AI inference clusters are increasingly constrained by instantaneous power, not just energy: grid operators condition new capacity on demand response, imposing time-varying power caps. Existing LLM serving systems optimize a static energy objective or shed fixed priority tiers under load; either way, goodput collapses when the power envelope moves. An LLM pipeline is not a uniform load: compute-bound prefill loses throughput almost linearly with GPU frequency, memory-bound answer decode sustains it down to\u00a00.57\u00d7\u00a0nominal, and reasoning’s thinking phase couples KV-cache capacity to scheduling — so a cap should be steered to where each watt costs the least performance. PowerSlider does so with a new Flex SLO contract that turns bounded user slack into an optimization constraint, prefill–think–answer disaggregation exposing per-stage frequency and KV control, and a Karush–Kuhn–Tucker (KKT) online solver re-solving within 7.7 ms of every cap change, backed by a consolidated fail-safe that power-gates drained instances when DVFS bottoms out on static power. On SGLang with production traces, PowerSlider sustains 78.3% online goodput at a 30% cap reduction versus 47.6% for the best of five baselines (1.64\u00d7), holds latency-critical tails within\u00a01.3\u00d7\u00a0of nominal (baselines:\u00a02.3-6\u00d7, up to\u00a012\u00d7), and delivers 92% mean goodput through a replayed CAISO grid-emergency day bottoming at\u00a00.41\u00d7\u00a0(54% at the trough; every baseline below 7%).<\/p>\n","protected":false},"excerpt":{"rendered":"
AI inference clusters are increasingly constrained by instantaneous power, not just energy: grid operators condition new capacity on demand response, imposing time-varying power caps. Existing LLM serving systems optimize a static energy objective or shed fixed priority tiers under load; either way, goodput collapses when the power envelope moves. An LLM pipeline is not a […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"text","value":"Yueying Li","user_id":0},{"type":"text","value":"Jiayang Chen","user_id":0},{"type":"text","value":"Yuanfan Chen","user_id":0},{"type":"text","value":"Leo Han","user_id":0},{"type":"user_nicename","value":"Haoran Qiu","user_id":"43428"},{"type":"user_nicename","value":"Esha Choukse","user_id":"40417"},{"type":"user_nicename","value":"Rodrigo Fonseca","user_id":"40429"},{"type":"text","value":"Udit Gupta","user_id":0}],"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"","msr_doi":"10.48550\/arXiv.2608.21719","msr_arxiv_id":"","msr_mag_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_release_tracker_id":"","msr_highlight_type":"","msr_date_display_format":"","msr_main_download_label":"","msr_external_link_label":"","msr_doi_label":"","msr_published_date":"2026-08-22","msr_startdate":"","msr_presentation_date":"","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_year":2026,"msr_month":8,"msr_day":22,"msr_microsoftintellectualproperty":false,"msr_pub_id":"","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2608.21719","label_id":252679,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"","msr_s2_pdf_url":"","msr_citation_count_updated":"","msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556,13547],"msr-publication-type":[193715],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[268188],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1184906","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-systems-and-networking","msr-locale-en_us"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2026-08-22","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"10.48550\/arXiv.2608.21719","msr_publication_uploader":[{"type":"url","title":"https:\/\/arxiv.org\/abs\/2608.21719","label_id":252679,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Yueying Li","user_id":0,"rest_url":false},{"type":"text","value":"Jiayang Chen","user_id":0,"rest_url":false},{"type":"text","value":"Yuanfan Chen","user_id":0,"rest_url":false},{"type":"text","value":"Leo Han","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Haoran Qiu","user_id":43428,"rest_url":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Haoran Qiu"},{"type":"user_nicename","value":"Esha Choukse","user_id":40417,"rest_url":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Esha Choukse"},{"type":"user_nicename","value":"Rodrigo Fonseca","user_id":40429,"rest_url":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Rodrigo Fonseca"},{"type":"text","value":"Udit Gupta","user_id":0,"rest_url":false}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[282170],"msr_project":[1017939],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"article","related_content":{"projects":[{"ID":1017939,"post_title":"Efficient AI","post_name":"efficient-ai","post_type":"msr-project","post_date":"2024-03-22 17:14:57","post_modified":"2026-03-11 10:49:36","post_status":"publish","permalink":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/project\/efficient-ai\/","post_excerpt":"Making Azure's big bet possible Recent innovations in generative large language models (LLMs) have made their applications and use-cases ubiquitous. This has led to large-scale deployments of these models, using complex, expensive, and power-hungry AI accelerators, most commonly GPUs. These developments make LLM training and inference efficiency an important challenge. In the Azure Research - Systems (opens in new tab) group we are working on improving the Azure infrastructure including hardware, power, and serving. Check…","_links":{"self":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/1017939"}]}}]},"_links":{"self":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1184906","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":2,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1184906\/revisions"}],"predecessor-version":[{"id":1184908,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1184906\/revisions\/1184908"}],"wp:attachment":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1184906"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=1184906"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1184906"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1184906"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1184906"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1184906"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1184906"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1184906"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1184906"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=1184906"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1184906"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1184906"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1184906"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1184906"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}