London salary data
What a product engineer earns in London
Not what the job ads say. These are the prices at which offers were actually accepted or lost, across startups from seed to Series B, between mid-2025 and mid-2026.
The short answer
If you want one line: a strong product engineer with two to four years of experience clears £95,000 to £120,000 base in London, plus equity. The full picture, by level:
| Level | Clears at | Outliers to | For context |
|---|---|---|---|
| Product engineer 2 to 4 years |
£95k to £120k | £140k | Payroll median for the same profile is about £72k. Advertised London median reached £100k this summer. |
| Senior 5 to 10+ years |
£120k to £155k | £175k | Payroll median at Series A or B is £112k, but that describes everyone who is not moving. |
| Staff | £135k to £170k | £200k | Published London staff guides say £110k to £165k. AI-native companies clear the guides routinely. |
| Any of the above with production LLM work |
+12% | and rising | Payroll-verified like-for-like premium. Advertised measures run higher, and the buyer pool changes entirely. |
One caveat that matters. £120k at four years is not the median. It is the top decile, concentrated in AI-native and heavily funded companies. But that is precisely the segment you are bidding against for the strongest people.
Why the benchmark data misleads you
Payroll data says the median mid-level engineer at a UK startup earns about £72,000. A founder looking at that number reasonably concludes that £90,000 is a strong offer.
Meanwhile the funded London startups competing for the same shortlist advertise £85,000 to £140,000 for the same profile. The top of the London startup market sits around £155,000. US AI companies hiring in London have stopped pretending there is a ceiling at all.
The candidates you actually want, the ones who pass your bar, are being bid at the 90th percentile. Benchmark at the median and you are not in the auction. Payroll medians describe people who already have jobs. They always did.
What that looks like on a real offer
This summer a Series A AI company in London, top-tier fund on the cap table, founders with the kind of track record engineers seek out, made an offer to an engineer with three years of commercial experience. He had gone into the process expecting £80,000.
The offer gave him a choice: £135,000 base with a standard option grant, or £115,000 base with a significantly larger one.
He took the equity. Which tells you the second thing about this market. The companies winning are not just paying more, they are structuring better.
There is a quiet tax logic here too. Between £100,000 and £125,000 the UK effective marginal rate hits 62 per cent, because the personal allowance tapers away. So £10,000 of base in that band is about £3,800 in your account. Sophisticated candidates increasingly push the difference into options instead. If a candidate ever asks to lower their base, this is why.
Where deals actually die
Twelve months of offer-stage outcomes, one pattern.
Below about £100k, deals die on money
Full stop. In one month we watched a £20,000 gap decide two deals at the same company. They lost a candidate at £75,000 who was holding a competing offer near £100,000, then closed a near-identical profile weeks later at £95,000. Same company, same role. Twenty thousand pounds between losing and hiring.
Above about £100k, money stops deciding and other things start
- The counter arrives the same week. A £120,000 engineer's employer matched a rival offer at his very next review and attached a promotion path. Deal dead. Assume the counter is coming within the week. I always took "half the people who accept a counter-offer are back on the market within months" for classic recruiter-speak. There is actually data suggesting it is true.
- At parity, benefits decide. Two offers a senior engineer described as equivalent on money were decided entirely by parental leave. Thirteen weeks at full pay versus statutory. The startup with the better mission lost.
- Even 20 per cent over ask does not close by itself. A Series B company offered about £120,000 to an engineer with three years of experience who had been looking for £100,000. He declined, for a rival offer with more ownership. Money gets you into the decision. Scope wins it.
The honest summary. Paying market gets you into the room. Under-paying takes you out of it before you knew you were in.
Equity: what normal looks like
The other half of the offer, and where most founders and candidates are flying blind.
| Stage | What normal looks like |
|---|---|
| Seed | First engineer about 1.5 per cent fully diluted, a figure that has not budged in years. By the fifth hire, roughly 0.33 per cent. |
| Series A | Individual grants sized from an option pool of about 12 per cent of the company. Pools of 10 to 15 per cent are typical in Europe. |
| Series B | Median new-hire grant worth about 8.6 per cent of base salary in annualised value. |
| Refreshers | A typical refresh is about 30 per cent of a new-hire grant. Nearly half of employees hold a second grant by year two. |
| Exercise windows | 90 days after leaving is still the default. Extended windows are spreading, but you have to ask. |
Two currents worth knowing. AI roles carry roughly a 12 per cent median pay premium in Europe and their initial equity grants grew 31 per cent in two years, about three times the general rate. If you are an AI-native company, your bands age in months rather than years.
And from April 2026 the UK dramatically expanded EMI, the tax-advantaged option scheme most startup engineers are granted under. If a company told you in 2025 that it was too big for EMI, that is mostly no longer true. I have written about what the EMI change actually did on the market page.
If you are hiring
First, be honest about what you are selling. You might believe your startup is the best in the world. You founded it, you should. But a candidate weighing offers sees what everyone outside sees: the fund on the cap table, the founders' track record, the brand, the traction. If the company across the table has more of those than you, no negotiator alive closes that gap on conviction alone.
- Benchmark against who you are actually competing with. Funded startups' advertised ranges, not payroll medians.
- Advertise the real band. Your cleared price is £20,000 to £30,000 above your JD, and the gap costs you the first conversation with exactly the people you want.
- Budget for the counter before you offer. It is highly likely coming within the week.
- If you cannot win on base, win on structure. Base and equity variants, long exercise windows, and the benefits that genuinely decide deals. Parental leave decided one of the offers above.
If you are a candidate
- Know the real bands, the ones in the table above rather than the ones on the job description. If an offer lands £30,000 under the market for your profile, that is information. It does not have to be an insult, but it is information.
- The best moves I saw all year were not "take the biggest number". They were choosing the larger option grant, trading base above £100,000 into equity, weighting exercise windows, and taking the role with real scope.
- Ask the six equity questions before you sign: percentage or share count and fully diluted total, strike price, latest valuation, pool size, refresh policy, exercise window. And whether the scheme is EMI.
The number is where the conversation starts, not where it ends.
Common questions
What is a good salary for a product engineer in London?
At two to four years of commercial experience, £95,000 to £120,000 base plus equity, with outliers to £140,000. Senior clears £120,000 to £155,000. Staff clears £135,000 to £170,000. Below about £100,000 you will lose candidates who are holding a second offer.
Why is the advertised range always lower than the real number?
Advertised bands are written to protect internal parity, then routinely broken for the strongest candidates. Across our client base the gap runs £20,000 to £30,000. Everyone's band is fiction at the top end, candidates know it, and they price accordingly.
Does AI or LLM experience raise the number?
Yes, about 12 per cent like-for-like on verified payroll data, and advertised measures run higher. But the premium follows proof rather than titles. Hiring managers report a wave of CVs rebadged "Senior AI Engineer" and filter them out in one interview. What moves you across the line is production LLM work, shipped and measured, in front of users.
How much equity should I be getting?
At seed, the first engineer takes a median of about 1.5 per cent fully diluted, falling to roughly 0.33 per cent by the fifth hire. At Series A, grants come from a pool of about 12 per cent of the company. At Series B the median new-hire grant is worth about 8.6 per cent of base in annualised value. Ask what the refresh policy is before you sign, not after.
Why would anyone negotiate their base salary down?
Between £100,000 and £125,000 the effective marginal rate is 62 per cent, because the personal allowance tapers away. £10,000 of base in that band is about £3,800 in your account. The same value in options, pension or an extended exercise window is worth materially more.
Where does this data come from?
Two places. First, offers and placements Ethiq worked on between mid-2025 and mid-2026: real offers from seed to Series B startups in London, anonymised, including the ones we lost. Second, published benchmark data from Ravio and Figures for payroll, cord and ITJobsWatch for advertised ranges, Wellfound and levels.fyi for candidate-reported figures, and Carta and Index Ventures for equity.
Where our data and the published data disagree, I have said so. The full report, including a PDF version, lives on Ethiq's site.
Read next
Nothing here is a job ad. Questions about the data, or an offer you are weighing up: james@ethiqrec.com.