Will Anthropic Stay On Top? Half Of Tech Leaders Didn’t
Today’s big story: Will 2027 Be A 2022 Repeat For Stocks?
Anthropic started trading on Nasdaq on Thursday. Its offer price was $83 a share, about $1.1 trillion. It closed its first day at $97.17, about $1.3 trillion.
ChatGPT was the AI app that everyone tried first. Yet Anthropic is the one that listed. So did OpenAI lose its lead, or is the race still early?
Tech has run races like this before. In 2004 the biggest name in internet search was Yahoo. In 2008 the biggest phone maker was Nokia. Both lost the lead within five years. We checked six such races to see how often the early leader held on.
The short answer: only two of six early leaders held on. The companies that passed them were already making money.
Three of the six early leaders lost the lead within five years. If you own a leader and it slips, you may have to sit through a fall in its stock. When the market has fallen, the Trefis High Quality (HQ) Portfolio has lost about 81 cents for every $1 its benchmark lost, on average. See the HQ record.

Does The Early Leader Usually Win?
Only two of six held. In each race, we picked the early leader. It is the company with the most sales in the race in the year the race went mainstream. Where a company reports it, we counted only the business that is the race. That means Nokia’s phones, Intel‘s data-center chips, Tesla‘s and BYD’s cars, and Uber‘s rides. Then we checked who led five years later.
Of six races, the early leader held two and lost three. One is too close to call. Yahoo, with $3.6 billion of sales in 2004, was the first to fall.
| Race, year | Early leader | 5 years on | Held? |
|---|---|---|---|
| Search, 2004 | Yahoo $3.6B, Google $3.2B | Google $23.7B, Yahoo $6.5B | Lost |
| Phones, 2008 | Nokia $51.7B, iPhone $1.8B | iPhone $91.3B | Lost |
| Social, 2012 | Facebook $5.1B, Twitter $0.3B | Facebook $40.7B, Twitter $2.4B | Held |
| Rides, 2019 | Uber $10.7B, Lyft $3.6B | Uber $25.1B, Lyft $5.8B | Held |
| EVs, 2020 | Tesla $29.5B, BYD $11.9B | BYD $90.2B, Tesla $82.1B | Too close to call |
| AI chips, 2020 | Intel $26.1B, Nvidia $6.8B | Nvidia $193.5B, Intel $16.9B | Lost |
Source: Trefis, annual sales in the race from company filings. Nokia counts Devices & Services, Apple the iPhone, Uber Mobility, Tesla automotive, BYD automobiles and related products (2025 also includes batteries), Intel data center (2025: Data Center and AI) and Nvidia Compute & Networking (fiscal years ending January 2021 and 2026). Nokia and BYD are converted to dollars at each year’s average rate.
Look at the electric-car row. Tesla’s car business was 2.5 times BYD’s in 2020. By 2025, BYD’s car-and-battery business was $8 billion bigger than Tesla’s car business. A 2.5-to-1 lead was gone in five years.
Three of the races are old enough to check at ten years. In all three, the five-year result did not flip back. In 2014, Google made $66.0 billion to Yahoo’s $4.6 billion.
Could You Have Seen It Coming?
Yes, by following the profit. Being bigger did not decide it. Yahoo had only 12% more sales than Google in 2004. Growing faster did not decide it either. Twitter’s sales more than doubled in 2013, while Facebook‘s grew 55%. Facebook still won by a wide margin.
What separated the races was whether the challenger already made money in the year the race went mainstream. We measure that with operating margin, the share of sales a company keeps as profit from running its business. Google kept 20% of its sales as operating profit in 2004. Twitter lost 24% in 2012.
| Challenger, year | Operating margin | Result |
|---|---|---|
| Google, 2004 | 20% | Passed Yahoo |
| Apple, 2008 | 19% | Passed Nokia |
| Nvidia, 2020 | 37% | Passed Intel |
| BYD, 2020 | 5% | Caught up with Tesla |
| Twitter, 2012 | -24% | Fell further behind |
| Lyft, 2019 | -75% | Fell further behind |
Source: Trefis, operating income divided by sales in the year the race went mainstream, from company filings. Apple and BYD are whole company, since neither reports profit for phones or cars alone. Nvidia is Compute & Networking. The others are whole company.
Every challenger that passed the leader was already profitable at the start. Google’s margin was 20%, Apple‘s was 19% and Nvidia‘s data-center business was at 37%. The two challengers that lost money stayed behind.
A challenger that earns money can keep spending on its product for years. It does not need anyone’s permission. A challenger that loses money has to raise it, and that money can run out. The leaders that lost were selling what customers used to want, such as Yahoo’s portal pages, Nokia’s keypads and Intel’s general-purpose chips. The challengers were paid well for what came next.
So What About OpenAI And Anthropic?
The test is cash. OpenAI does not publish its finances, so this test cannot yet be run on it. Anthropic now must publish. In our view, its filing looks like those of the challengers that won.
Its business brought in $6.6 billion of cash in the first nine months of 2026. Its sales were running at $81 billion a year in October. Customers who pay by usage made up 79% of its sales. More than 11,900 customers pay it $100,000 or more a year. Those are business customers, not consumers trying a free app.
The risk is in the same filing. Model versions released in the six months to September made over 80% of Anthropic’s sales in the September quarter. So the lead has to be won again every six months. And if Anthropic is now the early leader, the record gives it no better than even odds of still leading in 2031. We look at how it got here in Anthropic: $41 Million To $33 Billion. Why?
How Do You Invest When No One Knows Who Wins?
Own the race. In our view, the investors who did fine in every race above owned the whole field. An S&P 500 fund held Google as it pulled away from Yahoo. It held Nvidia as it passed Intel. Today Amazon, Google and Microsoft each own stakes in the AI labs and rent them computing power. We look at that in Anthropic Or OpenAI: Can Both Win?
In our view, a seven-figure portfolio does not need to predict the winner. Keep the index fund you already hold for the whole field. Add a small slice of the runner you believe in. Size it so that a 30% fall in it would barely dent your total. Then check one number each year: is the challenger making money?
The point to remember: half of the six early leaders lost the lead within five years. The companies that passed them were already making money. So do not bet on one runner. Own the race.
That only works if no single runner grows into too big a share of your money. If one already has, it helps to plan how you would step back from it.
Too Much Riding On One Stock?
If Anthropic is a big part of your wealth, so are its risk and its tax bill. What’s your exit plan? Our partner advisory firm offers a free 30-minute strategy review covering a floor under losses and a lower-tax way out.
For education, not a recommendation to buy or sell.
For questions about this article or underlying research, email help@trefis.com.