How to Negotiate a Raise as an Engineer in 2026 (and What Actually Works)

If you're an engineer reading this in 2026, you probably already suspect you're underpaid. The suspicion is usually right. Internal merit raises have shrunk to a median of roughly 3.5 percent for U.S. tech workers this year, and some surveys put actual base gains closer to the 1.5–2 percent range once you strip out the specialized AI roles. Meanwhile, external offers for the same level routinely land 10 to 30 percent higher. The gap between "loyal internal employee" and "same person with a new offer" has never been wider.

The playbook that worked in 2021 — walk into your one-on-one, say you've been thinking about your career, get a 15 percent bump on the spot — is dead. Managers in 2026 have layoff overhang, tighter budgets, and a rigid calibration process that quietly decides your raise months before you ever ask. But raises are still very much possible. They just require a different approach: written, timed, and grounded in evidence.

This guide is written for the mid-to-senior engineer who feels underpaid, feels awkward about asking, and doesn't want to burn a bridge. Skip the fluff, use the templates, and time it right.

The 30-Second Version

If you only read one section, read this. Here is what actually works in 2026:

1. Write a calibration memo — a one- to two-page document that quantifies your impact, benchmarks your comp against market data, and asks for a specific number. Deliver it 30 to 60 days before the calibration meeting your manager attends. This is the single most effective lever inside a big company.
2. Get a competing offer — if you'd genuinely take it. Roughly 60 percent of senior engineering offers now trigger a counter, and the counter wins about half the time. But bluffing is transparent and hurts you long-term. Only pursue this route if the outside offer is one you'd actually accept.
3. Time it to the calibration cycle. Ask off-cycle and you're asking your manager to fight for something the budget doesn't have. Ask three months before calibration and you're a case they can build.

Everything else in this article is a longer version of those three ideas, plus the scripts and email templates you can copy verbatim. If you want to see what target numbers look like at the top of the market, our engineering compensation bands guide has current ranges by level.

Why Raises Are Harder in 2026 Than 2021

Three structural things changed, and understanding them changes how you negotiate.

Budgets are flat. Merit-increase budgets across U.S. tech are running around 3.5 percent this year, down from 4 percent in 2025 and dramatically down from the 5–7 percent bumps that were common in 2021–2022. Some Robert Half survey data puts tech-specific base gains as low as 1.6 percent for 2026. The pool your manager is fighting over is smaller than it's been in five years.

Calibration is real. At any company larger than roughly 300 engineers, raises and promotions are decided in calibration meetings where managers gather, present cases, and negotiate a fixed budget across people. Your manager does not unilaterally set your comp. If you don't give them ammunition weeks before that meeting, you are not on the promotion list. Full stop.

Layoff overhang has weakened counter-offers. Managers who are worried about their own headcount are less willing to fight to retain you, because losing you might quietly help them make their next reorg number. Counter-offers still work at healthy companies, but they don't work everywhere anymore — and if you present one at a company that's about to reorg, you might just be volunteering for the list.

Against that backdrop, the engineers who are still getting real raises share three traits: they wrote something down, they timed it right, and they had external data to anchor to. Let's walk through each play.

Play 1: The Calibration-Cycle Memo

This is the most effective play at any company with a formal calibration process — which is nearly every company profiled in our Culture Directory, from Stripe and Databricks to Cloudflare and HubSpot. It works because you're not asking your manager to give you a raise. You're giving your manager the tools to argue for you against their peers.

Timing: Deliver the memo 30 to 60 days before calibration. If your company runs H1 and H2 cycles, the H2 calibration typically happens in October or November. Deliver in August or September. If you deliver during calibration week, you are too late. The decisions are already being formalized.

Structure: One to two pages, in whatever tool your team uses (Notion, Google Doc, a memo attached to Slack). Send it before a scheduled one-on-one, then walk through it live. Keep four sections:

1. Impact. Two to four projects from the past 6–12 months, each written as: what you did, the scope, the outcome (quantified where possible), and who benefited. This is not a task list. It's a scoped-impact summary. "Led the migration of the billing service to a new event pipeline, cutting P99 latency from 480ms to 90ms and eliminating a full class of dropped-event incidents" is worth more than a full paragraph of activity.
2. Scope creep. If your responsibilities have grown since your last review, name it explicitly. New on-call rotation? Mentoring three juniors? Owning a new subsystem? These are level-up signals that calibrators love, but only if you say them out loud.
3. Market benchmark. A single sentence: "Employee-reported compensation data for L5/senior engineers at companies of comparable size and stage centers around $X–$Y total comp. My current total comp is $Z, placing me roughly N percent below mid-band." Do not cite specific external platforms by name; frame the number as an industry benchmark. If you need to build the benchmark, our culture and compensation comparison tool and the AI engineer salary guide by level are good starting points.
4. The specific ask. Not "a raise." A specific base target, a specific equity refresh target if applicable, or a specific level (Senior 2, Staff, TL) if that's what you're pursuing. Vague asks get vague answers.

Here's a template you can adapt verbatim:

This memo does three things at once: it writes the case for your manager (which they now don't have to write themselves), it locks in a specific number so they can't fuzz the ask, and it demonstrates the kind of clear written communication that senior levels reward. It is a rare memo that does not at least start a serious conversation.

Play 2: The Competing-Offer Play (Handle With Care)

Competing offers still work in 2026, but they have become the highest-risk lever in the toolkit. The data is that roughly 60 percent of senior engineering offers now trigger a counter, and the counter wins about half the time. That's the good news. The bad news: the manager who counters you often quietly resets how they think about you for the next promotion cycle.

Rules for playing this hand well:

1. Only interview if you'd actually take the offer. If you can't picture yourself walking away, don't start. Managers can tell when you're bluffing, and the market is small.
2. Get the offer in writing first. A verbal "we'd probably offer around X" is not a competing offer. It's a rumor. You want a signed letter with a start date before you tell your current manager anything.
3. Bring it up privately, once. Not in a big meeting. Not over Slack. Book a 30-minute one-on-one and say the sentence below. Then stop talking.
4. Give a real window. "I need to make a decision by [date] — two weeks feels reasonable to me." Anything vaguer and you're not creating a decision, you're creating hope.

Here's the script that works:

Note what this script does not do. It doesn't name the offer as a threat. It doesn't demand a specific counter. It doesn't compare you to peers or make it personal. It states a fact, states a preference, and creates a decision window. That's the whole trick.

One more warning. If your company is a smaller startup or has been shedding headcount, be honest with yourself about whether the manager wants to fight for you. At larger, healthy engineering orgs like Cloudflare, Notion, or Anthropic, retention still matters. At companies quietly working through a headcount reduction, presenting an offer sometimes just resolves a decision the manager was already trying to make.

Play 3: The Retention Conversation (Senior/Staff+)

At Staff engineer and above, the dynamic changes. Companies expect senior engineers to be recruited constantly, and skilled managers know that if they don't have regular comp conversations with their staff engineers, they lose them.

The retention conversation is a lower-risk version of the competing-offer play. You are not presenting an offer. You are saying, in effect: "I'm getting inbound. I want to stay. Let's talk about what a competitive package looks like." Skilled managers will take this seriously, because they know a Staff-level backfill takes 6–9 months and costs more than the raise you're asking for.

Try something like:

This works because it invites your manager to solve the problem before it becomes a crisis. It also signals maturity: you're not weaponizing an offer, you're inviting a partnership. For senior engineers reading our engineering career ladders guide and thinking about their next level, this conversation is often the one that unblocks a Staff promotion sitting in "we're thinking about it" limbo.

Play 4: The Scope-Expansion Pivot

Sometimes a direct raise is off the table — frozen budget, recent round of layoffs, comp bands you're already at the top of. That does not mean the conversation is over. It means you pivot to non-cash levers that create long-term compounding value:

• Equity refresh. If you're at a pre-IPO company or on a 4-year new-hire RSU grant that's more than two years in, your effective total comp is already dropping as vesting winds down. Refresh grants are typically 25–30 percent of a new-hire grant at your level. This is the single most under-negotiated line item at big tech.
• Off-cycle title / level change. Promotion into a higher level often carries a mandatory raise into the new band. A Senior-to-Staff bump can be worth 15–25 percent of base by itself.
• Tech Lead or Manager transition. If you'd genuinely enjoy it, the TL/EM transition is another way to unlock a new comp band without a "raise" line item on the budget.
• Written promotion commitment. "We can't move base this cycle, but I want a written plan with the specific work needed for Staff by the next cycle, and a check-in in 90 days." A written commitment with milestones is worth taking. A vague "we'll see" is not.

The equity refresh in particular is worth a hard look. Refresh grants are usually smaller than initial hire grants — often a third to a fifth of the size — but they extend your total comp curve past year two, which is exactly when it would otherwise dip. If a raise isn't on the table, an early or larger-than-standard refresh is a very reasonable ask. Managers can often approve this out of a separate equity pool without touching the salary budget.

What Never Works

A few things that feel like arguments but are not:

• Long tenure alone. "I've been here five years" is a scope-of-loyalty argument, not a scope-of-impact argument. Nobody's raise budget is set on tenure. Tenure without a level jump is actually a red flag: it suggests you weren't promotable during that entire window.
• Personal reasons. New mortgage, new kid, cost-of-living increase. These are real problems, but they aren't your company's problems. Comp is set by market and impact, not need.
• Comparing to peers by name. "Alex makes more than me and I have more scope." Never. Ever. This tanks your relationship with your manager immediately, torches Alex's privacy, and reveals that you have inappropriate salary info. Anchor to market bands, not to specific peers.
• Nuclear ultimatums. "Give me X or I quit" is the fastest way to lose a job you were happy in. Give a decision window and options; don't hand your manager a suicide button.
• Asking during a bad quarter. If your company just missed a target, laid off 10 percent, or lost a key customer, wait. Not forever. Just past the crisis.

What to Anchor To (Without Fabricating Numbers)

You need external data to run any of these plays. But the goal is directional accuracy, not false precision. Total comp for senior engineers at large, well-funded tech companies in 2026 broadly lands in the $250K–$450K range for L5/senior, and stretches to $850K–$1.34M+ for L6/staff at frontier AI labs like OpenAI, Anthropic, and DeepMind. Fintech scale-ups like Stripe, Ramp, and Plaid anchor slightly below the frontier labs but well above the market median. Public cloud infra companies like Cloudflare sit lower on base but competitive on RSUs.

The important thing is to know your specific band. Our engineering compensation bands guide has ranges by level, and the AI engineer salary guide by level covers the AI-adjacent premium. Frame your ask in terms of a band, not a headline number. "Mid-band for my level at similar-stage companies" is far more defensible than "I read on the internet that senior engineers make $400K."

Responding to the Answers You'll Actually Get

Three responses account for probably 90 percent of what you'll hear. Rehearse the responses.

"We don't have the budget."

This is almost never a hard "no." It's a signal that your case wasn't compelling enough to reallocate someone else's budget. Pivot:

"Let me think about it and get back to you."

Fine, but pin down the timeline before you leave the meeting:

A low counter you don't want to accept.

Don't accept immediately, don't reject harshly. Buy 48 hours:

Then in the 48-hour window: decide whether the number plus the non-cash pieces (equity refresh, promotion commitment, scope changes) genuinely closes the gap. If it does, accept graciously in writing. If it doesn't, come back with a specific ask and a concrete alternative — and be prepared to actually walk if that's what you said you'd do.

The Startup vs Big-Tech Distinction

The plays above apply broadly, but the emphasis shifts by company stage.

At big tech (companies with formal calibration, comp bands, and finance-owned raise budgets), Play 1 — the calibration memo, timed correctly — is your single biggest lever. Managers have less discretion than you think, but they have real influence in the room. Give them the ammunition.

At scale-ups (post-Series C, still-growing, but formal HR processes), the calibration memo still works, and the retention conversation for Staff+ engineers is often the most effective play. Companies like Cursor, Perplexity, and Linear live in this zone.

At early-stage startups (Series A/B), formal calibration barely exists. Comp decisions are more discretionary, budgets are tighter, but total comp is more negotiable — especially equity. At this stage, an additional equity grant is often more valuable to you than a base bump, because the upside dwarfs the base. Ask founders and hiring managers about their equity refresh philosophy explicitly; many haven't thought about it and will give you a better answer than you expect.

For an honest look at how different companies handle career growth, culture, and compensation, browse our culture directory or use the side-by-side comparison tool to see how any two companies stack up on career opportunities, work-life balance, and comp.

Timing Playbook: When to Ask

• 30 to 60 days before calibration: The sweet spot. Deliver the memo, book the conversation, give your manager runway.
• 30 days after a big win: A launched product, a saved incident, a landed hire — these are natural triggers for an off-cycle conversation. Just don't drop it the same week; give the impact time to be recognized.
• Right after a promotion: Underused. The moment you're offered a promotion, the base is often the minimum of the new band. Push for mid-band. You will not seem ungrateful.
• Never during: A layoff round, a bad earnings quarter, the week your manager returns from a stressful leadership offsite, or the day of a re-org announcement. Read the room.

Browse Roles at Companies That Take Comp Seriously

If you run through this whole process and the honest conclusion is that your current company simply won't pay you what your level is worth, the market answer is straightforward: leave. Every senior engineer we know who made a real jump in the last 24 months did it by changing companies once and then negotiating hard at the new place.

If that's where you are, we track hundreds of open engineering roles from companies that actually pay competitively — frontier AI labs, developer-tools companies, fintech scale-ups. Every listing is filtered by culture signal so you can filter for the traits that matter to you.