In which verticals are margins massively improved due to AI? - Big Pharma?

[investmd]

If one agrees that ML is a technology that is here to stay and can dramatically improve efficiency, what industries are significantly positively impacted from an investment pov? 

 

I'd suggest that industries where COGS are v. high could benefit massively. They are unlikely to lower their price of goods sold but if high costs can be lowered, likely leads to higher margins, higher EPS and PE expansion. 

 

I'm thinking of Big Pharma - cost of bringing a new drug to market is approximately $2B. AI can help reduce costs by decreasing chance of failure through 2 main areas: 

1. Drug Development: Enrolling patients in clinical trials that have phenotype to respond positively to the drug - ie right patient at right time for the drug; simulating clinical trials and digital twins
2. Drug Discovery: Generative AI to predict a novel target and design a molecule to fit the target, high throughput screening to identify optimal compounds for further screening, predicting toxicity

 

So, Strategy 1 could be to buy a basket of Big Pharma - stable, international, dividend paying stocks and watch their EPS go up because costs go down. 

 

On the other hand, if we think AI can result in cost savings in many many verticals, Strategy 2 could be making a bullish argument to buy the S&P 500 index at a time when most value investors think the S&P 500 is "overpriced". 

 

Thoughts on Strategy 1 or 2? 

[Spekulatius]

If every company will be using it, there should be no profit margin improvement . I suspect that will be true for Pharma. However, I think there will be some companies who use AI well and others don’t . I am not sure how we know which ones are going to be winners or not since I have a hard time believing they will disclose enough details. AI could advantage large Pharma over smaller biotech startups but even that isn’t a given. my suspicion is that AI native startups will probably do a better job to harness AI than big clumsy organizations.

 

[UK]

5 hours ago, Spekulatius said:

If every company will be using it, there should be no profit margin improvement . I suspect that will be true for Pharma. However, I think there will be some companies who use AI well and others don’t . I am not sure how we know which ones are going to be winners or not since I have a hard time believing they will disclose enough details. AI could advantage large Pharma over smaller biotech startups but even that isn’t a given. my suspicion is that AI native startups will probably do a better job to harness AI than big clumsy organizations.

 

I also think, that longer term this should be competed out, so the winner longer term should be general economy via increased productivity and maybe some of AI providers, maybe some new, like Antrophic or Open AI, but also very likely usual suspects/current hyperscallers, like Google, depending on how market structure will develop, how much winner takes all it willl be and how capex intensive it eventually stays. I am also not sure about trying to pick winners for the long term in some of the current bottlenecks, like energy etc, which are commoditized, though this could work for a mid term. I think the main goal should be just not to get under this whole AI steamroller disruption with your investments, simillarly like it was with the internet. 

 

Edited February 7 by UK

[chenisheng]

What do you think about true luxury companies? Their margins probably will increase as their prices can be maintained/increased. Their marketing expenses, etc. can be lowered by generative AI 

[Marco Van Basten]

3 hours ago, chenisheng said:

What do you think about true luxury companies? Their margins probably will increase as their prices can be maintained/increased. Their marketing expenses, etc. can be lowered by generative AI 

So apparently, the resale premium of Hermes bags and Rolex watches has been going down, so perhaps the gap between supply and demand and the pricing room has been shrinking lately.  Similar things are supposedly happening in the Ferrari land.  (No personal knowledge of this sadly...)

[investmd]

3 hours ago, chenisheng said:

What do you think about true luxury companies? Their margins probably will increase as their prices can be maintained/increased. Their marketing expenses, etc. can be lowered by generative AI 

Hmmm... for luxury items they probably already have huge margins - that is the price they charge is way higher than the cost. Thus, saving a touch on marketing by using generative AI...my guess is doesn't move the needle. 

 

My guess is impact occurs where a)cost of goods are high and b)where cost of goods can be significantly affected by AI. 

 

V. v. rough back of the envelope numbers:

Le's assume annual revenue for Big Pharma at approximately $65B and expenses of $50B'ish meaning $15B of annual profit.  

• R&D makes up 20% of revenue ($13B), COGS 25% ($16B), SG&A 30% ($19B)
• AI could cut R&D costs by 20% (better prediction of which molecules will be successful drugs and decrease cost of trial costs), COGS by 10% (supply chain costs, improve yields) and maybe 10-15% across Sales, marketing and G&A resulting in maybe $6B of savings. 

 

Thus with same revenue of $65B, earnings can increase from $15B to $20B+, meaning 30% increase in earnings. 

 

Yes, one has to calculate increased cost of using AI tools etc....but back of envelope math. 

 

Downside is that AI doesn't decrease costs and one is left with stable, boring Big Pharma growing at 4-6% a year with 3-5% dividend yields. 

 

Is this a game where if one owns a basket of Big Pharma, Heads you win big and Tails you have little to lose? 

 

 

[Spekulatius]

4 hours ago, chenisheng said:

What do you think about true luxury companies? Their margins probably will increase as their prices can be maintained/increased. Their marketing expenses, etc. can be lowered by generative AI 

I don’ t think AI slob marketing will work for luxury, that one needs creativity and human touch.  l think the part that is the most affected are customer service or support roles. For competitive  industries like airlines or consumer goods  l think any cost savings will be quickly competed away. I think in niche industries we could see marking improvement. Think specialized components or services. Software cos are already  laying off customer facing positions which usually were the last one to be cut. that has now changed with AI chatbots etc. 

Edited February 7 by Spekulatius

[Spekulatius]

5 minutes ago, investmd said:

Hmmm... for luxury items they probably already have huge margins - that is the price they charge is way higher than the cost. Thus, saving a touch on marketing by using generative AI...my guess is doesn't move the needle. 

 

My guess is impact occurs where a)cost of goods are high and b)where cost of goods can be significantly affected by AI. 

 

V. v. rough back of the envelope numbers:

Le's assume annual revenue for Big Pharma at approximately $65B and expenses of $50B'ish meaning $15B of annual profit.  

• R&D makes up 20% of revenue ($13B), COGS 25% ($16B), SG&A 30% ($19B)
• AI could cut R&D costs by 20% (better prediction of which molecules will be successful drugs and decrease cost of trial costs), COGS by 10% (supply chain costs, improve yields) and maybe 10-15% across Sales, marketing and G&A resulting in maybe $6B of savings. 

 

Thus with same revenue of $65B, earnings can increase from $15B to $20B+, meaning 30% increase in earnings. 

 

Yes, one has to calculate increased cost of using AI tools etc....but back of envelope math. 

 

Downside is that AI doesn't decrease costs and one is left with stable, boring Big Pharma growing at 4-6% a year with 3-5% dividend yields. 

 

Is this a game where if one owns a basket of Big Pharma, Heads you win big and Tails you have little to lose? 

 

 

If every Pharma co reduces the cost to bring a drug to market by let say 20%, do you think it will improve profitability , or it stays the same or it will reduce it?

[investmd]

@Spekulatius: good q - My thesis is that cost of drugs paid by insurers doesn't come down as cost of production comes down because Big Pharma is an oligopoly that then has pricing power.  Yes, Big Pharma are susceptible to regulators, but probably not susceptible to some startup that is going to undercut them on price, like one may see in other industries like auto or entertainment. 

 

Thus, my guess is decreasing costs of production will improve profitability for all Big Pharma - ie if a currently life saving drug for lung cancer by Pharma A costs $30K/year and Pharma B develops the next even better life saving drug at a lower price, the new drug is not coming to market at under $30K/yr. If the new drug is proven to be more efficacious it will come to market at same price or higher. 

[Marco Van Basten]

4 hours ago, investmd said:

@Spekulatius: good q - My thesis is that cost of drugs paid by insurers doesn't come down as cost of production comes down because Big Pharma is an oligopoly that then has pricing power.  Yes, Big Pharma are susceptible to regulators, but probably not susceptible to some startup that is going to undercut them on price, like one may see in other industries like auto or entertainment. 

 

Thus, my guess is decreasing costs of production will improve profitability for all Big Pharma - ie if a currently life saving drug for lung cancer by Pharma A costs $30K/year and Pharma B develops the next even better life saving drug at a lower price, the new drug is not coming to market at under $30K/yr. If the new drug is proven to be more efficacious it will come to market at same price or higher. 

The problem for big Pharma is that the US subsidizes the cost of medicines for the rest of the world, we pay several times more for the same thing.  If Trump had brains, he'd mandate that the price of any medication cannot exceed twice what the average of France/Germany/UK/Switzerland/India/Africa pay for it.  I think this kind of law should happen, and most likely will happen, and that will justly devastate the profits of big pharma.  

[dwy000]

17 minutes ago, Marco Van Basten said:

The problem for big Pharma is that the US subsidizes the cost of medicines for the rest of the world, we pay several times more for the same thing.  If Trump had brains, he'd mandate that the price of any medication cannot exceed twice what the average of France/Germany/UK/Switzerland/India/Africa pay for it.  I think this kind of law should happen, and most likely will happen, and that will justly devastate the profits of big pharma.  

Not sure govt can mandate specific pricing (esp a republican govt!)  Didnt they already issue MFN for what CMS will pay for some drugs?  But the effect is expected to increase international prices as opposed to reducing US prices. 

[frommi]

My thinking is that everything that has a true moat with high switchting cost and where payroll is the biggest cost will benefit. Ummm maybe software?
 

[brobro777]

That's da thing - if you throw tons of money at something, whether AI or fiber optics in 2000, you gotta get massive margin improvements or benefits, and they gotta come pretty soon

 

If that don't happen, you're gonna look like assholes...

[frommi]

Yes, and look at what they tell us at Goldman after working with Anthropic's models:

 

https://www.cnbc.com/2026/02/06/anthropic-goldman-sachs-ai-model-accounting.html
Argenti said that it was “premature” to expect that the technology will lead to job losses for those workers.

And who benefits from this? Its the customers of Goldman, getting faster response times. More revenue? nope. Better margins? nope. Goldman spends more on software, but when everyone in the industry does it, they all get lower margins. Its a race to the bottom, Prof Damodaran is also of that opinion. 

[Spekulatius]

1 hour ago, frommi said:

Goldman spends more on software, but when everyone in the industry does it, they all get lower margins. Its a race to the bottom, Prof Damodaran is also of that opinion. 

It will work for business with little competition that have a lot to customer support function. Some software business likely belong to this bucket.

 

I just saw a newsblip WDAY is current laying off 2% of customer facing workers. Thats on top of a reduction in 2024. WDAY has been accused of being run inefficiently so there is that, but I think the opportunities to save costs here are larger than probably most opteht business. Customer won’t walk way that easily either.

https://www.theregister.com/2026/02/04/workday_layoffs_400_jobs/

 

Edited February 8 by Spekulatius