Protein Engineering Market

Some 9,180 protein engineering programs were running worldwide in 2025 by Douglas Insights count, set against a structural record that grew by 17,560 new entries released to the Protein Data Bank that year (wwPDB deposition statistics). The protein engineering market covers the instruments, reagents and consumables, protein design software and contract engineering services used to redesign or invent proteins for drugs, enzymes and materials. Douglas Insights estimates the market at USD 3.24 billion in 2025: 9,180 programs multiplied by an average annual spend of USD 352,600 per program. Revenue rises to USD 11.99 billion by 2035, a 13.99% annual rate. Codexis showed where the money is moving on 11 March 2026, when its results filing with the SEC reported fiscal 2025 revenue up 18.7% to USD 70.4 million, led by enzymatic siRNA manufacturing. The study belongs to our pharmaceuticals coverage and follows the Douglas Insights research methodology.

Which protein engineering product segment makes the money, reagents or design software?

Reagents and consumables earn 41.7% of protein engineering revenue, or USD 1.35 billion in 2025, because every design cycle burns expression kits, cloning enzymes, media and purification resins. Protein design software is the smallest product line at 13.3% but grows fastest, at 21.0% a year, as AI models move design upstream of the bench.

Reagents and consumables hold USD 1.35 billion because a single directed evolution campaign screens thousands of variants, and each variant needs DNA, a host cell and an assay well. The line grows 12.2% a year to USD 4.27 billion in 2035. Instruments, at USD 757.4 million or 23.4%, cover liquid handlers, chromatography systems, mass spectrometers and plate readers; replacement cycles of seven to ten years keep their growth to 9.7%, the slowest of the four. Contract engineering services take 21.6% of spend, USD 699.2 million, from drug developers that rent a partner’s library and screening capacity rather than build their own; they compound at 15.4% to USD 2.93 billion. Protein design software starts at USD 430.5 million and reaches USD 2.90 billion by 2035.

Software is the fastest-growing segment at 21.0% because the cost of a computational design round is a fraction of a wet-lab round, and the Nobel Foundation says over two million researchers in 190 countries had used AlphaFold2 by October 2024. Douglas Insights expects software to pass instruments in value before 2035, ending at USD 2.90 billion against USD 1.91 billion.

Why are biologics pipelines and AI design tools pulling protein engineering spend upward?

Program growth of 10.40% a year drives the protein engineering forecast. Biologics pipelines add 4.1 points, AI design adoption 3.6 points, enzymatic manufacturing 1.7 points and academic structural biology 1.0 point.

Biologics pipelines are the largest source of protein engineering demand, worth 4.1 points of the 10.40-point volume leg. Antibodies, bispecifics, fusion proteins and engineered cytokines all pass through affinity maturation, humanisation and developability work before a first human dose. Thermo Fisher Scientific reported USD 10.37 billion of 2025 revenue at its Life Sciences Solutions segment, out of USD 44.56 billion in total, in results released on 29 January 2026; much of that segment sells the expression systems and bioprocess reagents those programs consume. Douglas Insights counts 3,860 of the 9,180 active programs as antibody or other biologic optimisation work.

AI-guided design contributes 3.6 points. The 2024 Nobel Prize in Chemistry went half to David Baker for computational protein design and half to Demis Hassabis and John Jumper of Google DeepMind for protein structure prediction, and the prize committee notes that the team predicted the structure of virtually all 200 million known proteins. A design team that once waited months for a crystal structure now starts from a predicted model in minutes. Douglas Insights puts AI-first design programs at 1,420 in 2025, rising about 21% a year, in line with the software segment.

Enzymatic manufacturing adds 1.7 points. Codexis, which engineers enzymes with its directed evolution platform, recorded USD 44.4 million of research and development revenue and USD 26.0 million of product revenue in 2025, completed a USD 37.8 million technology transfer agreement with Merck and produced a 3-kilogram siRNA batch through its ECO Synthesis route, according to its fourth-quarter release dated 11 March 2026. Each such tech transfer converts a one-off protein engineering project into a recurring enzyme supply line. The company guided 2026 revenue to USD 72 million to USD 76 million.

Academic structural biology supplies the last 1.0 point. Depositions to the Protein Data Bank rose from 19,322 in 2024 to 20,975 in 2025, an 8.6% gain, and releases rose 14.9%, according to the Worldwide Protein Data Bank. Asia led 2025 depositions with 6,989, ahead of North America at 6,265 and Europe at 6,262. Each deposition is a protein engineering input for the next design cycle. The four driver contributions sum to 10.40 points; price per program adds 3.25 points, giving 13.99% revenue growth.

Which bottlenecks in expression screening and wet-lab validation slow protein engineering?

Wet-lab validation is the tightest protein engineering bottleneck, and Douglas Insights reckons three restraints together remove 2.6 points of program growth: validation capacity 1.3 points, biotech funding cycles 0.9 points and biosecurity compliance 0.4 points. Without them the volume leg would run near 13.0%.

Validation capacity removes 1.3 points. A computational round can propose 10,000 protein variants in a day, but expression, purification and assay of even 500 of them takes a typical lab two to four weeks. Design output has outrun bench throughput, so the extra designs queue rather than become new programs. Automated cloud labs help, but they still charge per plate.

Funding cycles remove 0.9 points. Venture-backed protein engineering start-ups depend on rounds that track biotech equity markets, and a weak year cancels programs outright. Absci, an AI drug creation company, states in its fiscal 2025 annual report that it has incurred significant losses since inception and expects further losses, a profile shared by most design-first peers.

Biosecurity compliance removes 0.4 points. Gene synthesis orders now carry sequence-of-concern screening, which adds days and paperwork to some de novo protein engineering projects. The cost is small per order, roughly 1% to 2% of a program budget in our model, but it falls hardest on small academic groups.

How much does an antibody or enzyme protein engineering program cost per year?

USD 352,600 a year is the average protein engineering program spend in 2025, by Douglas Insights estimate. Realised bands run from about USD 90,000 for an academic enzyme project to USD 2.5 million for a pharma antibody campaign. Spend per program rises 3.25% a year to about USD 485,500 in 2035.

The upper end of protein engineering pricing is anchored by drug creation. Absci reports in its 10-K filing that it took its first two programs from AI design to an investigational new drug (IND) filing in around two years at about USD 15 million per program, against an industry average of four to six years and more than USD 50 million. That ratio, about 3.3 to 1, is why pharma buyers pay software and service premiums. Price per program grows only 3.25% a year because cheaper computation offsets rising assay and labour costs.

Which companies supply the protein engineering toolchain, from Codexis to Thermo Fisher?

26.1% of 2025 protein engineering revenue sits with the top three suppliers, Thermo Fisher Scientific, Merck KGaA and Agilent Technologies, by Douglas Insights estimate. Their reagents and instruments are in nearly every lab, while specialists such as Codexis and Absci win on engineering know-how.

Thermo Fisher Scientific leads with an estimated 11.8% share, built on the expression and bioprocess catalogue inside a Life Sciences Solutions segment that Thermo Fisher reported at USD 10.37 billion of 2025 revenue. Merck KGaA follows at 8.9%, built on its Life Science reagents and Process Solutions unit for biologics. Agilent Technologies, at 5.4%, supplies the chromatography and analytical instruments through its Life Sciences and Diagnostics Markets group that confirm an engineered protein is what the design said.

Codexis holds about 2.2%, the ceiling implied by USD 70.4 million of 2025 revenue reported on 11 March 2026 in its SEC filing, and its position rests on enzyme directed evolution now aimed at siRNA manufacturing. Absci is built on the Origin-1 generative model and a lab-in-the-loop validation cycle, per its annual report. Google DeepMind does not sell protein engineering as a product line, but its AlphaFold work shapes the software segment that every other supplier competes in.

Sibling markets explain the rest of the field. Enzyme makers covered in our Industrial Enzymes Market report buy protein engineering capacity to improve thermal stability, and the contract partners in the Biologics Outsourcing Market report increasingly sell engineering upstream of manufacturing.

Where is protein engineering spend concentrated, and which region grows fastest?

North America leads protein engineering with USD 1.43 billion, or 44.3% of 2025 spend, because most biologic pipelines and AI design start-ups sit in Boston and the San Francisco Bay Area. Asia Pacific grows fastest, at 16.6% a year, as Chinese and South Korean biologics developers scale.

North America grows 13.1% a year to USD 4.91 billion by 2035. Europe holds USD 893.4 million, 27.6% of protein engineering revenue, anchored by Swiss, German, Danish and UK pharma and enzyme groups; it compounds at 12.8% to USD 2.98 billion. Asia Pacific starts at USD 725.1 million, 22.4% of the total, and reaches USD 3.37 billion, helped by the region’s lead in Protein Data Bank depositions (6,989 in 2025). Latin America accounts for USD 110.1 million and grows 14.5% to USD 426.4 million, mostly through Brazilian industrial enzyme and agricultural biotech programs. The Middle East and Africa is the wildcard: USD 74.4 million in 2025, rising 15.06% a year to USD 302.5 million as Gulf sovereign funds back genomics and biomanufacturing hubs.

Which biosecurity rules and synthesis screening standards reach protein engineering labs?

The 29 April 2024 Framework for Nucleic Acid Synthesis Screening is the main new rule touching protein engineering, because every designed protein starts as an ordered gene. Federally funded US researchers must buy synthetic DNA only from providers that screen for sequences of concern.

The framework published by the White House Office of Science and Technology Policy (OSTP) directed US research funding agencies to require compliance within 180 days of its release, with a grace period running to April 2025 and annual self-attestation updates due each 1 January. For protein engineering groups the practical cost is a vetted vendor list and a short delay on unusual sequences. Douglas Insights treats the rule as a 0.4-point drag on program growth, as set out above, rather than a ban on any design category.

Quality standards matter at the manufacturing edge. Codexis received ISO 9001 certification for its in-house manufacturing in 2025, according to its results release, a step that lets an engineered enzyme move from protein engineering project to a supplied pharmaceutical input. Engineered therapeutic proteins then follow the normal biologics approval route, which sits outside this market’s revenue.

What does protein engineering cover across rational design, directed evolution and AI-guided de novo design?

Protein engineering spans three methods: rational design at about 38% of 2025 programs, directed evolution at 46% and AI-guided de novo design at 16%, by Douglas Insights count of 9,180 programs. Revenue counted here is tool and service spend, not sales of the finished drugs or enzymes.

Rational design changes chosen residues from a known structure. Directed evolution, the method Codexis has built its business on, mutates and screens libraries over many rounds. AI-guided de novo design writes new sequences from a target shape, the approach recognised by the 2024 Nobel Prize. Broader cell programming sits in our Synthetic Biology Market report; this market stops at the protein.

How much of protein engineering spend comes from biopharmaceutical companies versus academic laboratories?

Biopharmaceutical companies account for about 61% of protein engineering spend in 2025, contract research organisations 17%, industrial and food enzyme producers 12% and academic laboratories 10%, in Douglas Insights end-user split. Academic labs run the most programs but spend the least on each.

Academic laboratories run about 3,300 of the 9,180 programs at an average near USD 98,000 a year. Biopharmaceutical companies run fewer programs at several times the budget, which is why their share of value is six times the academic share. Contract research organisations gain share fastest, at around 15% a year, as smaller biotechs outsource screening.

Which protein design method, directed evolution or AI-guided de novo design, gains program share by 2035?

AI-guided de novo design rises from 16% of protein engineering programs in 2025 to about 34% by 2035 in our base case, taking share mostly from rational design. Directed evolution holds near 41% because screening remains the proof step for any design.

The swing matters for revenue mix. A de novo program spends more on software and less on library consumables, so its rise is the reason the protein design software segment reaches USD 2.90 billion by 2035. Directed evolution keeps reagent volumes high. Hybrid workflows, where a model proposes and evolution polishes, are already the norm at Absci and similar groups.

How far can the protein engineering forecast move by 2035 if AI design adoption stalls or accelerates?

Our base case puts protein engineering at USD 11.99 billion in 2035 from 10.40% program growth and 3.25% price growth. A slower case reaches USD 8.54 billion and a faster case USD 15.97 billion, so the plausible span is wide.

The slower case assumes a two-year biotech funding drought and validation capacity that never catches up, cutting program growth to 7.6%. The faster case assumes more Codexis-style tech transfers like the one reported on 11 March 2026 in its filing, plus wider AI design use, lifting program growth to 12.9%. Add one point to annual program growth and the 2035 figure lands about USD 1.13 billion higher. Published forecasts that we read run from about 15.9% to 25.9% a year; our 13.99% is lower because we count tool and service spend only and model validation limits.

Douglas Exclusive: the Protein Engineering Program Scorecard

The Protein Engineering Program Scorecard is a Douglas Insights model, not an official ranking. It is built from 14 inputs: the 6 primary sources linked on this page, our count of 9,180 programs by 5 program types, and 3 spend bands from the pricing table. Each program type is scored 1 to 10 on 5 weighted criteria.

Antibody and biologic optimisation scores highest at 71.0 out of 100 because protein engineering programs there are large and well funded. Industrial enzymes and biocatalysts rank second at 67.5 on fast validation and clear regulation. RNA and gene editing enzymes follow at 66.5. AI-designed de novo binders score 65.5: top marks for growth, lowest for data availability. Vaccine antigens trail at 53.0. The finding: de novo binders sit only 5.5 points behind the leader, so a better validation dataset would move them to the top row.

How we built the protein engineering model from program counts and spend per program?

9,180 programs multiplied by USD 352,600 gives USD 3.24 billion for 2025, the protein engineering base. The count covers 31 countries and 5 regions, built from 47 data points including Protein Data Bank deposition totals and company filings.

Programs grow 10.40% a year to about 24,691 in 2035; price per program grows 3.25% to about USD 485,500. Multiplied, 24,691 programs times USD 485,500 equals USD 11.99 billion. Cross-check one: Codexis revenue of USD 70.4 million equals 2.2% of our 2025 base, in line with a specialist supplier. Cross-check two: our 2025 base sits within 6% of the midpoint of the published estimates we read, once their earlier base years are rolled forward.

Sources

1. U.S. SEC Codexis fourth quarter and fiscal year 2025 results (8-K exhibit 99.1) (2026)
2. U.S. SEC Thermo Fisher Scientific fourth quarter and full year 2025 results (2026)
3. U.S. SEC Absci Form 10-K fiscal 2025 (2026)
4. Nobel Prize Outreach Nobel Prize in Chemistry 2024, popular information (2024)
5. White House OSTP Framework for Nucleic Acid Synthesis Screening (2024)
6. Worldwide Protein Data Bank PDB deposition statistics (2026)