd\u54c1\u3092\u542b\u3080\u30c7\u30fc\u30bf\u30bb\u30f3\u30bf\u30fc\u5411\u3051PCIe GPU\u3002
– AMD\uff1a<\/strong>\u30a4\u30f3\u30b9\u30c6\u30a3\u30f3\u30af\u30c8PCIe\u30b3\u30f3\u30d4\u30e5\u30fc\u30c8\u30fb\u30ab\u30fc\u30c9\uff08MI8\uff08GCN 3.0\uff09\u3001MI6\uff08GCN 4.0\uff09\u3001MI25\uff08GCN 5.0\uff09\u3001MI100\uff08CDNA 1.0\uff09\u3001MI200\u30b7\u30ea\u30fc\u30ba\uff08CDNA 2.0\uff09\u3001MI300\u30b7\u30ea\u30fc\u30ba\uff08CDNA 3.0\uff09\u3001\u304a\u3088\u3073\u305d\u308c\u4ee5\u964d\u306e\u88fd\u54c1\u3092\u542b\u3080\uff09
–\u30a4\u30f3\u30c6\u30eb\uff1a<\/strong>Max \u30b7\u30ea\u30fc\u30ba (Xe-HPC) \u304a\u3088\u3073 Flex \u30b7\u30ea\u30fc\u30ba (Xe-HPG) \u3092\u542b\u3080\u30c7\u30fc\u30bf\u30bb\u30f3\u30bf\u30fc\u5411\u3051 PCIe GPU\u3002<\/td><\/tr>
\u30b3\u30cd\u30af\u30c6\u30a3\u30d3\u30c6\u30a3<\/strong><\/td>PCIe\u30b9\u30ed\u30c3\u30c8<\/strong><\/td>\u00d7\u53cc\u65b9\u5411\u7406\u8ad6\u5e2f\u57df\u5e451024 Gbps\u3001\u53cc\u65b9\u5411\u5b9f\u52b9\u30c7\u30fc\u30bf\u5e2f\u57df\u5e45\u7d04126 GB\/\u79d2\u3002 <\/td><\/tr>
<\/td>MCIO 8i\u30dd\u30fc\u30c8<\/strong><\/td>PCIe Gen 4 \u00d7 8\u7d4c\u7531\u3067\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30ab\u30fc\u30c9\u3092\u542b\u3080PCIe\u30c7\u30d0\u30a4\u30b9\u3092\u30b5\u30dd\u30fc\u30c8\u3002\u30cf\u30d6\u6a5f\u80fd\u3092\u6709\u52b9\u306b\u3059\u308b\u306b\u306f\u3001USB4 v2.0\u30dd\u30fc\u30c8\u3078\u306e\u540c\u6642\u63a5\u7d9a\u304c\u5fc5\u8981\u3067\u3059\u3002 <\/td><\/tr>
<\/td>USB4 v2.0\u30dd\u30fc\u30c8<\/strong><\/td>\u00d780Gbps\u5bfe\u79f0\u5e2f\u57df\u5e45\uff08TX\/RX\uff09\u307e\u305f\u306f120\/40Gbps\u975e\u5bfe\u79f0\u5e2f\u57df\u5e45\uff08TX\/RX\uff09\u3092\u30b5\u30dd\u30fc\u30c8\u3002\u6700\u5927100W\u306e\u96fb\u529b\u4f9b\u7d66\uff085V\/9V\/12V\/15V\/20V\uff09\u3067PD 3.0\u30d7\u30ed\u30c8\u30b3\u30eb\u3092\u30b5\u30dd\u30fc\u30c8\u3002MCIO 8i\u30dd\u30fc\u30c8\u3068\u4e00\u7dd2\u306b\u63a5\u7d9a\u3057\u305f\u5834\u5408\u306f\u3001\u30cf\u30d6\u306e\u62e1\u5f35\u3068PD\u96fb\u6e90\u306e\u307f\u3092\u51e6\u7406\u3057\u3001\u5358\u72ec\u3067\u63a5\u7d9a\u3057\u305f\u5834\u5408\u306f\u3001eGPU\u6a5f\u80fd\u3092\u6709\u52b9\u306b\u3057\u307e\u3059\u3002 <\/td><\/tr>
<\/td>USB\u30bf\u30a4\u30d7A<\/strong><\/td>\u00d72\u3001USB 3.2 Gen 2\u300110 Gbps\u3002<\/td><\/tr>
<\/td>NVMe M.2\u30b9\u30ed\u30c3\u30c8 <\/strong><\/td>\u00d71\u3001M.2 2280\u4ed5\u69d8\u3001PCIe 3.0 \u00d7 2\uff08USB 3.2 Gen 2\u304b\u3089\u5909\u63db\uff09\u3002\u5916\u4ed8\u3051\u30b9\u30c8\u30ec\u30fc\u30b8\u306e\u8d77\u52d5\u306b\u5bfe\u5fdc\u3002\u30ac\u30f3\u30e1\u30bf\u30eb\u306e\u30de\u30b0\u30cd\u30c3\u30c8\u5f0fSSD\u30ab\u30d0\u30fc\u3092\u63a1\u7528\u3002 <\/td><\/tr>
<\/td>\u30a4\u30fc\u30b5\u30cd\u30c3\u30c8\u30fb\u30dd\u30fc\u30c8<\/strong><\/td>\u00d71\u3001RJ45\u30a4\u30fc\u30b5\u30cd\u30c3\u30c8\u30dd\u30fc\u30c8\u300110\/100\/1000 Mbps\u30aa\u30fc\u30c8\u30cd\u30b4\u30b7\u30a8\u30fc\u30b7\u30e7\u30f3\u3002<\/td><\/tr>
\u30d1\u30ef\u30fc\u30b9\u30da\u30c3\u30af<\/strong><\/td>\u30bf\u30a4\u30d7<\/strong><\/td>ATX 3.1 GPU\u8981\u4ef6\u6e96\u62e0\u3001\u52b9\u738792%\u4ee5\u4e0a\u3002<\/td><\/tr>
<\/td>\u30a4\u30f3\u30d7\u30c3\u30c8<\/strong><\/td>100-240V\u300150\/60Hz\u3002<\/td><\/tr>
<\/td>\u7dcf\u5408\u51fa\u529b<\/strong><\/td>12V-2\u00d76\u3001\u5408\u8a0812V\u3001800W\u9023\u7d9a\u51fa\u529b\u5bfe\u5fdc\u3002<\/td><\/tr>
<\/td>\u96fb\u6e90\u30b9\u30a4\u30c3\u30c1<\/strong><\/td>\u767d\u8272LED\u30a4\u30f3\u30b8\u30b1\u30fc\u30bf\u30fc\uff08ON\uff1a\u70b9\u706f\uff0fOFF\uff1a\u6d88\u706f\uff09\u3002<\/td><\/tr>
<\/td>GPU\u96fb\u6e90\u30b3\u30cd\u30af\u30bf<\/strong><\/td>12V-2×6\u3002<\/td><\/tr>
<\/td>AC\u5165\u529b\u30bd\u30b1\u30c3\u30c8<\/strong><\/td>IEC 60320 C14\u30bd\u30b1\u30c3\u30c8\u3002<\/td><\/tr>
\u7d20\u6750\u3068\u4ed5\u4e0a\u3052<\/strong><\/td>\u5bf8\u6cd5<\/strong><\/td>157.3 \u00d7 119.8 \u00d7 182 mm<\/td><\/tr>
<\/td>\u30ab\u30e9\u30fc<\/strong><\/td>\u30ac\u30f3\u30e1\u30bf\u30eb<\/td><\/tr>
<\/td>\u6750\u6599<\/strong><\/td>\u30a2\u30eb\u30df\u30cb\u30a6\u30e0\u30fb\u30de\u30b0\u30cd\u30b7\u30a6\u30e0\u5408\u91d1\u3001PC 94V-0<\/td><\/tr>
<\/td>\u6b63\u5473\u91cd\u91cf<\/strong><\/td>1605 g<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

GPD BOX\u306b\u3088\u308a\u3001GPD\u306f\u30df\u30cbPC\u3078\u306e\u672c\u683c\u7684\u306a\u7b2c\u4e00\u6b69\u3092\u8e0f\u307f\u51fa\u3057\u305f<\/h2>\n\n

On the desktop side, GPD is using the GPD BOX<\/a> to enter a category it has largely avoided until now. The company is positioning it as its first mini PC, and the fuller official product page makes it easier to see why it thinks this is not just another tiny office box.<\/p>\n\n

\"GPD<\/a>
GPD BOX\u5bf8\u6cd5<\/figcaption><\/figure>\n\n

The processor lineup is now clearly defined. GPD lists two Intel Panther Lake parts, the Intel Core Ultra X7 358H and the Intel Core Ultra 7 356H, both in Intel’s Core Ultra Series 3 family. Memory options are listed at 32GB or 64GB of LPDDR5x 8533 MT\/s, which gives the GPD BOX<\/a> a more premium profile than a bargain mini PC.<\/p>\n\n

Storage and expansion are also stronger than average. GPD lists two M.2 slots, one PCIe Gen 5 x4 NVMe slot and one PCIe Gen 4 x2 slot, with 2280 SSD options from 512GB to 4TB. The BOX also includes an MCIO 8i to 8i port based on PCIe Gen 5 x8. In its own marketing image, GPD calls the BOX a world-first MCIO 8i design, quotes 512 Gbps bidirectional bandwidth, and says the MCIO 8i link offers four times the bandwidth of OCuLink based on PCIe 4.0 x4. That is very much part of how it wants the GPD BOX<\/a> to be perceived.<\/p>\n\n

\"GPD<\/a>
GPD BOX \u6b63\u9762\u56f3<\/figcaption><\/figure>\n\n

The I\/O layout is also notably rich. According to GPD’s official page, the GPD BOX<\/a> includes two USB4 v2.0 ports, one DisplayPort 2.1 output, one HDMI 2.1 port, four USB 3.2 Gen 2 Type-A ports, and two 2.5G LAN ports. GPD also describes the dual USB4 v2.0 setup in marketing material as 160 Gbps bidirectional bandwidth, which reinforces the idea that the BOX is being positioned around expansion and connectivity just as much as around raw CPU specifications.<\/p>\n\n

All of that helps explain GPD’s core pitch. The company is effectively saying that the GPD BOX<\/a> can serve as a compact desktop base, while the GPD G2<\/a> can extend it into a much more powerful gaming or creator-oriented machine. In other words, the BOX is not being sold as an isolated tiny PC, but as half of a modular performance concept.<\/p>\n\n

\"GPD<\/a>
GPD BOX\u80cc\u9762\u56f3<\/figcaption><\/figure>\n\n

\u3053\u308c\u306b\u3088\u308a\u3001BOX\u306f\u3088\u308a\u5e45\u5e83\u3044GPD\u88fd\u54c1\u7fa4\u306e\u4e2d\u3067\u660e\u78ba\u306a\u4f4d\u7f6e\u3092\u5360\u3081\u308b\u3053\u3068\u306b\u306a\u308b\u3002GPD WIN 5\u306e<\/a>\u3088\u3046\u306a\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u304c\u597d\u304d\u306a\u30d0\u30a4\u30e4\u30fc\u306b\u3068\u3063\u3066\u3001BOX\u306f\u3001\u3088\u308a\u30c7\u30b9\u30af\u306b\u8fd1\u3044\u30d5\u30a9\u30fc\u30e0\u30d5\u30a1\u30af\u30bf\u30fc\u3067\u3001\u540c\u3058\u30b3\u30f3\u30d1\u30af\u30c8\u306a\u6027\u80fd\u306e\u30a8\u30c8\u30b9\u306b\u5411\u304b\u3063\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b\u3002 <\/p>\n\n

GPD BOX\u306e\u6280\u8853\u4ed5\u69d8<\/h2>\n\n
\u30ab\u30c6\u30b4\u30ea\u30fc<\/strong><\/td>GPD BOX\uff08\u30a6\u30eb\u30c8\u30e9X7 358H\uff09<\/strong><\/td>GPD BOX\uff08\u30a6\u30eb\u30c8\u30e9\u30bb\u30d6\u30f3 356H\uff09<\/strong><\/td><\/tr><\/thead>
CPU\u3068iGPU<\/strong><\/td>\u30a4\u30f3\u30c6\u30eb\u00ae Core\u2122 Ultra X7 358H<\/strong>
– 16 \u30b3\u30a2\uff5c\u6700\u5927 4.8 GHz \u30bf\u30fc\u30dc\u30fb\u30d6\u30fc\u30b9\u30c8
– 12 Xe3 \u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u30fb\u30b3\u30a2\uff5c180 TOPS AI \u30d1\u30ef\u30fc
– 15-80W \u69cb\u6210\u53ef\u80fd TDP<\/td>		\u30a4\u30f3\u30c6\u30eb\u00ae Core\u2122 Ultra 7 356H<\/strong>
– 16 \u30b3\u30a2\uff5c\u6700\u5927 4.7 GHz \u30bf\u30fc\u30dc\u30fb\u30d6\u30fc\u30b9\u30c8
– 4 Xe3 \u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u30fb\u30b3\u30a2\uff5c100 TOPS AI Power
– 15-80W TDP \u8a2d\u5b9a\u53ef\u80fd<\/td><\/tr>
\u30b9\u30c8\u30ec\u30fc\u30b8\u3068\u62e1\u5f35<\/strong><\/td>– \u30e1\u30e2\u30ea32GB \/ 64GB LPDDR5x 8533 MT\/s
– \u30b9\u30c8\u30ec\u30fc\u30b8\uff1a1 \u00d7 M.2 PCIe Gen 5 \u00d7 4 \u30b9\u30ed\u30c3\u30c8
– \u62e1\u5f35: 1 \u00d7 M.2 PCIe Gen 4 \u00d7 2 \u30b9\u30ed\u30c3\u30c8
– \u30d7\u30ea\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\uff1a512GB – 4TB 2280 NVMe SSD <\/td>		– \u30e1\u30e2\u30ea32GB \/ 64GB LPDDR5x 8533 MT\/s
– \u30b9\u30c8\u30ec\u30fc\u30b8\uff1a1 \u00d7 M.2 PCIe Gen 5 \u00d7 4 \u30b9\u30ed\u30c3\u30c8
– \u62e1\u5f35: 1 \u00d7 M.2 PCIe Gen 4 \u00d7 2 \u30b9\u30ed\u30c3\u30c8
– \u30d7\u30ea\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\uff1a512GB – 4TB 2280 NVMe SSD <\/td><\/tr>
I\/O\u30dd\u30fc\u30c8<\/strong><\/td>– 2 \u00d7 USB4 v2.0 (80Gbps)
– 1 \u00d7 DisplayPort 2.1 (UHBR20)
– 1 \u00d7 HDMI 2.1 (FRL)
– 4 \u00d7 USB 3.2 Gen2 Type-A (10Gbps)
– 2 \u00d7 2.5G RJ45 LAN (10\/100\/1000\/2500 Mbps)
– 1 \u00d7 MCIO (PCIe 5.0 \u00d7 8)<\/td>		– 2 \u00d7 USB4 v2.0 (80Gbps)
– 1 \u00d7 DisplayPort 2.1 (UHBR20)
– 1 \u00d7 HDMI 2.1 (FRL)
– 4 \u00d7 USB 3.2 Gen2 Type-A (10Gbps)
– 2 \u00d7 2.5G RJ45 LAN (10\/100\/1000\/2500 Mbps)<\/td><\/tr>
\u30b2\u30fc\u30e0\u306e\u8155\u524d<\/strong><\/td>– \u7d71\u5408\uff1a4-Core Xe3 Graphics | 512 Stream Processors
– Radeon 780M \u306b\u5339\u6575\u3059\u308b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9
– \u5916\u90e8\u62e1\u5f35\uff1a\u30c7\u30e5\u30a2\u30eb\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30fc\u30b9GPU\u30b5\u30dd\u30fc\u30c8
– USB4 v2.0 (80Gbps)\uff1aPCIe 5.0 \u00d7 4 | ~31.5 GB\/s \u5e2f\u57df\u5e45
– MCIO Exclusive\uff1aPCIe 5.0 \u00d7 8 | ~63 GB\/s Bandwidth
– Ultimate Desktop Power: Support RTX 5090 \/ 5090D
– DLSS 4.0\u30b5\u30dd\u30fc\u30c8 | Up to 3x Frame Gen | 200-800 FPS\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9<\/td>		– \u7d71\u5408\uff1a12-Core Xe3 Graphics\uff5c1536 Stream Processors
– XeSS 3 Super Sampling\uff5cUp to 4x Frame Generation
– \u307b\u3068\u3093\u3069\u306e AAA \u30bf\u30a4\u30c8\u30eb\u3067 60-120 FPS \u3092\u5b9f\u73fe
– \u62e1\u5f35\u53ef\u80fd\uff1aUSB4 v2.0 (80Gbps) \u306b\u3088\u308b\u5916\u90e8 GPU
– PCIe 5.0 \u00d7 4 \u5171\u6709\u30ec\u30fc\u30f3 | ~31.5 GB\/s \u7406\u8ad6\u5e2f\u57df\u5e45
– NVIDIA GeForce RTX 5090 \/ 5090D \u307e\u3067\u30b5\u30dd\u30fc\u30c8
– DLSS 4.0 \u30b5\u30dd\u30fc\u30c8 | 3x \u30d5\u30ec\u30fc\u30e0\u751f\u6210 | \u6700\u5927 200-600 FPS<\/td><\/tr>
AI\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9<\/strong><\/td>– 180 TOPS Int8 Peak AI Compute
– \u30d9\u30f3\u30c1\u30de\u30fc\u30af\uff1aNVIDIA Jetson AGX Orin 64GB\u306b\u5339\u6575
– TDP 25W\u3067\uff1a\u5927\u5e45\u306a\u52b9\u7387\u5411\u4e0a
– LLM\u30b9\u30eb\u30fc\u30d7\u30c3\u30c8\uff1a\u6700\u59273.9\u500d\u9ad8\u901f
– \u97f3\u58f0\u5408\u6210\uff08TTS\uff09\u901f\u5ea6\uff1a\u6700\u59273.7\u500d\u9ad8\u901f
– \u30de\u30eb\u30c1\u30bf\u30b9\u30af\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30d3\u30b8\u30e7\u30f3\uff1a\u6700\u59275.4\u500d\u9ad8\u901f\u5316<\/td>		– 100 TOPS Int8 Aggregate AI Compute
– eGPU \u62e1\u5f35\uff1aMultimodal Interactive Agents
– Local LLM Inference, Scientific Computing & Simulation
– End-to-End Workflow\uff1a\u30e2\u30c7\u30eb\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u304b\u3089\u30c7\u30d7\u30ed\u30a4\u30e1\u30f3\u30c8\u307e\u3067
– \u30c7\u30fc\u30bf\u51e6\u7406\u3068AI\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u306e\u9ad8\u6027\u80fd\u30b5\u30dd\u30fc\u30c8<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

\u30d1\u30f3\u30b5\u30fc\u30fb\u30ec\u30a4\u30af\u306fGPD BOX\u306b\u660e\u78ba\u306a\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u30fb\u30a2\u30a4\u30c7\u30f3\u30c6\u30a3\u30c6\u30a3\u3092\u4e0e\u3048\u308b<\/h2>\n\n

Panther Lake is central to why the GPD BOX<\/a>looks more credible now than it did at first reveal. In Intel’s official platform announcement<\/a>, the company describes Panther Lake, now branded as Intel Core Ultra Series 3, as its first client SoC built on Intel 18A and aimed at AI PCs, gaming devices, and compact systems.<\/p>\n\n

GPD\u306f\u4eca\u56de\u3001BOX\u30922\u3064\u306e\u7279\u5b9a\u306e\u30c1\u30c3\u30d7\u306b\u7d50\u3073\u3064\u3051\u305f\u3002\u4e0a\u4f4d\u306e\u30a4\u30f3\u30c6\u30eb Core Ultra X7 358H\u306f\u300116\u30b3\u30a2\u3001\u6700\u59274.8GHz\u306e\u30d6\u30fc\u30b9\u30c8\u30b9\u30d4\u30fc\u30c9\u300112\u500b\u306eXe3\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b3\u30a2\u3001\u6700\u5927180TOPS\u306eAI\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u300115W\u304b\u308980W\u307e\u3067\u306e\u8a2d\u5b9a\u53ef\u80fd\u306aTDP\u3092\u5099\u3048\u3066\u3044\u308b\u3002\u307e\u305f\u3001\u30a4\u30f3\u30c6\u30eb Core Ultra 7 356H\u306f\u300116\u30b3\u30a2\u3001\u6700\u59274.7GHz\u306e\u30d6\u30fc\u30b9\u30c8\u901f\u5ea6\u30014\u3064\u306eXe3\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b3\u30a2\u3001\u6700\u5927100TOPS\u3001\u540c\u305815W\u304b\u308980W\u306e\u8a2d\u5b9a\u53ef\u80fd\u306a\u7bc4\u56f2\u3067\u63b2\u8f09\u3055\u308c\u3066\u3044\u308b\u3002 <\/p>\n\n

Those confirmed details make the GPD BOX<\/a> easier to frame against the wider mini PC market. It is no longer just a Panther Lake teaser with a nice chassis render. It now has clear CPU targets, clear GPU targets, a known memory ceiling, and a published connectivity story that leans hard into external graphics expansion.<\/p>\n\n

\"GPD<\/a>
GPD BOX\u80cc\u9762\u56f3<\/figcaption><\/figure>\n\n

GPD\u306f\u307e\u305f\u3001\u3053\u308c\u3089\u306e\u30c1\u30c3\u30d7\u306b\u975e\u5e38\u306b\u5177\u4f53\u7684\u306a\u6027\u80fd\u306b\u95a2\u3059\u308b\u8a00\u8449\u3092\u6dfb\u3048\u3066\u3044\u308b\u3002Ultra X7 358H\u306b\u3064\u3044\u3066\u306f\u3001\u7d71\u540812\u30b3\u30a2Xe3\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u3092\u642d\u8f09\u3057\u3001\u307b\u3068\u3093\u3069\u306eAAA\u30b2\u30fc\u30e0\u306760\uff5e120 FPS\u3092\u5b9f\u73fe\u3057\u3001XeSS 3\u306e\u30b5\u30dd\u30fc\u30c8\u3092\u5f37\u8abf\u3057\u3066\u3044\u307e\u3059\u3002Ultra 7 356H\u306b\u3064\u3044\u3066\u306f\u3001\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u6027\u80fd\u306fRadeon 780M\u30af\u30e9\u30b9\u306e\u7d71\u5408\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u306b\u5339\u6575\u3059\u308b\u3068\u3057\u3066\u3044\u308b\u3002\u3053\u308c\u3089\u306e\u4e3b\u5f35\u306f\u3001\u3088\u308a\u8a73\u7d30\u306a\u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306e\u88cf\u4ed8\u3051\u304c\u53d6\u308c\u308b\u307e\u3067\u306f\u614e\u91cd\u306b\u6271\u308f\u308c\u308b\u3079\u304d\u3067\u3059\u304c\u3001GPD\u304c\u3053\u306e2\u3064\u306e\u69cb\u6210\u3092\u3069\u306e\u3088\u3046\u306b\u7406\u89e3\u3057\u3066\u6b32\u3057\u3044\u304b\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002 <\/p>\n\n

The same is true for the AI story. GPD is pitching the higher-tier BOX as a 180 TOPS machine and the lower-tier system as a 100 TOPS design, then connecting those figures to local LLM work, scientific computing, simulation, and multimodal agent use cases. Whether those claims prove compelling in real-world workloads remains to be seen, but the company has clearly decided the GPD BOX<\/a> should be read as more than a small PC for light office tasks.<\/p>\n\n

\u78ba\u8a8d\u3055\u308c\u305fGPD BOX\u306e\u8a73\u7d30<\/th>GPD\u304c\u73fe\u5728\u767a\u8868\u3057\u3066\u3044\u308b\u3082\u306e<\/th>\u306a\u305c\u305d\u308c\u304c\u91cd\u8981\u306a\u306e\u304b<\/th><\/tr><\/thead>
CPU\u30aa\u30d7\u30b7\u30e7\u30f3<\/td>\u30a4\u30f3\u30c6\u30eb\u30fb\u30b3\u30a2\u30fb\u30a6\u30eb\u30c8\u30e9X7 358H\u307e\u305f\u306f\u30a4\u30f3\u30c6\u30eb\u30fb\u30b3\u30a2\u30fb\u30a6\u30eb\u30c8\u30e97 356H<\/td>BOX\u304c2\u3064\u306e\u7279\u5b9a\u306ePanther Lake Series 3\u30d7\u30ed\u30bb\u30c3\u30b5\u30fc\u3068\u95a2\u9023\u3057\u3066\u3044\u308b\u3053\u3068\u3092\u78ba\u8a8d<\/td><\/tr>
\u30b0\u30e9\u30d5\u30a3\u30c3\u30af<\/td>Xe3\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u3001X7\u30e2\u30c7\u30eb\u3067\u306f\u6700\u592712\u30b3\u30a2<\/td>\u7d71\u5408\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u3078\u306e\u610f\u6b32\u304cBOX\u306e\u30d4\u30c3\u30c1\u306e\u4e3b\u8981\u90e8\u5206\u3067\u3042\u308b\u3053\u3068\u3092\u793a\u3059\u3002<\/td><\/tr>
\u30e1\u30e2\u30ea\u30fc<\/td>32GB \u307e\u305f\u306f 64GB LPDDR5x 8533 MT\/s<\/td>BOX\u306f\u30a8\u30f3\u30c8\u30ea\u30fc\u30ec\u30d9\u30eb\u306e\u30b3\u30f3\u30d1\u30af\u30c8\u30b7\u30b9\u30c6\u30e0\u3088\u308a\u3082\u4e0a\u4f4d\u306b\u4f4d\u7f6e\u3059\u308b\u3002<\/td><\/tr>
\u30b9\u30c8\u30ec\u30fc\u30b8<\/td>PCIe Gen 5 x4 M.2\u30b9\u30ed\u30c3\u30c8\u00d71\u3001PCIe Gen 4 x2 M.2\u30b9\u30ed\u30c3\u30c8\u00d71<\/td>\u3088\u308a\u9ad8\u901f\u306a\u30b9\u30c8\u30ec\u30fc\u30b8\u30aa\u30d7\u30b7\u30e7\u30f3\u3092\u8ffd\u52a0\u3057\u3001\u3055\u307e\u3056\u307e\u306a\u69cb\u6210\u306b\u67d4\u8edf\u306b\u5bfe\u5fdc<\/td><\/tr>
\u5916\u90e8\u62e1\u5f35<\/td>PCIe Gen 5 x8\u306eMCIO 8i\u3001\u53cc\u65b9\u5411512Gbps\u5e2f\u57df\u5e45\u30684\u500d\u306eOCuLink\u5e2f\u57df\u5e45\u3067\u8ca9\u58f2\u3055\u308c\u308b<\/td>GPD\u304c\u5dee\u5225\u5316\u8981\u56e0\u3068\u3057\u3066MCIO\u3092\u91cd\u8996\u3057\u3066\u3044\u308b\u3053\u3068\u3092\u793a\u3059<\/td><\/tr>
\u30d3\u30c7\u30aa\u3068\u30cd\u30c3\u30c8\u30ef\u30fc\u30af<\/td>\u30c7\u30e5\u30a2\u30ebUSB4 v2.0\u3001DP 2.1\u3001HDMI 2.1\u3001USB-A\u00d74\u3001\u30c7\u30e5\u30a2\u30eb2.5G LAN<\/td>\u7b2c\u4e00\u4e16\u4ee3\u306e\u30df\u30cbPC\u3068\u3057\u3066\u306f\u3001BOX\u304c\u7570\u4f8b\u306b\u5145\u5b9f\u3057\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b<\/td><\/tr>
AI\u3068eGPU\u306e\u4f4d\u7f6e\u3065\u3051<\/td>\u6700\u5927180TOPS\u3001RTX 5090\u307e\u305f\u306f5090D\u30da\u30a2\u30ea\u30f3\u30b0\u306e\u30e1\u30c3\u30bb\u30fc\u30b8\u30f3\u30b0\u3092\u30b5\u30dd\u30fc\u30c8<\/td>GPD\u304c\u76ee\u6a19\u3068\u3059\u308b\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u30ec\u30d9\u30eb\u3092\u660e\u78ba\u306b\u3059\u308b\u3002<\/td><\/tr><\/tbody><\/table><\/figure>\n\n

\u306a\u305cGPD BOX\u3068GPD G2\u306e\u7d44\u307f\u5408\u308f\u305b\u304c\u672c\u5f53\u306e\u8a71\u306a\u306e\u304b\uff1f<\/h2>\n\n

\u3053\u3053\u3067\u672c\u5f53\u306b\u91cd\u8981\u306a\u306e\u306f\u3001GPD\u304c\u30df\u30cbPC\u3092\u767a\u58f2\u3057\u305f\u3053\u3068\u3067\u3082\u3001\u3088\u308a\u5e83\u5e2f\u57df\u306eeGPU\u30c9\u30c3\u30af\u3092\u767a\u58f2\u3057\u305f\u3053\u3068\u3067\u3082\u306a\u3044\u3002\u3088\u308a\u91cd\u8981\u306a\u70b9\u306f\u3001\u540c\u793e\u304c\u3053\u306e2\u3064\u3092\u30e2\u30b8\u30e5\u30fc\u30eb\u5f0f\u306e\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u30fb\u30b7\u30b9\u30c6\u30e0\u3068\u3057\u3066\u767a\u8868\u3057\u3066\u3044\u308b\u3053\u3068\u3060\u3002 <\/p>\n\n

\u5e2f\u57df\u5e45\u306f\u3001\u7279\u306b\u5c0f\u578b\u30c7\u30d0\u30a4\u30b9\u304b\u3089\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u30af\u30e9\u30b9\u306e\u7d50\u679c\u3092\u5f97\u305f\u3044\u5834\u5408\u306b\u3001\u5916\u90e8\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u306e\u4e3b\u306a\u5236\u9650\u306e1\u3064\u3068\u306a\u3063\u3066\u304d\u305f\u3002GPD\u306f\u3001MCIO\u304c\u3001\u7279\u306bUSB4 v2.0\u3084\u3088\u308a\u5f37\u529b\u306a\u30aa\u30f3\u30dc\u30fc\u30c9I\/O\u3068\u7d44\u307f\u5408\u308f\u305b\u308b\u3053\u3068\u304c\u3067\u304d\u308c\u3070\u3001\u65e7\u6765\u306e\u62e1\u5f35\u30a2\u30d7\u30ed\u30fc\u30c1\u3088\u308a\u3082\u3001\u3053\u306e\u554f\u984c\u306b\u5bfe\u3059\u308b\u3088\u308a\u8aac\u5f97\u529b\u306e\u3042\u308b\u7b54\u3048\u306b\u306a\u308b\u3053\u3068\u306b\u8ced\u3051\u3066\u3044\u308b\u3053\u3068\u306f\u660e\u3089\u304b\u3067\u3059\u3002 <\/p>\n\n

\"GPD<\/a>
GPD BOX\u3001\u30b9\u30bf\u30f3\u30c9\u3001\u96fb\u6e90\u30b1\u30fc\u30d6\u30eb\u4ed8\u304d<\/figcaption><\/figure>\n\n

If the GPD BOX<\/a> performs well as a compact daily-use desktop and the G2 delivers something close to the desktop-style expansion GPD is promising, the pairing could appeal to gamers, creators, developers, home lab users, and power users who want a flexible system without moving straight to a full-size tower.<\/p>\n\n

\u307e\u305f\u3001GPD\u306e\u4ed6\u306e\u8aad\u8005\u5c64\u3068\u3082\u3046\u307e\u304f\u5408\u81f4\u3057\u3066\u3044\u308b\u3002\u30a2\u30af\u30bb\u30b5\u30ea\u30fc\u30fb\u30ab\u30c6\u30b4\u30ea\u30fc\u3084<\/a>\u30d7\u30ec\u30df\u30a2\u30e0\u30fb\u30cf\u30f3\u30c9\u30d8\u30eb\u30c9\u3001\u30b3\u30f3\u30d1\u30af\u30c8\u306aGPD\u30b7\u30b9\u30c6\u30e0\u3092\u3059\u3067\u306b\u691c\u8a0e\u3057\u3066\u3044\u308b\u30d0\u30a4\u30e4\u30fc\u306f\u3001\u30b9\u30e2\u30fc\u30eb\u30fb\u30b9\u30bf\u30fc\u30c8\u3067\u5f8c\u304b\u3089\u30b9\u30b1\u30fc\u30eb\u30a2\u30c3\u30d7\u3059\u308b\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u3092\u6c17\u306b\u3059\u308b\u53ef\u80fd\u6027\u304c\u9ad8\u3044\u8aad\u8005\u3060\u3002 <\/p>\n\n

\u307e\u3060\u8a3c\u660e\u3055\u308c\u3066\u3044\u306a\u3044\u3053\u3068<\/h2>\n\n

\u30b9\u30da\u30c3\u30af\u30b7\u30fc\u30c8\u304c\u5145\u5b9f\u3057\u3066\u3044\u3066\u3082\u3001\u307e\u3060\u91cd\u8981\u306a\u30ae\u30e3\u30c3\u30d7\u304c\u3042\u308b\u3002GPD\u306f\u3001\u3053\u3053\u3067\u30ec\u30d3\u30e5\u30fc\u3057\u305f\u8cc7\u6599\u3067\u306f\u3001\u3044\u305a\u308c\u306e\u88fd\u54c1\u306b\u3064\u3044\u3066\u3082\u3001\u6700\u7d42\u7684\u306a\u4fa1\u683c\u3001\u767a\u58f2\u6642\u671f\u3001\u5730\u57df\u3054\u3068\u306e\u5728\u5eab\u72b6\u6cc1\u3092\u307e\u3060\u78ba\u8a8d\u3057\u3066\u3044\u306a\u3044\u3002 <\/p>\n\n

More importantly, the biggest remaining questions now sit around proof rather than specification. We still need fuller benchmark context for the RTX 4090 2% loss claim on the GPD G2<\/a>, the broader 4090 or 5090-class support messaging around the dock, the BOX 60 to 120 FPS gaming claims, the Radeon 780M comparison, and the RTX 5090 or 5090D pairing language tied to the GPD BOX<\/a>.<\/p>\n\n

\u3060\u304b\u3089\u3068\u3044\u3063\u3066\u3001\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306e\u8a71\u304c\u5f31\u3044\u308f\u3051\u3067\u306f\u306a\u3044\u3002\u3080\u3057\u308d\u3001\u30b9\u30da\u30c3\u30af\u306e\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u306b\u3088\u3063\u3066\u3001\u3053\u306e\u30b3\u30f3\u30bb\u30d7\u30c8\u3092\u771f\u5263\u306b\u53d7\u3051\u6b62\u3081\u3084\u3059\u304f\u306a\u3063\u3066\u3044\u308b\u3002\u3057\u304b\u3057\u3001\u8aad\u8005\u306f\u78ba\u8a8d\u3055\u308c\u305f\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u306e\u30ea\u30b9\u30c8\u3068\u3001\u307e\u3060\u72ec\u81ea\u306e\u30c6\u30b9\u30c8\u304c\u5fc5\u8981\u306a\u6027\u80fd\u306e\u7d50\u8ad6\u3068\u3092\u5206\u3051\u3066\u8003\u3048\u308b\u3079\u304d\u3060\u3068\u3044\u3046\u3053\u3068\u3060\u3002 <\/p>\n\n

For now, the announcement works best as a strong statement of intent. GPD is making it clear that it wants to compete more aggressively in compact performance computing, and the GPD BOX<\/a> plus GPD G2<\/a> combination is its clearest sign yet of how it plans to do it.<\/p>\n\n

\u6700\u8fd1\u306eGPD\u306e\u52e2\u3044\u3092\u3059\u3067\u306b\u8ffd\u3063\u3066\u3044\u308b\u4eba\u306b\u3068\u3063\u3066\u3082\u3001\u3053\u306e\u30cb\u30e5\u30fc\u30b9\u306f\u300cGPD WIN 5 – Everything We Know So Far\uff08GPD\u306eWIN 5 – \u3053\u308c\u307e\u3067\u306e\u5168\u60c5\u5831<\/a>\uff09\u300d\u306e\u3088\u3046\u306a\u5831\u9053\u3068\u81ea\u7136\u306b\u4e26\u3093\u3067\u3044\u308b\u3002\u30d5\u30a9\u30fc\u30e0\u30d5\u30a1\u30af\u30bf\u30fc\u306f\u7570\u306a\u308b\u304b\u3082\u3057\u308c\u306a\u3044\u304c\u3001\u6839\u5e95\u306b\u3042\u308b\u30e1\u30c3\u30bb\u30fc\u30b8\u306f\u304a\u306a\u3058\u307f\u3060\u3002 <\/p>\n\n

\u3088\u304f\u3042\u308b\u3054\u8cea\u554f<\/h2>\n\n

GPD G2\u3068\u306f\u4f55\u3067\u3059\u304b\uff1f<\/h3>\n\n

The GPD G2<\/a> is an eGPU dock that GPD describes as the world’s first MCIO dual-port eGPU dock. Its official specs page now lists a PCIe Gen 5 x16 slot, MCIO 8i support for PCIe Gen 4 x8 devices, USB4 v2.0, two USB-A ports, Gigabit Ethernet, an M.2 2280 slot, and up to 100W PD charging.<\/p>\n\n

GPD BOX\u3068\u306f\u4f55\u3067\u3059\u304b\uff1f<\/h3>\n\n

The GPD BOX<\/a> is a newly announced Panther Lake compact desktop that GPD is positioning as its first mini PC. GPD has listed Core Ultra X7 358H and Core Ultra 7 356H options, up to 64GB LPDDR5x memory, dual M.2 storage, MCIO 8i, dual USB4 v2.0, and up to 512 Gbps bidirectional MCIO bandwidth.<\/p>\n\n

\u306a\u305c\u30d1\u30f3\u30b5\u30fc\u6e56\u304cGPD BOX\u306b\u3068\u3063\u3066\u91cd\u8981\u306a\u306e\u304b\uff1f<\/h3>\n\n

Panther Lake\u304c\u91cd\u8981\u306a\u306e\u306f\u3001\u30a4\u30f3\u30c6\u30eb\u304c\u3053\u308c\u3092AI PC\u3001\u30b2\u30fc\u30df\u30f3\u30b0\u30c7\u30d0\u30a4\u30b9\u3001\u30b3\u30f3\u30d1\u30af\u30c8\u30b7\u30b9\u30c6\u30e0\u5411\u3051\u306e\u6700\u65b0\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u3068\u4f4d\u7f6e\u4ed8\u3051\u3066\u3044\u308b\u304b\u3089\u3060\u3002GPD\u306f\u4eca\u56de\u3001BOX\u30922\u3064\u306e\u7279\u5b9a\u306eCore Ultra Series 3\u30c1\u30c3\u30d7\u306b\u7d50\u3073\u3064\u3051\u305f\u3053\u3068\u3067\u3001\u30df\u30cbPC\u304c\u610f\u56f3\u3059\u308b\u6027\u80fd\u306e\u7bc4\u56f2\u304c\u3088\u308a\u660e\u78ba\u306b\u306a\u3063\u305f\u3002 <\/p>\n\n

GPD BOX\u306fGPD G2\u3068\u7d44\u307f\u5408\u308f\u305b\u3089\u308c\u307e\u3059\u304b\uff1f<\/h3>\n\n

Yes. In fact, that pairing is one of the main points of the announcement. GPD is explicitly presenting the GPD BOX<\/a> and GPD G2<\/a> as a combined setup for a much more powerful compact gaming or creator PC.<\/p>\n\n

GPD\u306f\u4fa1\u683c\u3084\u767a\u58f2\u6642\u671f\u3092\u78ba\u8a8d\u3057\u3066\u3044\u307e\u3059\u304b\uff1f<\/h3>\n\n

\u3044\u3044\u3048\u3001\u3053\u3053\u3067\u30ec\u30d3\u30e5\u30fc\u3057\u305f\u8cc7\u6599\u306b\u306f\u3042\u308a\u307e\u305b\u3093\u3002\u4fa1\u683c\u3001\u767a\u58f2\u6642\u671f\u3001\u5730\u57df\u306b\u3064\u3044\u3066\u306f\u307e\u3060\u78ba\u8a8d\u3059\u308b\u5fc5\u8981\u304c\u3042\u308b\u3002 <\/p>\n\n

RTX 4090\u3001RTX 5090\u30af\u30e9\u30b9\u3001BOX\u30b2\u30fc\u30df\u30f3\u30b0\u306e\u30af\u30ec\u30fc\u30e0\u306f\u6700\u7d42\u7684\u306a\u3082\u306e\u3068\u3057\u3066\u6271\u308f\u308c\u308b\u3079\u304d\u304b\uff1f<\/h3>\n\n

\u3044\u3084\u3001\u307e\u3060\u3060\u3002GPD\u306f\u305d\u306e\u3088\u3046\u306a\u4e3b\u5f35\u3092\u767a\u8868\u3057\u3066\u3044\u308b\u304c\u3001\u73fe\u5728\u5165\u624b\u53ef\u80fd\u306a\u767a\u8868\u8cc7\u6599\u306b\u306f\u3001\u72ec\u81ea\u306b\u691c\u8a3c\u3059\u308b\u305f\u3081\u306b\u5fc5\u8981\u306a\u5b8c\u5168\u306a\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u624b\u6cd5\u3084\u30c6\u30b9\u30c8\u6761\u4ef6\u304c\u8a18\u8f09\u3055\u308c\u3066\u3044\u306a\u3044\u3002 <\/p>\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

GPD\u306f\u3001\u30df\u30cbPC\u5e02\u5834\u3078\u306e\u521d\u53c2\u5165\u3068\u306a\u308bGPD BOX\u3068\u7d44\u307f\u5408\u308f\u305b\u3066\u3001\u6700\u5148\u7aef\u306eMCIO\u304a\u3088\u3073USB4 v2.0\u5916\u4ed8\u3051\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30b9\u30fb\u30c9\u30c3\u30af\u3067\u3042\u308bGPD G2\u3092\u6b63\u5f0f\u306b\u767a\u8868\u3057\u305f\u3002\u3053\u306e\u8a18\u4e8b\u3067\u306f\u3001\u4e2d\u6838\u3068\u306a\u308b\u4ed5\u69d8\u3001Intel Panther Lake\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u306e\u53ef\u80fd\u6027\u3001\u305d\u3057\u3066\u73fe\u5728\u78ba\u8a8d\u5f85\u3061\u3068\u306a\u3063\u3066\u3044\u308b\u91cd\u8981\u60c5\u5831\u306e\u5206\u6790\u306b\u3064\u3044\u3066\u89e3\u8aac\u3059\u308b\u3002 <\/p>\n","protected":false},"author":76,"featured_media":1280557,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1166,983],"tags":[23106,23108,23107,1167,23109],"class_list":["post-1280546","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-1166","category-983","tag-g2","tag-gpd-box","tag-23107","tag-1167","tag-23109"],"acf":[],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/posts\/1280546","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/users\/76"}],"replies":[{"embeddable":true,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/comments?post=1280546"}],"version-history":[{"count":2,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/posts\/1280546\/revisions"}],"predecessor-version":[{"id":1280585,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/posts\/1280546\/revisions\/1280585"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/media\/1280557"}],"wp:attachment":[{"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/media?parent=1280546"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/categories?post=1280546"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/gpdstore.net\/ja\/wp-json\/wp\/v2\/tags?post=1280546"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}