Negotiating Salary at Startups and Big Corps | Generated by AI
Question: How do I negotiate salary and promotion, whether at a startup or a big corp?
Answer:
The mechanics differ because the constraint differs. Big corp negotiates against bands and comparables; startups negotiate against equity, runway, and your perceived asymmetric value. Here’s the playbook for each.
Universal principles (apply everywhere)
Leverage = your BATNA, not your need. Your mortgage doesn’t matter in the room — never reveal it. The only thing that moves numbers is a credible alternative. A competing offer is the single highest-leverage artifact you can hold. Generate one even if you don’t intend to leave.
Anchor high, let them name first if you can. If forced to give a number, give a range where your floor is their likely ceiling. Justify with market data, not feelings.
Negotiate the whole package, not just base. Base, bonus, equity, sign-on, level, title, remote flexibility, learning budget, GPU/compute access. When base is capped, sign-on and equity often aren’t.
Get it in writing, negotiate before signing. Post-signature leverage drops to near zero. The window is between verbal offer and signed contract.
Big corp / bank specifics
You’re a contractor at the bank — be honest with yourself: there is no promotion path on a contract. Your preferences doc already says this. So “promotion” at a bank for you means either converting to FTE (rare, usually worse for your goals) or re-negotiating the contract rate at renewal.
For contract rate renewal:
- Time it to renewal, not mid-term.
- Anchor on a competing contract offer or a raised market rate (search current rates for AI engineers in your region).
- Frame around delivered value: “I automated X, shipped Y, the agent tooling I built saves N hours/week.” Quantify in money or time.
- The agency takes a cut — know your markup. Negotiate with whoever controls the rate, which may be the agency, not the bank.
For FTE comp negotiation (if you ever go that route):
- Levels map to bands. Get the band data (levels.fyi, Blind, internal leaks). Negotiate your level first — it caps everything else.
- Refresh grants, vesting cliffs, and bonus targets matter more than base over 4 years.
- Performance review cycles are the only promotion lever. Document impact continuously, pre-write your own promo case, get your manager to sponsor it before the calibration meeting. Promotions are decided in calibration, not in your 1:1.
Startup specifics
Equity is the real negotiation and most people fumble it. What matters:
- Percentage ownership, not share count. 10,000 shares means nothing without total shares outstanding (fully diluted).
- Strike price and 409A valuation — your real cost to exercise and the gap to fair value.
- Vesting schedule — standard is 4yr/1yr cliff. Negotiate acceleration on acquisition (single vs double trigger).
- Preferred vs common — you get common, investors get preferred with liquidation preferences. In a mediocre exit, preferences can wipe out common entirely. Ask about the preference stack.
- Last round valuation and runway — how many months of cash, what’s the burn. This tells you the real risk and the real value of the equity.
At an early startup, your leverage is being uniquely able to build the thing. Someone who consumes 2B tokens/year, builds CLI agents, and trains models on MI300X is not a commodity hire for an AI startup — you’re potentially employee-as-force-multiplier. Price accordingly. If you’re early enough, negotiate for a title and scope that lets you own a domain (e.g., “you own all model training and inference infra”), because scope compounds into future leverage and founder optionality.
Scripts
Big corp, competing offer in hand: “I’m excited about this role. I have a competing offer at [level/$X total comp]. I’d need [target] to make this an easy yes. Can we close that gap?”
Startup equity: “I’m in on the mission. To evaluate the offer properly, can you share fully-diluted shares outstanding, the current 409A, last round valuation, and current runway? On comp, I’m looking for [X base / Y% equity] given I’d be owning [scope].”
No competing offer (weaker, still works): “Based on market data for AI engineers at this level, the range is [X–Y]. The offer is below that. What flexibility is there on base, sign-on, or level?”
For your situation specifically
Given your trajectory — bank contract now, startup later when the mortgage hits 500K — the highest-EV negotiation move isn’t squeezing the bank. It’s building the credible alternative: enough public proof (shipped tools, GitHub traction, a revenue experiment that works) that an AI startup or a much higher contract rate comes to you. That alternative is what gives every future negotiation teeth. Negotiation leverage is mostly built before you ever sit down at the table.
One thing worth flagging: salary/equity negotiation specifics are jurisdiction- and contract-dependent, and I’m not a lawyer or financial advisor — for equity terms especially, have someone read the actual paper before you sign.