Resources.
Straight answers for anyone buying technology, from data center space to AI. Free, open and dated.
Guides.
The old way of buying data center space is dead
Why calling data center providers for capacity reports fails at record-low vacancy, how space really frees up, and why bidding every provider at once wins now.
Make providers compete
Why buying data center space one provider at a time is slow and leaves you no leverage. What the sales runaround looks like, and how to get providers competing at once for a better price, in days, not months.
Why directories aren't neutral
Provider websites, data center directories, AI answers and salespeople each lean a predictable way. Here's what each gets wrong, real examples from our own directory, and what a view on your side looks like.
How to run a data center bid
How to build your plan for data center space first, invite the right providers, and get quotes you can actually compare.
How much power to plan for
How to turn a rack count into a kW number for data center space (colocation). Measured draw versus the label, what power per rack changes, and why power sets your price.
Racks vs cage vs suite vs hall
The four ways data center space (colocation) is sold: racks, a cage, a private suite, or a whole hall. What separates them, what "retail" and "wholesale" mean, and how to tell which you need.
What your contract should include
The checklist for a data center (colocation) contract: term and renewal, the annual increase, connection fees, on-site support rates, and what an SLA has to say to mean anything. Structure only, no prices.
Carrier-neutral vs carrier-specific
What a carrier-neutral data center really means, why almost every provider claims it, how to check which networks are in your building, and how connection fees work.
Planning your data center move
A step-by-step plan for moving into data center space (colocation): what to settle before you sign, why running both sites beats a single cutover, and realistic timelines.
How to get data center pricing
Why no data center provider publishes a price list, why you shouldn't trust a price you find online, and how to get comparable quotes without the sales-call runaround.
How to negotiate data center space
What's actually negotiable in a data center (colocation) contract, why the first quote is only an opening number, and why competition, not a tougher tone, is your leverage.
How to compare data center quotes
How to line up data center (colocation) quotes that arrive in different shapes: per kW or per cabinet, term, yearly increases, connection fees and power billing. Plus what really lowers your cost.
Beyond Equinix & Digital Realty
Equinix, Digital Realty, QTS and CyrusOne are the four data center names everyone hears. We track 124 providers. The one that fits you best is often one you've never heard of.
What AI answers can't tell you
AI answers about data center space already name mid-size providers and directories. What they can't tell you is who has space or what it costs today. Here's what they give you, and how to get the rest.
Data center glossary
Plain definitions of the data center (colocation) terms in quotes, contracts and provider marketing, including the ones most often misused, like Uptime Tier and cloud on-ramps.
Markets.
Markets we cover
All 152 markets on one page, grouped by region. Find yours in ten seconds, then get the providers there competing for your project.
Northeast & Mid-Atlantic data centers
The world's most connected data center market lives here, and it's short on power. Getting connected is the slow part.
Southeast data centers
Lower power costs, Atlanta's carrier hub and Miami's door to Latin America. Your big call is inland or coast.
Central US data centers
Reach the whole country at lower power costs than either coast. Watch the strained grids in Ohio and Colorado.
Southwest data centers
Power at a third to half of California's price. The catch: cheap power isn't the same as available power.
Northwest data centers
Silicon Valley's network reach, or cheap inland power and room to grow. The Northwest's power-cost gap decides it.
Canada data centers
Quebec hydro is some of the cheapest data center power in North America. Quebec, BC and Alberta all ration big new connections.
Europe data centers
Power, more than price, is Europe's bottleneck. Dublin, Amsterdam, London and Frankfurt have all restricted new grid connections.
Asia-Pacific data centers
Singapore rations new capacity in fixed, green-gated rounds. Johor, across the border, takes the overflow, and it's tightening too.
Latin America data centers
São Paulo anchors the region. Querétaro grows fastest, and it's running into power and water limits.
Middle East & Africa data centers
Here the first question is whether the grid stays on. Dubai's rarely blinks; Lagos runs on backup.
Phoenix data centers
Who runs data centers in metro Phoenix, including providers most shortlists miss. Which buyer each one fits, what the power squeeze means for you, and two big names that aren't here at all.
Northern Virginia (Ashburn) data centers
The world's biggest data center market, mapped by the town each building is really in: Ashburn, Sterling, Reston, Manassas and Herndon. Who's there, where the cloud connects, and what Loudoun's power squeeze means for you.
Dallas data centers
Who runs data centers across Dallas–Fort Worth, including the mid-size providers in Plano, Richardson and Allen that rarely make a shortlist. The Infomart's three competing providers, confirmed connections, and which capacity isn't built yet.
Los Angeles data centers
Data center space in Los Angeles, built around One Wilshire. Who runs what across downtown, the South Bay and Orange County, what's confirmed about cloud connections, and where our verified list stops.
Cloud.
Which cloud is best for an AI product?
How training and inference change the platform, why your data often picks the cloud for you, and what to do if the data has to stay in the country.
Should we bring workloads back from the public cloud?
Which workloads are worth bringing back from the public cloud, where they can go, and what the move costs and saves.
Which GPU cloud provider is best for us, and how do we compare them?
What to compare in a GPU cloud besides the hourly price, how training and inference change the choice, and what the contract terms hide.
How do we get GPU capacity, and who actually has it?
The three ways to get GPUs, what "capacity" really means, and how to get in line without committing to more than you need.
We need to lower our AWS, Google or Azure bill. How do we get better terms?
Better terms on AWS, Azure or Google: private pricing, reserved instances and cost optimization, negotiated from your real workload, starting now.
Broadcom raised our VMware bill. How do we get leverage back without replatforming?
Keep VMware and get the licensing under control through a top-tier VMware service provider, or replatform when it fits. Two paths, both bid for you.
Network.
We still have copper lines. How do we upgrade to fiber?
Which of your sites can get fiber, what happens to alarm, elevator and fax lines, and what carriers will and won't commit to.
Can we cut our network costs without changing providers?
Where a network bill leaks, how to make your current carrier compete for the renewal, and when switching is worth it after all.
How do we consolidate dozens of sites and carriers onto one contract?
What goes in the inventory, whether one provider can really cover every site, and how to move without paying twice.
How do we deploy AI across our network, securely and in every region?
Private paths to the clouds that run your models, one security stack for every user, and keeping regulated data in its region.
How do we secure AI agents on our network?
Why agents are a different security problem, the controls that apply today, and what to ask anyone who says they secure agents.
MPLS, internet, SD-WAN or SASE: how should they work together?
What MPLS, internet, SD-WAN and SASE each do for which applications, how they work in tandem, and the roadmap to one network.
How do we give employees safe access to AI tools?
What your network can see and stop when people use AI tools, whether to block or allow them, and how it fits your AI policy.
Satellite or 5G: which backup line should each site get?
What a backup line has to survive, how 5G, fixed wireless, satellite and a second wired line differ, and how to run backup across many sites.
Security.
Our cyber insurance renewal wants MDR, MFA and tested backups. Where do we start?
What cyber insurers ask for at renewal, which gaps decide the quote, and how to show the insurer it's done.
What security gaps do we have, and do we need an assessment to find them?
What a security assessment covers, how to read the results, and what happens to the gaps after the report.
We have too many security tools. How do we cut down to fewer vendors?
One platform or separate tools, what to keep and what to retire, and whether one provider can run all of it.
We need a SOC and a NOC. What are our options, and can AI run them?
In-house, co-managed or outsourced; whether one provider should run the SOC and the NOC; and what AI in a SOC does today.
What does a 24/7 MDR service cost, and what drives the price?
How MDR is priced, what makes the price go up or down, and whether you need MDR, an MSSP or your own SOC.
XDR, MDR, EDR: what's the difference, and which do we buy?
What EDR, XDR and MDR each do, whether you need the tool, the team or both, and where a SIEM fits.
Voice.
What does a cloud phone system cost per user, and what drives it?
How a cloud phone system is priced, what moves the price per user, and the one-time costs that sit outside it.
We have phone systems everywhere. How do we consolidate them, and the vendors, into one?
What goes in the voice inventory, whether one provider can cover every site, country and Teams user, and how to move site by site.
How do we add phone calling to Microsoft Teams?
The three ways to put calling in Teams, what going direct with Microsoft leaves you owning, and the two kinds of provider that do it with you.
We need a new phone system. What are our best options?
Three reasons for a new phone system (end of life, an upgrade, consolidation), what to decide before anyone bids, and how to keep your numbers and the front desk working.
Our cloud phone provider is hard to work with. How do we switch?
What to measure before you leave a cloud phone provider, what your contract lets you do, and how to move without the phones going down.
Contact center.
Agentic AI in the contact center: what's real and what's a demo?
What agentic AI means in a contact center, how to tell a real deployment from a demo, and what a proposal should prove.
AI for customer experience: which kinds fit our contact center, and how should we pay?
Which kinds of contact center AI fit which customer experience jobs, the four ways each is priced, and how to get the right one at the right price.
How do we use AI in our contact center, and where do we start?
What AI does in a contact center, how to choose a first pilot on the platform you run, and which numbers show whether it's working.
How do we move our contact center to the cloud, AI included?
What to decide before choosing a contact center platform, whether AI comes with the move or after, and how to cut over without dropping calls.
Can we modernize our contact center without ripping and replacing it?
What can sit on top of a contact center platform you keep, what to move first, and when replacing the platform beats building on it.
We're not getting full value from our contact center platform. How do we fix it?
Where contact center value leaks, what you can fix without switching platforms, and when to bid the renewal.
AI.
We pay for Copilot, ChatGPT Enterprise and Claude. Do we need all three?
Why companies pay for several AI subscriptions, how to decide which to keep, and whether one governed way in can replace some seats.
How do we pick an AI model provider for our product?
How AI model pricing is structured, what data terms to read about prompts and outputs, and why the region a model runs in matters.
How do we get our AI spend under control?
Where AI spend hides, whether to pay for seats or usage, and what one governed way in looks like.
How do we get an AI pilot into production?
Why AI pilots stall, which job should go first, and what an AI proposal should commit to before you sign.
How do we get employees to use AI safely, without the cost running away?
What goes in an AI usage policy, why one approved way in beats a list of banned tools, and how to tell whether adoption is working.
What is an AI gateway, and do we need one?
What an AI gateway does, when you need one and when it's overkill, and how it differs from a security tool.
Wireless & IoT.
How do we connect thousands of IoT devices without getting locked to one carrier?
eSIM, multi-carrier SIM or one carrier, what drives the per-device price, and what every IoT connectivity proposal has to answer.
We have thousands of phone lines and the bill makes no sense. How do we fix it?
Where a corporate mobile bill goes wrong, whether to switch carriers or renegotiate, and what managed mobility adds.
We have thousands of IoT devices deployed. How do we manage them and keep the cost down?
What an IoT connectivity management platform does, where fleet cost leaks, and how to switch carriers on devices already in the field.
How do we manage AT&T, Verizon and T-Mobile under one plan?
Why companies end up on several mobile carriers, how to decide who's on which, and how to keep the rates right between renewals.
What is private 5G, and when does it beat Wi-Fi?
What private 5G and CBRS are, where private 5G beats Wi-Fi and where it doesn't, and who builds and prices it.
What does a full managed mobility service include?
What a full managed mobility service covers, whether one provider or two should run it, and how it's priced.