Raise the smallest facility that meets your goals — you can always upsize, you can never un-pay the fees. Size to haircut projections, not the plan you pitch VCs. Then negotiate in writing: specific asks get specific changes; the terms that move are pricing, fees, covenant levels, no-call length, and warrant size; the terms that don't are the structural kit. Watch for late givebacks, and read every new draft as a new document — because that's what it is.
The first big mistake: taking the biggest number on the table
When term sheets show up, the temptation is to take the biggest headline. A $100M facility sounds like a $100M facility in the press release. Founders are trained to maximize the number.
Resist it. Raise the smallest facility that meets your goals. You can always upsize. You can never un-pay the fees.
The mechanics are all in Part 6 — you pay for commitment, not usage. Structuring fees scale with committed amount. Unused fees accrue from day one on everything you haven't drawn. And the softer costs are worse: facilities carry draw windows, minimum utilization expectations, and sometimes mandatory draw schedules, so an oversized facility doesn't just cost money — it puts you behind plan with your lender every month. That's the relationship dynamic you least want heading into your first trigger conversation, your first amendment request, your first anything. An undersized facility makes you the borrower who keeps beating plan. Guess which one gets the good amendment.
There's a covenant angle too, and it's under-appreciated: your projections become your covenants. The origination plan you show during diligence calibrates the facility size, the draw schedule, the minimum-utilization expectations, and often the trigger levels. Pitch your VC-deck growth case to a lender and you have created a legal obligation to hit your VC-deck growth case. This is the gun pointed at your own foot.
The practical move: haircut your projections, then size the facility to the haircut. Take your origination plan, cut it meaningfully — a third is not paranoid — and raise against the smaller number: a facility you have clear line of sight to drawing in full. If you beat plan, congratulations: upsizing a performing facility is the easiest conversation in capital markets. Lenders want to put more money behind a book that's working; predictable deployment into proven performance is their whole model. But downsizing an oversized facility, or bleeding unused fees while you limp toward a draw target? There's no good version of that conversation.
Small facility, fully drawn, performing, upsized. That's the flywheel. It beats a big facility you're paying rent on to impress people.
a third is not paranoid→ Size the facility to the haircut→ Draw it fully→ Beat plan→ Upsize ↺
the easiest conversation in capital markets
Test facility sizes against your haircut draw curve before you pick a number.
Negotiate on paper
Now, how to actually move the terms. The single highest-leverage habit is embarrassingly simple: write your asks down. A one-page document per lender — term, current position, your ask, your reasoning — shared with your counsel and, in cleaned-up form, with the lender.
Why it works: your asks stop being vibes and become an agenda. Nothing gets forgotten across the six-week negotiation. Your lawyer bills fewer hours reconstructing what you wanted. And on the other side, a specific, reasoned ask is something a deal lead can take to committee — "borrower requests the floor at 2.5% given the forward curve, and the no-call at 18 months with a volume-based step-off" is a memo paragraph; "they feel the terms are aggressive" is nothing. In my experience the hit rate on written, specific, reasoned asks is startlingly high. The hit rate on general grumbling is zero.
What moves. Across every negotiation I've seen or heard credibly described, the same terms flex:
- Pricing — spread, and especially the floor. Floors are half-forgotten by everyone; asks to cut them often just... work.
- Upfront and unused fees — frequently halved or restructured (deferred to later tranches, waived on the first tranche, converted to draw fees) when challenged.
- Covenant levels — minimum-cash amounts, trigger thresholds, and above all ramp-up holidays: covenant relief until the book reaches critical mass, granted routinely when asked "given the nascency of the program." Ask in exactly those words.
- No-call length and shape — hard periods shortened, or fitted with performance-based step-offs ("fee drops away above $X of outstandings").
- Warrant size and structure — the difference between the opening ask and the close is often the largest single dollar item in the deal. Milestone vesting (pay for capital that actually funds) is a standard get.
- Exclusivity scope — capacity rights stepped down sooner, ROFRs narrowed to ROFOs, escape hatches added (the exclusivity-dies-when-funding-stops clause from Part 5).
- Cure periods, notice periods, definitions — cheap for the lender to give, enormously valuable to you at 2am some future Tuesday.
What doesn't move. The structural kit: the SPV, the first-priority lien, the account control, the backup servicer, the bad-acts guaranty, the reporting obligations, the audit rights. These are the asset class, not this lender's aggression. Spending negotiating capital here marks you as someone who doesn't know the market — and worse, it spends credibility you need for the asks that can move. Concede the kit early and loudly; bank the goodwill.
Concede the kit early and clearly; spend your negotiating capital on the asks that can move.
Watch the late givebacks. Between the first draft and the signature draft, terms drift in both directions — your asks get granted, and new lender-favorable items quietly appear: a fee that wasn't in draft one, a tightened definition, a new condition on the step-up you negotiated. This isn't bad faith; it's how their committee approved your concessions — by charging for them elsewhere. The defense is mechanical: diff every draft. Make your counsel produce a redline against the previous version and a running list of changes-since-first-draft. Read every parenthetical; a "ninety (60) days" drafting conflict is funny until it's your cure period in a workout.
Read the long form as a new document. The friendly two-page indicative terms and the fifteen-page full term sheet are different genres. The long form is usually where pricing improves (they've done more work, they want the deal) and where recourse and control quietly expand — personal guaranties appearing, ROFRs materializing, key-person clauses arriving, deposits becoming non-refundable. Both drifts are normal. Neither is binding until you sign. Re-run the full Part 5 read on every draft as if you'd never seen the deal before.
And use your one moment. Your leverage peaks in the window when you hold multiple live sheets and haven't signed anything. Every mechanism in this section works ten times better inside that window than after it. Which is the entire subject of the next part — because the day you sign, the auction ends, and a different game begins.