From Pills to Living Cells: Why Pharma Builds So Many Kinds of Drugs, and How Complex Each One Really Is

Prepared by Richstorm.co

Key Takeaways

• Pharmaceutical companies build many different types of drugs, called modalities, because no single approach can reach every kind of disease target inside the body.

• The modalities in use today span an enormous range of manufacturing difficulty, from small molecule pills to living, patient-specific cell therapies.

• As a rule, complexity tracks with how many separate technologies have to be combined into one product, not with how new or advanced the science sounds.

• Complexity generally increases as a modality moves from a single chemical or biologic entity toward a hybrid of multiple manufacturing disciplines, or toward a living, patient-specific product.

Why Pharma Builds So Many Different Kinds of Drugs

It would be simpler for everyone if one type of drug could treat every disease. It cannot, because diseases are not one problem. Some are caused by a single faulty protein that a small molecule can block. Others require a level of precision that only an engineered antibody can deliver. Some require an entirely missing gene to be replaced, which no chemical compound can do at all. The modality has to match the biology of the problem, not the other way around.

A useful way to think about it: a small molecule drug is like a key cut for a specific lock. It is cheap and fast to make, and it can slip inside a cell to reach targets that are otherwise hidden. But it cannot address a missing or broken gene, and it usually cannot distinguish between a diseased cell and a healthy one nearly as well as a more targeted therapy can.

An antibody is more like a trained scout. It can recognize one specific marker with high precision and stay in the body for weeks, but it is generally too large to get inside most cells and cannot fix an underlying genetic problem either. Cell and gene therapies exist because some diseases cannot be solved by blocking or targeting anything from the outside. They require either replacing a broken gene permanently or introducing an entirely new, living, reprogrammed cell into the body.

In practice, most large pharmaceutical companies now run multiple modalities side by side, choosing whichever one best fits the biology of a given disease, rather than betting everything on one type of drug.

Ranking Every Modality by How Hard It Is to Build

Complexity in drug development generally comes down to one thing: how many distinct manufacturing technologies have to be combined into a single working product, and how much of that process happens inside a living system rather than a controlled chemical reaction.

Worth noting upfront: this specific ordering isn't the only reasonable one. Cell and gene therapy rank above antibody-drug conjugates here because they involve irreversible, one-shot manufacturing with far less room for error, but a ranking weighted toward how many entirely separate disciplines a single company needs in-house would put ADCs higher instead, since they force together biologics fermentation and industrial toxic-chemical synthesis, two capabilities almost no company has by default. Both orderings are defensible; this one just makes the criterion explicit rather than leaving it implied.

Ranked roughly from simplest to hardest to build:

1. Small Molecules

These are chemically synthesized compounds, produced through well-established chemistry that the industry has refined for more than a century. They are the least complex modality to manufacture, which is part of why they remain the largest share of the pharmaceutical market by volume. They treat an enormous range of conditions, from infections to high blood pressure to many cancers, and their biggest advantage is that they can cross into places biologics cannot reach, including the brain.

2. Monoclonal Antibodies (mAbs)

A monoclonal antibody is a single, lab-grown protein engineered to recognize one specific target. It is a biologic, meaning it is grown inside living cells rather than synthesized chemically, which is already a step up in manufacturing complexity from a small molecule. Antibodies treat autoimmune diseases, cancers, and a growing list of chronic conditions, largely by blocking or flagging a specific protein for the immune system to act on.

3. Fusion Proteins

A fusion protein combines two different protein components into a single molecule, engineered as one continuous piece. It sits at roughly the same complexity level as a standard antibody, since it is still a single biologic entity produced through one manufacturing process, just with a more complex underlying design.

4. Bispecific and Multispecific Antibodies

These are antibodies engineered to recognize two or more different targets at once, instead of just one. The manufacturing process is still a single biologic production line, but designing an antibody that can bind two separate targets correctly, without the two halves interfering with each other, is a much harder engineering problem than a standard antibody. This is the design complexity we discussed in an earlier article on bispecific antibody-drug conjugates.

5. RNA and mRNA Therapeutics

An mRNA therapy works differently from anything above it. Instead of delivering a finished protein, it delivers genetic instructions that tell the patient's own cells to produce a specific protein themselves, whether that is a piece of a virus for a vaccine or a therapeutic protein the body is missing. The molecule itself is not the hard part. The hard part is delivery: the mRNA has to be wrapped in a protective fatty shell called a lipid nanoparticle to survive in the body long enough to reach the right cells and get taken up successfully. Getting that delivery vehicle right, consistently, at scale, is its own significant manufacturing discipline.

6. Antibody-Drug Conjugates (ADCs)

As covered in detail in our earlier article on this modality, an ADC combines an antibody, a chemical linker, and a highly toxic small-molecule payload into one product. It requires three separate manufacturing processes, biologic and chemical, joined together in a conjugation step, followed by purification to remove leftover unattached antibody and excess toxin. This hybrid biologic-plus-chemistry structure is what pushes ADCs meaningfully above a standard antibody in difficulty.

7. Radioligand Conjugates (RDCs)

A radioligand conjugate follows the same basic design as an ADC, an antibody or similar targeting molecule attached to a payload, except the payload is a radioactive isotope instead of a toxic chemical. That single substitution adds an entirely new layer of complexity on top of everything an ADC already requires: production has to happen in licensed radiopharmaceutical facilities with radiation shielding, specialized handling, and, because radioactive material decays over time, extremely tight timelines between manufacturing and getting the drug into a patient.

8. Cell Therapy (CAR-T)

CAR-T therapy takes a patient's own immune cells, genetically re-engineers them outside the body to recognize and attack cancer, then infuses them back into that same patient. This is where manufacturing stops being a repeatable industrial process and becomes something closer to a customized, individual medical procedure performed for every single patient. Each batch is literally a different person's living cells, which means no two manufacturing runs are ever quite identical, and the entire process, from drawing the patient's blood to delivering the finished product, typically takes two to six weeks under tight logistical coordination.

9. Gene Therapy

Gene therapy delivers a corrected or replacement gene directly into a patient's cells, usually carried by an engineered, deactivated virus, most often a class called AAV. This sits at the top of the complexity ranking because it combines nearly everything difficult about the categories above it: it requires growing living viral particles at scale, a purification process that has to separate genuinely therapeutic viral particles from empty, non-functional ones that are chemically almost identical, and the therapy is usually intended as a single, permanent, one-time treatment, which means there is far less room for error than with a drug a patient can simply stop taking if something goes wrong.

The Ranking at a Glance

Where This Ranking Matters Most

This complexity ranking isn't just an academic exercise. It maps directly onto where artificial intelligence can and can't help today: the modalities near the top of this list, where the challenge is choosing the right molecule, are exactly where AI-driven discovery is producing real, verifiable results. The modalities near the bottom, where the challenge is physically building something consistently at scale, are exactly where AI's progress has been much more limited so far. A companion piece on RichStorm looks specifically at where AI actually helps today, and who already controls the manufacturing capacity that AI hasn't been able to open up.