{"id":27635,"date":"2026-03-05T04:00:00","date_gmt":"2026-03-05T11:00:00","guid":{"rendered":"https:\/\/ambiq.com\/?post_type=news&p=27635"},"modified":"2026-03-08T00:43:45","modified_gmt":"2026-03-08T07:43:45","slug":"ambiq-reveals-12nm-spot-platform-enabling-operation-down-to-300mv-for-atomiq","status":"publish","type":"news","link":"https:\/\/ambiq.com\/tw\/news\/ambiq-reveals-12nm-spot-platform-enabling-operation-down-to-300mv-for-atomiq\/","title":{"rendered":"Ambiq \u63a8\u51fa 12nm SPOT\u00ae \u5e73\u53f0\uff0c\u4f7f Atomiq™ SoC \u53ef\u5728\u4f4e\u81f3 300mV \u96fb\u58d3\u4e0b\u904b\u884c"},"content":{"rendered":"\n
\"\"
<\/figcaption><\/figure>\n\n\n\n

\u7f8e\u570b\u5fb7\u5dde\u5967\u65af\u6c40\uff0c2026\u5e743\u67085\u65e5<\/strong> \u2014 \u8d85\u4f4e\u529f\u8017\u908a\u7de3 AI \u534a\u5c0e\u9ad4\u89e3\u6c7a\u65b9\u6848\u9818\u5c0e\u5ee0\u5546 Ambiq Micro, Inc.\uff08Ambiq\u00ae\uff09 <\/strong>\u4eca\u65e5\u516c\u5e03\u5373\u5c07\u63a8\u51fa\u7684 Atomiq™ \u7cfb\u7d71\u55ae\u6676\u7247\uff08SoC\uff09 \u7684\u6700\u65b0\u6280\u8853\u7d30\u7bc0\u3002Atomiq \u63a1\u7528 Ambiq \u5168\u65b0\u8a2d\u8a08\u7684 12nm SPOT\u00ae \u5e73\u53f0\uff0c\u4e26\u57fa\u65bc \u53f0\u7a4d\u96fb\uff08TSMC\uff09N12e\u00ae \u88fd\u7a0b\u6280\u8853\u6253\u9020\uff0c\u5c0e\u5165\u5168\u65b0\u7684 \u8d85\u4f4e\u529f\u8017\uff08Ultra-Low-Power, ULP\uff09\u904b\u4f5c\u6a21\u5f0f\u3002\u8a72\u6a21\u5f0f\u53ef\u5728 \u4f4e\u81f3 300mV \u7684\u96fb\u58d3\u4e0b\u904b\u884c\uff0c\u5275\u4e0b Ambiq \u6b77\u4f86 \u6700\u4f4e\u904b\u4f5c\u96fb\u58d3 \u7684\u65b0\u91cc\u7a0b\u7891\u3002<\/p>\n\n\n\n

\u6b64\u9805\u767c\u5e03\u6a19\u8a8c\u8457 SPOT\u00ae \u6280\u8853\u9996\u6b21\u898f\u5283\u5c0e\u5165 FinFET \u88fd\u7a0b\u7bc0\u9ede\uff0c\u540c\u6642\u4e5f\u5c55\u73fe Atomiq™ \u5e73\u53f0\u958b\u767c\u7684\u91cd\u8981\u9032\u5c55\u3002\u76ee\u524d Atomiq110 \u5df2\u671d\u5411\u9996\u6279\u6a23\u54c1\u63a8\u51fa\u7a69\u6b65\u63a8\u9032\uff0c\u4e26\u9810\u8a08\u65bc 2027 \u5e74\u6b63\u5f0f\u91cf\u7522\u3002<\/p>\n\n\n\n

FinFET + SPOT\uff1a\u958b\u555f\u5168\u65b0\u529f\u8017\u6548\u80fd\u65b0\u5883\u754c<\/strong><\/p>\n\n\n\n

\u900f\u904e\u7d50\u5408 Ambiq \u5c08\u6709\u7684 \u6b21\u81e8\u754c\uff08Subthreshold\uff09\u96fb\u8def\u8a2d\u8a08\u6280\u8853<\/strong> \u8207 \u53f0\u7a4d\u96fb\u5148\u9032\u7684 12nm FinFET \u88fd\u7a0b<\/strong>\uff0cAtomiq \u5f97\u4ee5\u5be6\u73fe \u9ad8\u901f\u6b21\u81e8\u754c\u8207\u8fd1\u81e8\u754c\uff08Near-threshold\uff09\u904b\u4f5c\u6548\u80fd<\/strong>\uff0c\u800c\u9019\u5728\u50b3\u7d71 \u5e73\u9762 CMOS \u88fd\u7a0b<\/strong> \u4e2d\u904e\u53bb\u5e7e\u4e4e\u96e3\u4ee5\u5be6\u73fe\u3002<\/p>\n\n\n\n

\u5168\u65b0 ULP\uff08\u8d85\u4f4e\u529f\u8017\uff09\u6a21\u5f0f \u53ef\u5728 \u4f4e\u81f3 300mV \u96fb\u58d3 \u4e0b\u904b\u4f5c\uff0c\u540c\u6642\u7dad\u6301\u7a69\u5b9a\u6548\u80fd\uff0c\u4e26\u64f4\u5c55 ULP\u3001LP\uff08\u4f4e\u529f\u8017\uff09\u8207 HP\uff08\u9ad8\u6548\u80fd\uff09\u6a21\u5f0f \u4e4b\u9593\u7684\u52d5\u614b\u7bc4\u570d\u3002\u6b64\u9805\u6280\u8853\u9032\u5c55\u8b93 Ambiq \u80fd\u5920\u5be6\u73fe \u6b77\u4f86\u6700\u4f4e\u904b\u4f5c\u96fb\u58d3\u8207\u66f4\u9ad8\u80fd\u6e90\u6548\u7387\u3002<\/p>\n\n\n\n

\u5c08\u70ba\u908a\u7de3 AI \u61c9\u7528\u6253\u9020<\/strong><\/p>\n\n\n\n

Atomiq \u5c08\u70ba AI \u5de5\u4f5c\u8ca0\u8f09 \u9032\u884c\u6700\u4f73\u5316\u8a2d\u8a08\u3002\u76f8\u8f03\u65bc\u55ae\u7d14\u4f9d\u8cf4\u66f4\u9ad8\u6642\u8108\u901f\u5ea6\uff0cAI \u904b\u7b97\u66f4\u9069\u5408\u900f\u904e \u5e73\u884c\u8655\u7406\u67b6\u69cb \u4f86\u63d0\u5347\u6548\u7387\u3002\u6b64\u524d\u5df2\u516c\u5e03\u7684 Arm\u00ae Ethos™-U85 NPU \u53ef\u540c\u6642\u57f7\u884c \u6578\u767e\u500b\u4e58\u52a0\u904b\u7b97\uff08MAC\uff09\uff0c\u5373\u4f7f\u5728 ULP\uff08\u8d85\u4f4e\u529f\u8017\uff09\u6a21\u5f0f \u4e0b\u3001\u6642\u8108\u4f4e\u81f3 100 MHz \u7684\u60c5\u6cc1\u4e0b\uff0c\u4ecd\u80fd\u63d0\u4f9b \u6bcf\u79d2\u6578\u767e\u5104\u6b21\u904b\u7b97\uff08tens of billions of operations per second\uff09 \u7684 AI \u904b\u7b97\u80fd\u529b\u3002<\/p>\n\n\n\n

\u5728\u63a5\u8fd1 300mV \u96fb\u58d3\u7684 ULP\uff08\u8d85\u4f4e\u529f\u8017\uff09\u6a21\u5f0f \u4e0b\u57f7\u884c AI \u63a8\u8ad6\uff0c\u9810\u671f\u53ef\u5927\u5e45\u964d\u4f4e \u908a\u7de3 AI \u61c9\u7528\u7684\u529f\u8017\uff0c\u540c\u6642\u4ecd\u53ef\u642d\u914d \u6a19\u6e96\u958b\u767c\u5de5\u5177\u8207\u751f\u614b\u7cfb\u7d71\u4f7f\u7528\u3002\u9019\u9805\u8d85\u4f4e\u529f\u8017\u7279\u6027\u540c\u6642\u652f\u63f4 SoC \u5167\u7684\u591a\u500b\u5b50\u7cfb\u7d71\uff0c\u5305\u62ec \u8655\u7406\u5668\u3001NPU \u4ee5\u53ca\u591a\u5a92\u9ad4\u8655\u7406\u5f15\u64ce\u3002<\/p>\n\n\n\n

\u5e73\u53f0\u5275\u65b0\u5be6\u73fe\u8d85\u4f4e\u529f\u8017\u904b\u4f5c <\/strong><\/p>\n\n\n\n

\u8981\u5be6\u73fe \u4f4e\u81f3 300mV \u7684\u904b\u4f5c\u96fb\u58d3\uff0c\u9700\u8981\u5728\u67b6\u69cb\u8a2d\u8a08\u4e0a\u9032\u884c\u91cd\u5927\u5275\u65b0\u3002Atomiq \u6574\u5408\u4e86\u5168\u65b0\u8a2d\u8a08\u7684 \u6676\u7247\u5167\u96fb\u6e90\u7ba1\u7406\u55ae\u5143\uff08PMU\uff09\uff0c\u4e26\u63a1\u7528 \u9ad8\u5ea6\u53ef\u914d\u7f6e\u7684\u591a\u901a\u9053\u55ae\u96fb\u611f\u591a\u8f38\u51fa\uff08Single-Inductor Multiple-Output, SIMO\uff09\u964d\u58d3\u8f49\u63db\u5668\u3002\u6b64\u81ea\u9069\u61c9\u96fb\u6e90\u7cfb\u7d71\u53ef\u52d5\u614b\u8abf\u6574\u4ee5\u56e0\u61c9 \u88fd\u7a0b\uff08process\uff09\u3001\u6eab\u5ea6\uff08temperature\uff09\u8207\u5de5\u4f5c\u8ca0\u8f09\uff08workload\uff09 \u7684\u8b8a\u5316\uff0c\u5728 \u5948\u5b89\u57f9\uff08nA\uff09\u81f3\u6beb\u5b89\u57f9\uff08mA\uff09 \u7684\u8ca0\u8f09\u7bc4\u570d\u5167\u63d0\u4f9b \u516d\u7d44\u7368\u7acb\u96fb\u58d3\u8ecc\u3002<\/p>\n\n\n\n

\u70ba\u5be6\u73fe\u5982\u6b64\u5927\u5e45\u5ea6\u7684\u96fb\u58d3\u7e2e\u653e\uff0cAmbiq \u4ea6\u958b\u767c\u4e86 \u5168\u65b0\u7684\u5ba2\u88fd\u5316\u57fa\u790e IP\uff08foundation IP\uff09\uff0c\u5c08\u70ba \u53ef\u9760\u7684\u8d85\u4f4e\u96fb\u58d3\u904b\u4f5c \u6240\u8a2d\u8a08\u300212nm SPOT\u00ae \u5e73\u53f0 \u5df2\u5728 Ambiq \u5be6\u9a57\u5ba4\u5b8c\u6210 \u6676\u7247\u9a57\u8b49\uff08silicon validation\uff09\uff0c\u9019\u4e5f\u6a19\u8a8c\u8457 Atomiq \u5546\u696d\u5316\u9032\u7a0b\u7684\u91cd\u8981\u6280\u8853\u91cc\u7a0b\u7891\u3002<\/p>\n\n\n\n

\u9ad8\u968e\u4e3b\u7ba1\u8a55\u8ad6<\/strong><\/p>\n\n\n\n

\u300cAtomiq \u4ee3\u8868\u4e86 SPOT \u6280\u8853\u591a\u5e74\u6301\u7e8c\u7cbe\u9032\u8207\u67b6\u69cb\u5275\u65b0\u7684\u6210\u679c\uff0c\u300dAmbiq \u5275\u8fa6\u4eba\u66a8\u6280\u8853\u9577 Scott Hanson<\/strong> \u8868\u793a\u3002\u300c\u5c07 SPOT \u6280\u8853\u5c0e\u5165 12nm FinFET <\/strong>\u88fd\u7a0b\u7bc0\u9ede\uff0c\u4f7f\u6211\u5011\u80fd\u5920\u5728\u63d0\u4f9b \u6b21\u4e16\u4ee3\u908a\u7de3 AI \u6240\u9700\u6548\u80fd \u7684\u540c\u6642\uff0c\u5c07\u904b\u4f5c\u96fb\u58d3\u964d\u81f3\u524d\u6240\u672a\u6709\u7684\u4f4e\u6c34\u6e96\u3002\u300d<\/p>\n\n\n\n

\u300c\u96a8\u8457 I<\/strong>oT \u88dd\u7f6e \u6301\u7e8c\u6f14\u9032\u4ee5\u56e0\u61c9 AI \u6642\u4ee3 \u7684\u9700\u6c42\uff0c\u5e02\u5834\u5c0d\u65bc \u66f4\u9ad8\u6548\u80fd\u3001\u6700\u4f73\u5316\u529f\u8017\u6548\u7387\u4ee5\u53ca\u66f4\u4f4e\u6f0f\u96fb\u6d41 \u7684\u5148\u9032\u534a\u5c0e\u9ad4\u6280\u8853\u9700\u6c42\u4e5f\u65e5\u76ca\u63d0\u5347\uff0c\u300d\u53f0\u7a4d\u96fb\u5317\u7f8e\u696d\u52d9\u7ba1\u7406\u526f\u7e3d\u88c1 Dr. Lucas Tsai<\/strong> \u8868\u793a\u3002\u300c\u6211\u5011\u6301\u7e8c\u8207 Ambiq \u7b49\u7522\u696d\u5275\u65b0\u8005\u5408\u4f5c\uff0c\u900f\u904e\u53f0\u7a4d\u96fb \u696d\u754c\u9818\u5148\u7684 ULP IoT \u5e73\u53f0\uff0c\u5171\u540c\u63a8\u52d5 \u8d85\u4f4e\u529f\u8017\u534a\u5c0e\u9ad4\u89e3\u6c7a\u65b9\u6848 \u7684\u6280\u8853\u9032\u5c55\u3002\u300d<\/p>\n\n\n\n

\u5c55\u671b\u672a\u4f86<\/strong> <\/p>\n\n\n\n

Atomiq110 SoC<\/strong> \u9810\u8a08\u5c07\u65bc 2027 <\/strong>\u5e74\u9032\u5165\u91cf\u7522\u3002Ambiq \u4ea6\u8a08\u756b\u5728\u672a\u4f86\u6578\u500b\u6708\u5167\u516c\u5e03\u66f4\u591a\u6280\u8853\u5275\u65b0\uff0c\u4ee5\u89e3\u6c7a AI \u7cfb\u7d71\u4ecd\u9762\u81e8\u7684\u91cd\u8981\u529f\u8017\u6311\u6230\u4e4b\u4e00 : \u8a18\u61b6\u9ad4\u6548\u7387\uff08memory efficiency\uff09\u3002<\/p>\n\n\n\n

\u6b61\u8fce\u849e\u81e8 Embedded World 2026<\/strong>\uff0c\u81f3 Hall 4, Booth 581<\/strong><\/a> \u53c3\u89c0 Ambiq \u5c55\u4f4d\uff0c\u9032\u4e00\u6b65\u4e86\u89e3 Atomiq\u3002\u5982\u9700\u66f4\u591a\u8cc7\u8a0a\u6216\u8a02\u95b1 Atomiq \u6700\u65b0\u6d88\u606f\u901a\u77e5\uff0c\u8acb\u9020\u8a2a Atomiq \u7522\u54c1\u9801\u9762<\/a>\u3002<\/p>\n\n\n\n

\u95dc\u65bc Ambiq<\/strong> <\/p>\n\n\n\n

Ambiq \u7684\u4f7f\u547d\u662f\u900f\u904e\u63d0\u4f9b\u6700\u4f4e\u529f\u8017\u7684\u534a\u5c0e\u9ad4\u89e3\u6c7a\u65b9\u6848\u4f86\u5be6\u73fe\u7121\u8655\u4e0d\u5728\u7684\u4eba\u5de5\u667a\u6167 \uff08AI\uff09 \u548c\u667a\u6167\u6280\u8853 \u3002 \u6211\u5011\u5e6b\u52a9\u5ba2\u6236\u5728\u529f\u8017\u6311\u6230\u6700\u70ba\u56b4\u5cfb\u7684\u908a\u7de3\u7aef\u5be6\u73fe\u4eba\u5de5\u667a\u6167 \uff08AI\uff09 \u8a08\u7b97\u3002 \u6211\u5011\u7684\u6280\u8853\u5275\u65b0\u57fa\u65bc\u5c08\u5229\u548c\u5c08\u6709\u6280\u8853\u2014\u2014\u4e9e\u95be\u503c\u529f\u8017\u512a\u5316\u6280\u8853 \uff08SPOT\u00ae\uff09 \uff0c\u8207\u50b3\u7d71\u534a\u5c0e\u9ad4\u8a2d\u8a08\u76f8\u6bd4\uff0c\u5f9e\u6839\u672c\u4e0a\u5c07\u6676\u7247\u529f\u8017\u964d\u4f4e\u4e86\u6578\u500d\u3002 \u622a\u81f3\u76ee\u524d\uff0c\u6211\u5011\u5df2\u7d93\u70ba\u8d85\u904e2.7\u5104\u53f0\u8a2d\u5099\u63d0\u4f9b\u652f\u63f4\u3002 \u6709\u95dc\u66f4\u591a\u8cc7\u8a0a\uff0c\u8acb\u8a2a\u554f www.ambiq.com<\/a>\u3002<\/p>\n\n\n\n

\u806f\u7e6b\u65b9\u5f0f<\/strong><\/strong><\/p>\n\n\n\n

Charlene Wan\u00a0<\/p>\n\n\n\n

\u4f01\u696d\u884c\u92b7\u526f\u7e3d\u88c1 <\/p>\n\n\n\n

cwan@ambiq.com<\/a> <\/p>\n\n\n\n

\u6295\u8cc7\u8005\u95dc\u4fc2\u806f\u7d61<\/strong> <\/p>\n\n\n\n

Teneo\u00a0\u00a0<\/p>\n\n\n\n

Christina Coronios<\/p>\n\n\n\n

christina.coronios@teneo.com<\/a> <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

\u7f8e\u570b\u5fb7\u5dde\u5967\u65af\u6c40\uff0c2026\u5e743\u67085\u65e5 \u2014 \u8d85\u4f4e\u529f\u8017\u908a\u7de3 AI \u534a\u5c0e\u9ad4\u89e3\u6c7a\u65b9\u6848\u9818\u5c0e\u5ee0\u5546 Ambiq Micro, Inc.\uff08Ambiq\u00ae\uff09 \u4eca\u65e5\u516c\u5e03\u5373\u5c07\u63a8\u51fa\u7684 Atomiq™<\/p>\n","protected":false},"featured_media":0,"template":"","news-category":[],"class_list":["post-27635","news","type-news","status-publish","hentry"],"acf":[],"_links":{"self":[{"href":"https:\/\/ambiq.com\/tw\/wp-json\/wp\/v2\/news\/27635","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ambiq.com\/tw\/wp-json\/wp\/v2\/news"}],"about":[{"href":"https:\/\/ambiq.com\/tw\/wp-json\/wp\/v2\/types\/news"}],"wp:attachment":[{"href":"https:\/\/ambiq.com\/tw\/wp-json\/wp\/v2\/media?parent=27635"}],"wp:term":[{"taxonomy":"news-category","embeddable":true,"href":"https:\/\/ambiq.com\/tw\/wp-json\/wp\/v2\/news-category?post=27635"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}