{"id":117256,"date":"2025-05-25T10:12:09","date_gmt":"2025-05-25T10:12:09","guid":{"rendered":"https:\/\/kamucalisani.net\/?p=117256"},"modified":"2025-05-25T10:12:09","modified_gmt":"2025-05-25T10:12:09","slug":"amazon-kendi-yapay-zeka-cipi-icin-kollari-sivadi-nvidia-hisseleri-dustu","status":"publish","type":"post","link":"https:\/\/kamucalisani.net\/index.php\/2025\/05\/25\/amazon-kendi-yapay-zeka-cipi-icin-kollari-sivadi-nvidia-hisseleri-dustu\/","title":{"rendered":"Amazon, kendi yapay zeka \u00e7ipi i\u00e7in kollar\u0131 s\u0131vad\u0131, Nvidia hisseleri d\u00fc\u015ft\u00fc"},"content":{"rendered":"

<\/span> Amazon<\/strong>, de\u011feri 100 milyar dolar\u0131 a\u015fan yapay zeka \u00e7ipi<\/strong> pazar\u0131nda Nvidia<\/strong>\u2019n\u0131n hanedanl\u0131\u011f\u0131na meydan okumak i\u00e7in \u00e7abalar\u0131n\u0131 art\u0131r\u0131yor. Amazon, halihaz\u0131rda kendi bulut platformu AWS\u2019de Nvidia\u2019n\u0131n \u00e7iplerinin yan\u0131nda kendi \u00e7iplerini de kullan\u0131yor. Ancak firman\u0131n gelecek planlar\u0131 aras\u0131nda daha fazla ba\u011f\u0131ms\u0131zl\u0131k hedefi bulunuyor. Buna giden yol ise \u015firket i\u00e7inde geli\u015ftirilen yapay zeka \u00e7iplerinden ge\u00e7iyor. <\/figure>\n

Bloomberg’in haberine g\u00f6re, Amazon m\u00fchendisleri, k\u00fc\u00e7\u00fck ancak yo\u011fun bir m\u00fchendislik laboratuvar\u0131nda, en yeni yapay zeka \u00e7ipi Trainium 2<\/strong>‘yi m\u00fcmk\u00fcn olan en k\u0131sa s\u00fcrede piyasaya s\u00fcrmeye \u00e7al\u0131\u015f\u0131yor. \u00c7ip sekt\u00f6r\u00fcne y\u0131llar\u0131n\u0131 veren Rami Sinno<\/strong>, Trainium 2 projesine \u00f6nc\u00fcl\u00fck ediyor ve Amazon’un \u00e7ip tasar\u0131m\u0131n\u0131 ve testlerini y\u00f6netiyor. Sinno, Amazon AI yar\u0131 iletkenlerinin ilk iki neslinin geli\u015ftirilmesine yard\u0131mc\u0131 olmu\u015ftu, \u015fimdi ise \u00fc\u00e7\u00fcnc\u00fc nesil yapay zeka \u00e7ipi Trainium 2 \u00fczerinde \u00e7al\u0131\u015f\u0131yor. Nvidia hisseleri bu rapor sonras\u0131nda y\u00fczde 4 kadar d\u00fc\u015ft\u00fc.<\/strong><\/p>\n

Nvidia\u2019ya kar\u015f\u0131 uygun maliyetli alternatif olacak<\/b><\/p>\n

Aktar\u0131lanlara g\u00f6re Nvidia’n\u0131n yapay zeka donan\u0131m\u0131ndaki \u00f6nemli liderli\u011fine ra\u011fmen, Amazon’un d\u0131\u015f tedarik\u00e7ilere ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 azaltma ve rekabet\u00e7i yapay zeka \u00e7\u00f6z\u00fcmleri sunma hedefleri ilgi g\u00f6rmeye ba\u015flad\u0131. Amazon’un en yeni yapay zeka \u00e7ipi Trainium 2, \u00f6nceki modele g\u00f6re 4 kat daha fazla hesaplama g\u00fcc\u00fc, 2 kat verimlilik ve 3 kat daha fazla bellek<\/strong> ile geli\u015fmi\u015f performans sunuyor.<\/p>\n

<\/span> Trainium 2 \u00fczerinde \u00e7al\u0131\u015fan ekip, dahili bile\u015fenleri azaltarak ve \u0131s\u0131 y\u00f6netimini iyile\u015ftirerek tasar\u0131m\u0131n\u0131 optimize etti. Elbette Trainium 2, Nvidia\u2019n\u0131n H200 veya yakla\u015fan Blackwell B200 gibi \u00fcst seviye yapay zeka \u00e7ipleriyle kafa kafaya yar\u0131\u015fmayacak, zaten \u00f6yle bir amac\u0131 yok. Bunun yerine \u00e7ip, Amazon’un yapay zeka yeteneklerini \u00f6l\u00e7eklendirme ve Nvidia donan\u0131m\u0131na uygun maliyetli bir alternatif<\/strong> sunma stratejisinin merkezinde yer al\u0131yor. <\/figure>\n

\u00a0 <\/p>\n

Anthropic<\/strong> de dahil olmak \u00fczere \u00f6nde gelen yapay zeka giri\u015fimleri, i\u015f y\u00fckleri i\u00e7in Amazon’un Trainium yongalar\u0131n\u0131 benimsiyor. Ge\u00e7ti\u011fimiz hafta Amazon, Anthropic’e sa\u011flad\u0131\u011f\u0131 fonu 8 milyar dolara<\/strong> \u00e7\u0131kararak yapay zeka alan\u0131ndaki ilerlemeler konusundaki i\u015fbirli\u011fini peki\u015ftirmi\u015f ve AWS\u2019yi Anthropic’in bulut sa\u011flay\u0131c\u0131s\u0131 yapm\u0131\u015ft\u0131. Amazon ayr\u0131ca i\u015fletmelere uygun maliyetli yapay zeka olu\u015fturma \u00e7\u00f6z\u00fcmleri sunmak i\u00e7in Databricks ile be\u015f y\u0131ll\u0131k bir anla\u015fma imzalam\u0131\u015ft\u0131. Databricks, Trainium’un kullan\u0131m\u0131ndan elde edilen tasarruflar\u0131 m\u00fc\u015fterilere aktarmay\u0131 ve Nvidia’n\u0131n yapay zeka donan\u0131m pazar\u0131ndaki hakimiyetine meydan okumay\u0131 ama\u00e7l\u0131yor.<\/p>\n

Yaz\u0131l\u0131m sorun olmaya devam ediyor<\/b><\/p>\n

\u00d6te yandan Amazon, Nvidia’n\u0131n CUDA<\/strong> sayesinde \u00fcst\u00fcn oldu\u011fu bir alan olan yaz\u0131l\u0131m geli\u015ftirmede engellerle kar\u015f\u0131la\u015f\u0131yor. Hen\u00fcz ilk a\u015famalar\u0131nda olan Neuron SDK<\/strong>‘s\u0131n\u0131n, Nvidia’n\u0131n yapay zeka geli\u015ftiricilerine sundu\u011fu kolayl\u0131k ve esneklikle e\u015fle\u015fecek \u015fekilde geli\u015fmesi gerekiyor. Esas\u0131nda Nvidia\u2019n\u0131n g\u00fcc\u00fcn\u00fcn as\u0131l kayna\u011f\u0131 donan\u0131m g\u00fcc\u00fc olmad\u0131, e\u011fer \u00f6yle olsayd\u0131 AMD\u2019nin yapay zeka \u00e7ipler \u00e7ok daha iyi konumlarda olurdu. Bloomberg’e g\u00f6re, Neuron SDK yaz\u0131l\u0131m paketi daha kullan\u0131c\u0131 dostu bir yaz\u0131l\u0131m olsa bile, Nvidia’dan Amazon’a ge\u00e7mek y\u00fczlerce saatlik geli\u015ftirme s\u00fcresi gerektirebilir.<\/p>\n\n

Kaynak\u00a0 :\u00a0https:\/\/www.donanimhaber.com\/amazon-nvidia-nin-yapay-zeka-cipleriyle-rekabet-edecek–184595<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

Amazon, de\u011feri 100 milyar dolar\u0131 a\u015fan yapay zeka \u00e7ipi pazar\u0131nda Nvidia\u2019n\u0131n hanedanl\u0131\u011f\u0131na meydan okumak i\u00e7in \u00e7abalar\u0131n\u0131 art\u0131r\u0131yor. Amazon, halihaz\u0131rda kendi bulut platformu AWS\u2019de Nvidia\u2019n\u0131n \u00e7iplerinin yan\u0131nda kendi \u00e7iplerini de kullan\u0131yor. Ancak …<\/p>\n","protected":false},"author":1,"featured_media":117257,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[533,480,164,571],"class_list":["post-117256","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-teknoloji","tag-amazon","tag-cip","tag-nvidia","tag-yapay-zeka"],"_links":{"self":[{"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/posts\/117256","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/comments?post=117256"}],"version-history":[{"count":1,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/posts\/117256\/revisions"}],"predecessor-version":[{"id":117260,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/posts\/117256\/revisions\/117260"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/media\/117257"}],"wp:attachment":[{"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/media?parent=117256"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/categories?post=117256"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/kamucalisani.net\/index.php\/wp-json\/wp\/v2\/tags?post=117256"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}