I'm fixing my portfolio mess up

TSMC added to the portfolio

Dear Investors,

Did you notice that you didn’t get a newsletter last weekend?

If not, tsk tsk.

The missed newsletter was intentional. I decided that we will experiment with the format and timing of the newsletter.

I’m leaning toward higher quality rather than higher quantity.

Quantity is the social media way, but I believe that neither the reader nor the writer benefit from this.

I want each edition of ROA to be worthy of your time, so bear with me as we try a few new things.

Sincerely, Raj

Microchips don't glow, but they kind of look like this

The mess

About six months ago, I skewed the portfolio weights significantly. Our top two stocks became 60% of the portfolio. This wasn’t meant to last long, but it did. Since then the portfolio performance has been like riding a bucking bronco. You can see the volatility in the graph below. It became more squiggly.

Our largest holding, Nvidia, is a volatile stock and it kept doing its thing. Our second largest holding, Novo Nordisk, hit a bad time. The market has lost some belief in the healthcare companies leading the weight loss and diabetes revolution. Consequently the share price drifted lower.

Extraordinary Companies portfolio performance

Novo Nordisk

I've seen arguments from professional investors against healthcare stocks, including Novo Nordisk. The concern is about them being one-hit wonders where the stock price peaks and then declines. The problem with arguments like this is that they are anecdotally convincing but they ignore the company’s actual performance.

I don’t like to use the N-word to describe myself, but I can be a bit of a Nerd 🤓. The problem that causes here is that I don’t easily buy into the one-hit wonder story because I cannot see it in the numbers.

Novo’s financial performance was and is currently brilliant. These days that is largely due to sales from it’s patented diabetes and weight loss drugs, Ozempic and Wegovy. The future financial performance will depend on Novo’s ability to sell new drugs that are in its pipeline.

Drugs have to go through 4 phases of clinical trials before they can be sold to the public. The problem is that failure rate through the phases is reasonably high. So not all drugs make it all the way through phase 4.

Novo has valid patents until 2032. Thereafter, Ozempic and Wegovy can be copied and made by generic drug manufacturers. When that happens, the price of these drugs will drop and so will Novo’s profitability.

Investors are concerned that Novo won’t have anything new that makes it through clinical trials, which can maintain profitability. But Novo actually has things in its pipeline and no one really knows what will happen by 2032.

There is a logical fallacy with this type of reasoning. Investors penalise Novo because we know about its patents and clinical trials pipeline. But why don’t we penalise Nvidia? How do we know if they will be making the best GPUs in 2032? How do we know that any tech company will maintain its position 7 years hence?

We’re penalising Novo because we have slightly better visibility into its future product lines. Almost seems like we should be doing the opposite.

Here at ROA, we are interested in owning the best companies in the world. The question is should we abandon Novo, one of the best companies in the world, because its price is dropping or should we be buying more?

My belief is that it is erroneous to let price determine how you invest. But at the same time, Novo is slamming the portfolio’s performance.

Fortunately, I have a really cool boss and he doesn't mind me blowing some cash on Novo. Although between us, I will admit that sometimes, he is a bit of a nerd.

Anyway, for now, I’m hanging onto Novo Nordisk.

Nvidia

Then we come to this week’s problem. A Chinese company released an artificial intelligence (AI) called Deepseek. They claim that it cost $5 million to create compared to $100 million for ChatGPT.

The entire tech world freaked out and the value of semiconductor companies took a major dip this week. Nvidia led the fall with a drop of about 20%.

The Chinese company claimed that they developed Deepseek using old Nvidia GPU chips. Due to sanctions imposed by the USA, the Chinese have not been able to access the latest Nvidia GPUs. The thinking is that if they can make a better AI at a fraction of the cost using old technology, then the prospects for chipmakers is not that bright.

Two obvious questions arise. The first is are we comparing apples to apples regarding the costs? Some doubt exists on whether the Chinese costing includes everything. The second question is whether the use of Nvidia GPUs is being downplayed, so as not to alert the US that China has gotten hold of new GPUs through other channels.

Neither of these questions concern me.

If people are able to make better AI using less computing power, it will not lead to lower sales of Nvidia GPUs, rather it will result in more people developing AI applications. Historically, we have examples of this in the electronics industry.

The faster and cheaper telecommunications equipment gets, the more and more applications appear and the more data you use. Think about your own usage. You probably have access to data over fiber optic cable or 5G mobile data. Are you using more or less of it than a decade ago when it was slower and more costly?

What I am suggesting is that the demand for Nvidia GPUs wont decline over time, even if AI gets more efficient. What will happen is we will create more AI and use it more freely.

Data is so cheap these days, that most of us don't even bother to type in a website address (URL), you just search for the company's name in Google and then click the first search result, right? You wouldn’t do that if data was slow and expensive.

By the way, don’t be surprised if in future, some of the best GPUs in the world come from China. That’s what happens when you restrict technology, the other guys figure it out for themselves.

Fixing the portfolio

Anyway, this dip in tech prices gave us a chance to start fixing the portfolio weightings. We did this by making our first purchase in quite a while.

We bought Taiwan Semiconductor Manufacturing Company (TSMC). TSMC produces Nvidia’s GPU chips. TSMC will introduce a 2 nanometer manufacturing process in 2025. This will be the most advanced technology in the semiconductor industry in terms of energy efficiency and chip density.

If you're wondering what 2 nanometer means, it means that the smallest features on the chip will be 2 billionths of a meter in size. For comparison, a human hair is 100,000 nanometers wide.

Back in my days as a semiconductor engineer, the smallest feature size I used was 350 nanometers. That was a little while ago, but it is huge in comparison to what TSMC can do.

The plan to bring the portfolio weights back into line will happen through the purchase of new Extraordinary Companies over time. While that happens, we will ponder Novo Nordisk and Nvidia.

Now let’s look at what we bought.

Stock information

Name:

Taiwan Semiconductor Manufacturing Corporation

Ticker:

TSM (USA) & 2330 (Taiwan)

Country:

Taiwan

Exchange:

New York & Taiwan Stock Exchange

Industry:

Semiconductors & Semiconductor Equip

Market capitalisation:

$900 billion

Financial performance

Key points from the financial analysis:

  • Return on capital employed: It is averaging 24%, which is excellent. Our rule of thumb is that 15% is good.

  • FCF margin: It is averaging 20%, which is excellent.

  • Revenue growth: Healthy and expected to remain good as the demand GPUs is high.

  • Stock-based compensation: Low usage of SBC, which is good, because we don’t want the company being given away.

  • Valuation: The FCF yield is 3.1%. That is not cheap but fair for this sector.

  • Capital expenditure: Capex is high. This is a feature of semiconductor manufacturing. It advances at a high pace, so upgrades are constant. The mitigation for this is that the ROCE is high, so the capex is money well spent.

  • Equity / assets: The company is financed with 65% equity and 35% debt. We don’t want too much debt, but this is reasonable.

TSMC financial analysis

Conclusion

TSMC is an Extraordinary Company.

Thank you to the world for freaking out due to Deepseek because it gave us an opportunity to purchase TSMC at a discount. A few days ago it dropped 15%, we purchased the stock and it is already up 7%.

We’re on the prowl for more Extraordinary Companies, so if you come across anything, drop me an email.

As always, be nice to your favourite nerds.

I would like to hear what you’re up to.
DM me on social media or email me.

Reply

or to participate.