When Orders Pile Up and You're Short-Handed
If inquiries keep coming in while you still haven't delivered on last month's orders, you start wondering whether to pause new sales altogether. Any solo entrepreneur has lived through this scene at least once. Something worth watching happened at Microsoft last week that speaks directly to this moment: a company that used to ask whether AI demand was coming has started asking how fast it can deliver on the orders it already has. When the question changes, so does the number a company checks every day.
What Happened at Microsoft Last Week
Start with the numbers. Microsoft's backlog — contracts it has signed with customers but hasn't yet booked as revenue — has reached $678 billion. Roughly 30% of that is assumed to convert to cash within the next 12 months. Azure grew 43% last quarter, and Microsoft Cloud revenue came in at $59.3 billion. Meanwhile, the company paid out $6.8 billion in dividends against $27.2 billion in AI-related spending. In other words, most of the cash Microsoft brought in went not to shareholders but to data centers and AI infrastructure.
What underwrites this spending is the customer list. In just four months, Microsoft signed a run of contracts across strikingly different sectors: four Frontier Company clients in May, the UK's NHS in July, the Pentagon in August, and Saudi Arabia's HUMAIN this week — enterprise, healthcare, defense, and the Middle East, one after another. Add to that Meta's confirmation as a major Azure AI customer, and the list now includes clients spending hundreds of millions of dollars a year.
Why a Giant Changed Its Question
Sum up this pattern in one line: order first, invest later. Microsoft locks in the contract first, then spends against it. It's the $678 billion backlog that lets the company justify AI spending running at roughly four times its dividend payout. The debate over whether demand exists is already over.
It's worth pausing on what a backlog actually is. It isn't the result of halting sales — it's the result of pushing delivery dates further out while continuing to take on contracts. If only about 30% converts to cash within 12 months, that means most of the rest is scheduled even further out. This isn't a build-first, sell-later model; it's a model where what will sell gets locked in before it gets built.
The catch comes next. A contract doesn't turn into money the day it's signed. Revenue only gets booked once the data center is built, the hardware is installed, and the customer actually starts using it. That's why Microsoft has shifted its focus from winning orders to fulfilling them — because how fast the backlog converts into revenue is what drives quarterly results. And that delivery speed depends on more than construction timelines; it also hinges on variables like each customer's contract term, price renegotiations, and delivery delays.
How the customer base grows is worth watching too. Large deals move through a roughly three-year cycle: announcement, initial rollout, and large-scale expansion. By that math, the customers announced in May won't generate meaningful revenue until 2027 or 2028. Which means if Microsoft fails to deliver on the orders it holds now on schedule, the next round of expansion contracts slips right along with it.
The Principle a Solo Founder Can Borrow
The scale is nothing alike, but the principle is the same. Money comes from delivery, not from the contract itself. Buy equipment or hire staff the moment inquiries pick up, and you're left holding fixed costs the moment demand cools. Stop taking orders because you're backed up, and you're staring at an empty calendar three months later. Take a one-person furniture workshop as an example: instead of turning away inquiries because orders are backed up, it's better to push the delivery date out a few weeks and collect a deposit. That total deposit amount is your backlog, and it's the number that tells you whether to bring on help or invest in more equipment. The order-first, invest-later strategy is the path between these two traps, and it runs in this order:
Turn inquiries into deposit-backed orders to lock in real demand, invest in subcontractors or equipment only within that amount, and check the delivery dates written into your contracts every week — that's what creates room to take on the next round of expanded orders. Just as Microsoft tracks what share of its backlog converts to cash, a solo founder should track, as a number, what share of this month's orders actually got delivered. Execution speed is the next bottleneck, and it needs to be designed for at the very moment orders start piling up.
One Thing to Try This Week
This week, list out every order you currently hold and build a table showing each one's delivery date and how much of it you've delivered so far. If that table shows more backlogged orders than new ones, it's not time to stop selling — it's time to change how you deliver.
