Boards still plan product investment off last year's 35% R&D-of-revenue bar. That was the CY-24 median in the same Aleph × Benchmarkit cut. CY-2025 actuals already moved.
Across 342 B2B SaaS and AI-native companies reporting CY-2025 actuals, the R&D Expenses as % of Revenue cut (N=192) prints a median of 27%. That is an 8-point drop from 35% in CY-24. The 75th percentile sits at 40%. The lean floor (bottom quartile) sits at 10%.
This B2Bcentr report shows which sample produced the 27% median, how R&D differs from full S&M spend, what growth cohorts and ARR scale do to the number, and why a 35% board slide now mis-sets the product plan.
Key Takeaways
- Median R&D is 27% of revenue (N=192). Boards still quote ~35%.
- YoY: 35% to 27%. An 8-point drop in one year.
- 75th percentile prints 40%; bottom quartile / lean floor is 10%.
- High-growth (>50%) cohort spends 33%; slow-growth (<10%) spends 26%.
- Distinct from S&M: full S&M median is 35% of revenue (N=196).
Which sample produced the 27%
Primary source: Benchmarkit 2026 SaaS & AI-Native Metrics (research partnership with Aleph). Overall file: 342 B2B SaaS and AI-native software companies, CY-2025 actuals. The R&D Expenses as % of Revenue cut is N=192. Context write-up: Aleph's 2026 SaaS benchmarks takeaways. Guide page: Aleph 2026 SaaS benchmarks.
Definition used here: Research and Development expenses as a percentage of revenue (product and engineering on the P&L, not Sales & Marketing). At 27%, the median private company spends a little over a quarter of revenue on R&D. That is not S&M %, not Gross Margin, and not ARR per employee.
| Cut | Number | Sample |
|---|---|---|
| Median R&D % of revenue | 27% | Aleph × Benchmarkit 2026; N=192; CY-2025 |
| CY-24 median (same trend cut) | 35% | YoY: 8-point decline into CY-25 |
| 75th percentile (CY-25) | 40% | Earlier-stage or heavy AI product investment |
| Bottom quartile / lean floor | 10% | Product maturity or AI coding leverage |
| High-growth (>50%) cohort | 33% median R&D | Highest growth-band R&D in the cut |
| Slow-growth (<10%) cohort | 26% median R&D | Underinvestment risk cited near 16% |
| Scale: sub-$5M → >$100M ARR | ~32% → ~22% | ARR-band cut N=96; leverage at $20M–$50M |
| Top quartile R&D efficiency (exec summary) | ~22% | Cited as AI-assisted engineering productivity |
| Overall file size | 342 companies | B2B SaaS and AI-native; CY-2025 actuals |
| S&M (separate post) | 35% of revenue | N=196; distinct from product R&D |
Why the ~35% folklore still shows up in board decks
Operator folklore still treats last year's 35% R&D median as the planning default. Older packs round that into a clean third-of-revenue product target and keep using it as a floor even after CY-25 compressed.
Benchmarkit's own key insight on this cut is direct: median R&D at 27% of revenue is consistent with meaningful product investment in B2B SaaS, and it reflects a balance between product commitment and revenue efficiency across the benchmark population. Total-population benchmarks still need care: sample composition is material, and R&D % correlates with growth rate and with company size.
Here's why that matters. If your plan still assumes a 35% floor while the private median is 27%, you will overfund the product line relative to peers or under-credit the efficiency the market already booked in CY-25. Read this next to B2Bcentr's S&M spend report (35% of revenue on sales and marketing) so you do not confuse a growth-function line with product R&D.
R&D is not S&M, and it is not Gross Margin
Full S&M answers how much of revenue the combined sales and marketing function consumes. In B2Bcentr's S&M spend report, Aleph × Benchmarkit's median is 35% of revenue on S&M (N=196). R&D answers how much of revenue product and engineering consume. Different numerator, different planning use.
Gross Margin answers what share of revenue remains after cost of goods. In B2Bcentr's Gross Margin report, the same Aleph file prints an 80% median. You can print a healthy 27% R&D line while still clearing an 80% gross margin. Do not swap the two metrics in one sentence.
ARR per employee and Magic Number are siblings on the efficiency side. ARR per employee tracks how much revenue each head produces (see B2Bcentr's ARR per employee report). Magic Number tracks how efficiently last period's S&M converts into this period's ARR growth (see the Magic Number report). Use R&D % of revenue for the product P&L shape. Use S&M % for the growth function. Use Gross Margin for contribution after COGS.
Growth cohorts, ARR scale, and the efficiency story
By 2025 growth rate, the >50% growth cohort carries the highest median R&D at 33%. That matches the product thesis: companies that treat product differentiation as the primary growth driver allocate capital accordingly.
The slow-growth band is the risk tell. Companies growing under 10% print a 26% median R&D. Benchmarkit flags companies at just 16% R&D as at elevated risk of competitive disadvantage through underinvestment, vendor consolidation, or AI-native upstarts that prioritize product above all else.
By ARR scale (N=96 in the ARR-band cut), R&D as a percentage of revenue declines from about 32% at sub-$5M to about 22% above $100M. The sharpest efficiency transition sits in the $20M–$50M band (about 22% vs about 32% at $5M–$20M), which Benchmarkit reads as the R&D leverage inflection point alongside GTM efficiency.
YoY, median R&D fell from 35% in CY-24 to 27% in CY-25. Aleph's exec summary pairs that with top-quartile R&D efficiency near 22%, described as achievable through AI-assisted engineering productivity. The 75th percentile at 40% still marks the heavy-investment end of the distribution.
Read it next to Rule of 40, ARR growth, and S&M spend
Median Rule of 40 is 25%, not the investor 40% folklore bar (Rule of 40 report). A large part of the CY-25 Rule of 40 jump came from cutting investment lines, including S&M and R&D, not only from stronger durability. Aleph's takeaways are blunt on that point: efficiency improved because companies reduced S&M and R&D investments, while median growth fell and GRR weakened.
Median YoY ARR growth is 20%, not the old 40% folklore bar (ARR growth report). A 27% R&D line at 20% growth is a different operating reality than a 35% R&D line at 40% growth. Pair the spend % with the growth bar before you celebrate "efficiency."
The median company freed roughly 8% of revenue from R&D in a single year (35% to 27%), plus cuts from S&M and G&A, while losing growth points. That is a one-time efficiency dividend if you bank it. It is a reinvestment budget if you put it into durable product advantage and AI-specific engineering.
On AI specifically, the same file prints a small directional cut for AI Expenses as % of R&D (N=46): median about 10% of R&D allocated specifically to AI capabilities, with top quartile at 30%+. Treat that as early-stage and directional, not as a board floor.
B2Bcentr's take
Who this does not work for
Pre-PMF teams with almost no revenue base cannot manage to a 27% R&D % the same way a scaled ARR book can. Early teams should still prioritize product discovery and technical differentiation, then graduate to R&D % of revenue once the P&L is meaningful.
This number also breaks if you redefine R&D to exclude contractor engineering, platform teams, or AI infra cost, or if you paste a full S&M 35% into a product R&D slide. Stick to R&D Expenses as % of revenue the way Benchmarkit reports it. Match growth cohort and ARR band before you copy a peer median.
What to do Monday
- Pull trailing-year R&D Expenses as % of revenue and keep S&M on a separate line.
- Put 27% median, 40% 75th percentile, and 10% lean floor on one board slide next to S&M 35%.
- Split R&D % by growth cohort and ARR band. Do not manage a slow-growth book to a >50% cohort's 33%.
- If R&D sits near 16% with sub-10% growth, audit product roadmap and AI-native competitive risk before cutting again.
- Re-check Rule of 40 (25%), 20% ARR growth, Gross Margin (80%), and ARR per employee on the same pack.
Next number for the same efficiency pack: how much of revenue full Sales & Marketing consumes when boards still quote 40–50%.
FAQ
What is a good SaaS R&D spend as a percent of revenue in 2026?
Aleph × Benchmarkit's 2026 file prints a 27% median R&D Expenses as % of Revenue (N=192). The 75th percentile sits at 40%. The bottom quartile / lean floor sits at 10%. Treat last year's 35% as folklore relative to CY-2025 actuals, not as the current private median.
Is 35% of revenue on R&D still a realistic planning assumption?
Treat 35% as last year's median (CY-24), not as the 2025 private median. The same cut shows the median at 27% and an 8-point YoY drop. Heavy-investment books still show up near the 40% 75th percentile.
How is R&D different from S&M spend as a percent of revenue?
R&D is product and engineering on the P&L as a percent of revenue (27% median here). Full S&M is sales plus marketing; Aleph × Benchmarkit's median is 35% of revenue on S&M (N=196). Do not paste one number into the other slide.
Does growth rate change R&D spend as a percent of revenue?
Yes in this sample. Companies growing above 50% print a 33% median R&D. Slow-growth companies under 10% print 26%. Benchmarkit flags books near 16% R&D as at elevated underinvestment risk.
How does R&D % of revenue change with company size?
In the ARR-band cut (N=96), R&D declines from about 32% at sub-$5M ARR to about 22% above $100M. The sharpest leverage transition sits around $20M–$50M.
Did R&D spend fall year over year?
Yes in this cut. Median R&D fell from 35% in CY-24 to 27% in CY-25, an 8-point drop in one year. Aleph's exec summary also cites top-quartile R&D efficiency near 22% via AI-assisted engineering productivity.
How much of R&D is specifically AI spend?
A small directional cut (N=46) prints a median of about 10% of R&D allocated specifically to AI capabilities, with top quartile at 30%+. Treat it as early-stage, not as a board floor.
Where can I read the full Aleph × Benchmarkit sample?
Start with the Benchmarkit 2026 report hub and Aleph's takeaways post. Guide page: 2026 SaaS benchmarks.
Primary source: Benchmarkit 2026 SaaS & AI-Native Metrics (Aleph partnership; 342 companies; R&D Expenses as % of Revenue cut N=192; CY-2025). Context: Aleph takeaways and Aleph guide page.