[{"data":1,"prerenderedAt":615},["ShallowReactive",2],{"sanity-YhX1HDaXpmfKQzIs7IY6mVn1RZUa4WyhbNA6-REKhbw":3},{"data":4,"sourceMap":-1},{"_createdAt":5,"_id":6,"_rev":7,"_type":8,"_updatedAt":9,"author":10,"category":20,"content":26,"draft":606,"mainImage":607,"publishedAt":610,"slug":611,"summary":613,"title":614},"2026-10-30T14:12:30Z","eb78e81f-be50-49ac-a824-d9dbf0d04fd3","wekmjS1uhPZNv5mJc94QbO","blog","2026-10-30T14:19:18Z",{"_id":11,"image":12,"name":17,"role":18,"slug":19},"1bdfb84b-6a55-4c46-a0ca-5448da3ce161",{"_type":13,"asset":14},"image",{"_ref":15,"_type":16},"image-789f75d971c7736e0250517cef7b59a7de9ae8be-1877x2469-jpg","reference","Virginia Senf","Growth Lead",null,{"_id":21,"slug":22,"title":25},"304a4be8-8a19-4976-8903-48ed811e951f",{"_type":23,"current":24},"slug","speckle-aec-voices","Speckle AEC voices",[27,38,46,54,62,71,79,94,106,118,130,142,150,158,166,174,186,198,206,214,222,233,241,252,260,268,276,287,295,303,311,319,327,338,346,354,362,373,384,392,400,408,419,427,438,446,454,489,497,505,513,526,538,550,562,574,582],{"_key":28,"_type":29,"children":30,"markDefs":36,"style":37},"772f75214ead","block",[31],{"_key":32,"_type":33,"marks":34,"text":35},"ac47c5b2b9dd","span",[],"On a data center project, the most valuable thing you can save isn't money. It's time.",[],"normal",{"_key":39,"_type":29,"children":40,"markDefs":45,"style":37},"5e5d3cae6fac",[41],{"_key":42,"_type":33,"marks":43,"text":44},"fa53eb8a40fb",[],"Every week a facility isn't energized is a week of capacity that isn't online, compute that isn't sold, and a customer commitment that's at risk. For hyperscalers, colocation providers and the developers building for them, the business case is written in megawatts delivered by a date. A 5% saving on design hours barely registers next to a month pulled forward on the schedule.",[],{"_key":47,"_type":29,"children":48,"markDefs":53,"style":37},"20c307e7604d",[49],{"_key":50,"_type":33,"marks":51,"text":52},"4fb93ea4b018",[],"That changes how owners and delivery teams should think about technology. The question isn't \"how do we make our people more efficient?\" It's \"what is sitting on the critical path, and how do we take it off?\" More often than most teams realize, the answer is model data: getting it, checking it, coordinating it and trusting it.",[],{"_key":55,"_type":29,"children":56,"markDefs":61,"style":37},"8ce57bc76b37",[57],{"_key":58,"_type":33,"marks":59,"text":60},"9471f4d3afcd",[],"This post covers five practices we see working on fast-track data center programs, and how Speckle's data platform and new agentic capabilities support each one.",[],{"_key":63,"_type":29,"children":64,"markDefs":69,"style":70},"ba2bf53dcc57",[65],{"_key":66,"_type":33,"marks":67,"text":68},"29eaf7e34edc",[],"Why data center delivery stalls",[],"h2",{"_key":72,"_type":29,"children":73,"markDefs":78,"style":37},"3cb835c2efef",[74],{"_key":75,"_type":33,"marks":76,"text":77},"a4ab2db5bff0",[],"Data centers look repeatable on paper. In practice, every site is a variation: different power and cooling configurations, local codes, rack densities, and phasing. That combination of repetition and variation creates a specific set of bottlenecks.",[],{"_key":80,"_type":29,"children":81,"level":91,"listItem":92,"markDefs":93,"style":37},"9b99e043f147",[82,87],{"_key":83,"_type":33,"marks":84,"text":86},"4f9beb3c8570",[85],"strong","Compressed, overlapping schedules.",{"_key":88,"_type":33,"marks":89,"text":90},"5f4323920067",[]," Design, procurement and construction run in parallel. Decisions are made on incomplete information, and the model changes underneath everyone while work is already underway.",1,"bullet",[],{"_key":95,"_type":29,"children":96,"level":91,"listItem":92,"markDefs":105,"style":37},"034dced6f731",[97,101],{"_key":98,"_type":33,"marks":99,"text":100},"51940e9d2cfa",[85],"Extreme MEP density.",{"_key":102,"_type":33,"marks":103,"text":104},"e92d42ececa0",[]," Power distribution, cooling, cable trays and fire protection compete for the same space. Coordination is harder and clashes are more frequent than on most building types.",[],{"_key":107,"_type":29,"children":108,"level":91,"listItem":92,"markDefs":117,"style":37},"3b0e1f58655f",[109,113],{"_key":110,"_type":33,"marks":111,"text":112},"c5e9152a3afa",[85],"Prefabrication raises the stakes.",{"_key":114,"_type":33,"marks":115,"text":116},"5a60dba94669",[]," Skids, modular electrical rooms and prefabricated MEP racks only save time if the model is right before fabrication starts. An error found on site costs weeks, not hours.",[],{"_key":119,"_type":29,"children":120,"level":91,"listItem":92,"markDefs":129,"style":37},"99f7824cd8e8",[121,125],{"_key":122,"_type":33,"marks":123,"text":124},"414be71896ce",[85],"Fragmented tools and teams.",{"_key":126,"_type":33,"marks":127,"text":128},"6aac0b54135b",[]," Architects, MEP engineers, structural engineers, trade contractors and fabricators all work in different authoring tools and file formats. Information moves through exports, uploads and email.",[],{"_key":131,"_type":29,"children":132,"level":91,"listItem":92,"markDefs":141,"style":37},"ecf1e08368b1",[133,137],{"_key":134,"_type":33,"marks":135,"text":136},"7d3a4ad99000",[85],"Coordination is a human bottleneck.",{"_key":138,"_type":33,"marks":139,"text":140},"709c62d2210f",[]," Experienced VDC and coordination managers are scarce. Much of their week goes on running clash tests, triaging results, chasing model updates and checking whether data meets standards, rather than resolving the problems that matter.",[],{"_key":143,"_type":29,"children":144,"markDefs":149,"style":37},"af57be3f2fdd",[145],{"_key":146,"_type":33,"marks":147,"text":148},"1a624efbb8ac",[],"None of these are solved by working harder. They're solved by making the right information available sooner, and by taking repetitive work off the people your schedule depends on.",[],{"_key":151,"_type":29,"children":152,"markDefs":157,"style":70},"7ddb4a83b63d",[153],{"_key":154,"_type":33,"marks":155,"text":156},"1aedd91d1066",[],"Practice 1: Treat the program as a product, not a series of projects",[],{"_key":159,"_type":29,"children":160,"markDefs":165,"style":37},"9a9044cfbe0f",[161],{"_key":162,"_type":33,"marks":163,"text":164},"7d96701b97d5",[],"The fastest data center programs don't start each site from zero. They treat their design as a product: a set of proven building blocks, such as data halls, electrical rooms, cooling plant and support spaces, that are adapted to each site rather than redesigned.",[],{"_key":167,"_type":29,"children":168,"markDefs":173,"style":37},"44ea8ec4a742",[169],{"_key":170,"_type":33,"marks":171,"text":172},"ed3ff57a1663",[],"That only works if the design data behind those blocks is structured, versioned and reusable. Most teams have the intent but not the infrastructure. The \"standard\" lives in a folder of files, and knowing which version was used on which site depends on who you ask.",[],{"_key":175,"_type":29,"children":176,"markDefs":185,"style":37},"de1d0bbd7194",[177,181],{"_key":178,"_type":33,"marks":179,"text":180},"d86f94c6bcca",[85],"How Speckle helps.",{"_key":182,"_type":33,"marks":183,"text":184},"6fc91282e2d1",[]," Speckle captures model data from your authoring tools as structured, versioned objects rather than static files. Every change is tracked, so you can see exactly how a site design differs from the reference design and why. Teams can reuse validated components across sites and start each new project from a known-good baseline, instead of from a copy of the last job.",[],{"_key":187,"_type":29,"children":188,"markDefs":197,"style":37},"69a8ca9a4456",[189,193],{"_key":190,"_type":33,"marks":191,"text":192},"66c8bfcfdf38",[85],"The time saving:",{"_key":194,"_type":33,"marks":195,"text":196},"8a8a3d5231da",[]," shorter early design phases and fewer surprises when a \"standard\" turns out not to be standard.",[],{"_key":199,"_type":29,"children":200,"markDefs":205,"style":70},"5bcd8c020c63",[201],{"_key":202,"_type":33,"marks":203,"text":204},"d9686d604101",[],"Practice 2: Work from live model data, not exports",[],{"_key":207,"_type":29,"children":208,"markDefs":213,"style":37},"9711c1a1d9ac",[209],{"_key":210,"_type":33,"marks":211,"text":212},"6fc1e13e1d77",[],"On a fast-track project, the gap between \"the model changed\" and \"everyone knows the model changed\" is where time disappears. Teams wait for weekly uploads. Coordination runs against last week's export. A fabricator builds from a model that was superseded days ago.",[],{"_key":215,"_type":29,"children":216,"markDefs":221,"style":37},"549e8cb9e1b7",[217],{"_key":218,"_type":33,"marks":219,"text":220},"56fcfef6e337",[],"The fix is to work from the source. Instead of passing around derivatives, such as exported files, screenshots and spreadsheets pulled from models, every stakeholder should see the same current data, in the tools they already use.",[],{"_key":223,"_type":29,"children":224,"markDefs":232,"style":37},"1bdf3f188bd2",[225,228],{"_key":226,"_type":33,"marks":227,"text":180},"bea29c1cfa80",[85],{"_key":229,"_type":33,"marks":230,"text":231},"f3d8dd54dd58",[]," Speckle is a vendor-neutral data layer that connects the tools your project team already uses, including Revit, Rhino, Tekla, Navisworks, IFC and common data environments like Autodesk Construction Cloud and Bentley ProjectWise. Model data flows between them as structured objects, so the MEP engineer's latest routing is visible to the structural team and the trade contractor without a manual export-and-upload cycle.",[],{"_key":234,"_type":29,"children":235,"markDefs":240,"style":37},"ce64201f129f",[236],{"_key":237,"_type":33,"marks":238,"text":239},"fe7ade3be319",[],"Speckle's architecture also removes the practical model size limits that slow teams down on large, dense facilities. Full campus models load in the browser, so owners, PMs and site teams can review the current state without specialist software. And with Speckle Intelligence, anyone can ask questions of the live 3D model in plain language, such as \"how many CRAH units are in Hall 2?\" or \"which electrical rooms changed since last Tuesday?\", and get answers from the model itself, not from a report that's already out of date.",[],{"_key":242,"_type":29,"children":243,"markDefs":251,"style":37},"037cdcb8f59e",[244,247],{"_key":245,"_type":33,"marks":246,"text":192},"e4c9d71fc87c",[85],{"_key":248,"_type":33,"marks":249,"text":250},"013f606dd4a3",[]," days removed from every information handoff, multiplied across every handoff on the project.",[],{"_key":253,"_type":29,"children":254,"markDefs":259,"style":70},"b3e92ba54d42",[255],{"_key":256,"_type":33,"marks":257,"text":258},"367f6c335125",[],"Practice 3: Put coordination on agents and supercharge your best VDC managers",[],{"_key":261,"_type":29,"children":262,"markDefs":267,"style":37},"8cb9abf2f85b",[263],{"_key":264,"_type":33,"marks":265,"text":266},"2dff557a70b1",[],"On most data center projects, coordination capacity is the constraint you can't hire your way out of. There aren't enough experienced VDC managers, and the ones you have spend too much of their time on work that is necessary but repetitive: setting up and running clash tests, filtering thousands of results down to the few hundred that matter, grouping them, assigning them to the right trade, and checking whether last week's issues were actually fixed.",[],{"_key":269,"_type":29,"children":270,"markDefs":275,"style":37},"906e12f47e87",[271],{"_key":272,"_type":33,"marks":273,"text":274},"72d0a0209469",[],"That work is exactly where AI agents can take on real jobs, not just answer questions.",[],{"_key":277,"_type":29,"children":278,"markDefs":286,"style":37},"55b510eb296e",[279,282],{"_key":280,"_type":33,"marks":281,"text":180},"95c341f2871b",[85],{"_key":283,"_type":33,"marks":284,"text":285},"afd0da0f649b",[]," Speckle is building an agentic platform that executes critical coordination work directly on live project data. Agents run clash detection continuously as models update, rather than waiting for the weekly coordination cycle. They triage and group results, filter out noise and known tolerances, route issues to the responsible discipline, and track them through to resolution. Your VDC manager reviews, decides and leads the coordination meeting, with the groundwork already done.",[],{"_key":288,"_type":29,"children":289,"markDefs":294,"style":37},"26c2c5778be0",[290],{"_key":291,"_type":33,"marks":292,"text":293},"c6c06f382832",[],"The goal isn't to replace your coordination team. It's to give your best VDC manager the output of a much larger team, so you can do more work with the people you already have. On a data center program, that means:",[],{"_key":296,"_type":29,"children":297,"level":91,"listItem":92,"markDefs":302,"style":37},"8ac420578f1c",[298],{"_key":299,"_type":33,"marks":300,"text":301},"971051ba247e",[],"Clashes found within hours of a model change, not at the next coordination meeting",[],{"_key":304,"_type":29,"children":305,"level":91,"listItem":92,"markDefs":310,"style":37},"09459c34d9e3",[306],{"_key":307,"_type":33,"marks":308,"text":309},"f8483e85abdd",[],"Coordination meetings spent on decisions, not on sorting through results",[],{"_key":312,"_type":29,"children":313,"level":91,"listItem":92,"markDefs":318,"style":37},"928ae369217f",[314],{"_key":315,"_type":33,"marks":316,"text":317},"cb4a6238ce96",[],"Prefabrication packages released sooner, with more confidence",[],{"_key":320,"_type":29,"children":321,"level":91,"listItem":92,"markDefs":326,"style":37},"1831097b805a",[322],{"_key":323,"_type":33,"marks":324,"text":325},"fd6b9f25bd13",[],"The ability to take on more sites in parallel without a matching increase in VDC headcount",[],{"_key":328,"_type":29,"children":329,"markDefs":337,"style":37},"5a1913a7d628",[330,333],{"_key":331,"_type":33,"marks":332,"text":192},"189ad3fa31ac",[85],{"_key":334,"_type":33,"marks":335,"text":336},"d21c7843c100",[]," shorter coordination cycles on the critical path, and capacity to run more projects at once.",[],{"_key":339,"_type":29,"children":340,"markDefs":345,"style":70},"cb27e372741b",[341],{"_key":342,"_type":33,"marks":343,"text":344},"10ff18716ea3",[],"Practice 4: Make model QA/QC continuous, not a gate",[],{"_key":347,"_type":29,"children":348,"markDefs":353,"style":37},"82d5576a23f1",[349],{"_key":350,"_type":33,"marks":351,"text":352},"bc3b656ce475",[],"On most projects, model quality is checked at milestones. A model is submitted, reviewed, sent back with comments, fixed and resubmitted. Each loop can take a week or more, and problems found late tend to be the expensive ones: missing parameters on equipment, components that don't match the reference design, or models that don't meet the owner's standards for handover and operations.",[],{"_key":355,"_type":29,"children":356,"markDefs":361,"style":37},"1f0a3206d98f",[357],{"_key":358,"_type":33,"marks":359,"text":360},"0a6a810a99d2",[],"On a data center, those gaps don't just delay a milestone. They delay procurement, fabrication and commissioning, and they surface again at handover when the owner's operations team needs reliable asset data from day one.",[],{"_key":363,"_type":29,"children":364,"markDefs":372,"style":37},"61c4c028fdbd",[365,368],{"_key":366,"_type":33,"marks":367,"text":180},"5faba537a0bd",[85],{"_key":369,"_type":33,"marks":370,"text":371},"665858706df5",[]," Speckle checks model data against your standards every time it's published, not only at submission. Agents for data conditioning and model QA/QC flag missing or inconsistent information as it happens, so issues are fixed by the person who created them, while the context is still fresh. Contractors running large, distributed modeling teams get a live view of model quality and progress across every team and trade, without waiting for a manual review.",[],{"_key":374,"_type":29,"children":375,"markDefs":383,"style":37},"06427d89d4c9",[376,379],{"_key":377,"_type":33,"marks":378,"text":192},"8fac965d2327",[85],{"_key":380,"_type":33,"marks":381,"text":382},"09c63a045aad",[]," fewer resubmission loops, cleaner prefabrication data, and a handover that doesn't turn into a months-long data cleanup exercise.",[],{"_key":385,"_type":29,"children":386,"markDefs":391,"style":70},"ed5d670cea2f",[387],{"_key":388,"_type":33,"marks":389,"text":390},"aea356c4b58d",[],"Practice 5: Give owners intelligence across the whole portfolio",[],{"_key":393,"_type":29,"children":394,"markDefs":399,"style":37},"91320002612f",[395],{"_key":396,"_type":33,"marks":397,"text":398},"8601c79e90f1",[],"Owners building data centers at scale are increasingly asking questions their delivery partners struggle to answer quickly. How does this site compare to the last five? Where is our design drifting from the standard? Which configurations get to power fastest? What will this site look like at the next phase of expansion?",[],{"_key":401,"_type":29,"children":402,"markDefs":407,"style":37},"e8368de338aa",[403],{"_key":404,"_type":33,"marks":405,"text":406},"44e87dd92fb0",[],"Those answers sit inside dozens of models, held by different consultants and contractors, in different tools. Pulling them together usually means a manual data request and a spreadsheet that's stale on arrival.",[],{"_key":409,"_type":29,"children":410,"markDefs":418,"style":37},"e0d855e58a02",[411,414],{"_key":412,"_type":33,"marks":413,"text":180},"dfa88904d9b1",[85],{"_key":415,"_type":33,"marks":416,"text":417},"3af0accf69b4",[]," Because Speckle structures model data consistently across projects and tools, owners can benchmark designs across their entire portfolio, not just one project at a time. That data can feed directly into the analytics and AI platforms owners already run, such as Databricks or Microsoft Fabric, so design data sits alongside cost, schedule and operational data instead of being locked away in project files. Owners own their data, regardless of which partner or tool produced it.",[],{"_key":420,"_type":29,"children":421,"markDefs":426,"style":37},"a7b46226cc6d",[422],{"_key":423,"_type":33,"marks":424,"text":425},"4a3a4387cd84",[],"That intelligence also speeds up the earliest, highest-leverage decisions. Faster site evaluation and test-fitting, based on what has actually been built before, shortens the path from land to a design ready for permitting.",[],{"_key":428,"_type":29,"children":429,"markDefs":437,"style":37},"9e726cfa46c1",[430,433],{"_key":431,"_type":33,"marks":432,"text":192},"5ead92fe2a96",[85],{"_key":434,"_type":33,"marks":435,"text":436},"eacd8aebd5ef",[]," faster early-stage decisions, faster answers from delivery partners, and a design standard that gets better with every site.",[],{"_key":439,"_type":29,"children":440,"markDefs":445,"style":70},"d12bf62d5d2e",[441],{"_key":442,"_type":33,"marks":443,"text":444},"e65eecac39d7",[],"What this means for owners and for contractors",[],{"_key":447,"_type":29,"children":448,"markDefs":453,"style":37},"e569d06300b5",[449],{"_key":450,"_type":33,"marks":451,"text":452},"3100b35fcec1",[],"Owners and delivery teams feel schedule pressure from different sides, but they're solving the same problem. Here's how the value lands for each.",[],{"_key":455,"_type":456,"rows":457},"6c3288e73fd2","table",[458,465,471,477,483],{"_key":459,"_type":460,"cells":461},"d0d0ad9b-0ee9-42d1-9c4b-34c1b2959ca0","tableRow",[462,463,464],"","Owners and developers","General contractors and delivery teams",{"_key":466,"_type":460,"cells":467},"0a2efc26-38c0-43bf-a321-2d261f15beeb",[468,469,470],"What's at stake","Capacity online by a committed date; speed of rollout across many sites","Hitting aggressive milestones; winning the next program from the same client",{"_key":472,"_type":460,"cells":473},"3b1e215c-34d5-4512-a4d8-7693173d418b",[474,475,476],"Biggest time drain","Waiting on partners for answers; re-learning lessons on every site","Coordination cycles, rework, and late changes before prefabrication",{"_key":478,"_type":460,"cells":479},"b4a206e6-87ed-495b-9466-fd1eb54cbd80",[480,481,482],"Where Speckle helps most","Portfolio benchmarking, design standards, and owning project data across partners","Agentic clash detection and coordination, continuous QA/QC, live model access for site teams",{"_key":484,"_type":460,"cells":485},"932909f2-a7f3-44d8-86ac-2f063d605773",[486,487,488],"What changes","Faster decisions and a design standard that improves with every site","More sites delivered in parallel with the same VDC team",{"_key":490,"_type":29,"children":491,"markDefs":496,"style":37},"ba49de0ec6fb",[492],{"_key":493,"_type":33,"marks":494,"text":495},"5e24b0438577",[],"For contractors in particular, speed is a competitive advantage. Owners are consolidating their programs with the partners who can prove they deliver fastest and most predictably. The contractor that can show live coordination status, model quality and progress across every site, on demand, is the one that earns the next campus.",[],{"_key":498,"_type":29,"children":499,"markDefs":504,"style":70},"ff92c386cb13",[500],{"_key":501,"_type":33,"marks":502,"text":503},"964c7a044f4d",[],"Where to start",[],{"_key":506,"_type":29,"children":507,"markDefs":512,"style":37},"3069c2692a17",[508],{"_key":509,"_type":33,"marks":510,"text":511},"2202d5e4621b",[],"You don't need to transform your whole delivery process at once. The teams moving fastest start with one live project and one clear bottleneck.",[],{"_key":514,"_type":29,"children":515,"level":91,"listItem":524,"markDefs":525,"style":37},"141e0db6ffb5",[516,520],{"_key":517,"_type":33,"marks":518,"text":519},"62aac48eb4cb",[85],"Map your critical path.",{"_key":521,"_type":33,"marks":522,"text":523},"b36e985fb0b8",[]," Identify where model data holds up the schedule today: coordination cycles, resubmissions, prefabrication sign-off or partner data requests.","number",[],{"_key":527,"_type":29,"children":528,"level":91,"listItem":524,"markDefs":537,"style":37},"4c5a9e2f8292",[529,533],{"_key":530,"_type":33,"marks":531,"text":532},"cac02428983a",[85],"Connect the tools you already use.",{"_key":534,"_type":33,"marks":535,"text":536},"2a2fe1ccb58f",[]," Bring live model data from your project's authoring tools and CDE into one place, without asking anyone to change how they work.",[],{"_key":539,"_type":29,"children":540,"level":91,"listItem":524,"markDefs":549,"style":37},"4b969c6777df",[541,545],{"_key":542,"_type":33,"marks":543,"text":544},"0a66bc535177",[85],"Automate one coordination workflow.",{"_key":546,"_type":33,"marks":547,"text":548},"a8e03f5377ec",[]," Start with clash detection and triage on a live project, and measure how much faster issues are found and closed.",[],{"_key":551,"_type":29,"children":552,"level":91,"listItem":524,"markDefs":561,"style":37},"9c71ed3782c5",[553,557],{"_key":554,"_type":33,"marks":555,"text":556},"7e8af00aaae1",[85],"Set your standard once.",{"_key":558,"_type":33,"marks":559,"text":560},"a42148c83268",[]," Define the model quality checks that matter for prefabrication, handover and operations, and run them continuously.",[],{"_key":563,"_type":29,"children":564,"level":91,"listItem":524,"markDefs":573,"style":37},"bfc8b2df7a30",[565,569],{"_key":566,"_type":33,"marks":567,"text":568},"eb0ff6ec7391",[85],"Scale across the program.",{"_key":570,"_type":33,"marks":571,"text":572},"1a0f9321ea0d",[]," Once it works on one site, roll it out across the portfolio so every new project starts faster than the last.",[],{"_key":575,"_type":29,"children":576,"markDefs":581,"style":37},"ecf02494a6ad",[577],{"_key":578,"_type":33,"marks":579,"text":580},"14cc60ec8e3d",[],"On data center projects, the winners won't be the teams that work harder. They'll be the teams whose data moves as fast as their schedule.",[],{"_key":583,"_type":29,"children":584,"markDefs":602,"style":37},"1a9dd3d4ea49",[585,589,593,598],{"_key":586,"_type":33,"marks":587,"text":588},"ad63e50efa17",[85],"Building or delivering data centers?",{"_key":590,"_type":33,"marks":591,"text":592},"1250035339e8",[]," Speckle is working with a small group of AEC leaders to put agentic coordination to work on live projects. ",{"_key":594,"_type":33,"marks":595,"text":597},"a3b82f132704",[596],"e331a2fbb26e","Talk to our team",{"_key":599,"_type":33,"marks":600,"text":601},"30a031d41e0e",[]," about what that could look like on your next site.",[603],{"_key":596,"_type":604,"href":605},"link","https://speckle.systems",false,{"_type":13,"asset":608},{"_ref":609,"_type":16},"image-5edbd5dd56605164b6db76b88ab6397d56edf02a-2400x1254-png","2026-10-30",{"_type":23,"current":612},"speed-to-power-how-data-center-teams-are-compressing-delivery-with-connected-data-and-ai-agents","Five practices we see working on fast-track data center programs, and how Speckle's data platform and new agentic capabilities support each one.","Speed to Power: How Data Center Teams Are Compressing Delivery with Connected Data and AI Agents",1793370352859]