Heading into October 2026, semiconductors remain the hottest address in AI investing, but in the host's view the stocks everyone watches are expensive; the real opportunity hides in four overlooked chip stocks . The claim is crisp: his discounted cash flow work on Broadcom, TSMC, Nvidia and Qualcomm puts their fair values 31 to 48% above current prices. He spends a few minutes on each name, laying out how he built the math, which assumptions it rests on, and what he himself is doing. The horizon is the same for all four: the next 12 to 18 months.
Inside the calculation: cash flow and discount rate
The valuation method is identical across all four names: project the free cash flow the business will generate from today into the very long run, discount it back to today, and divide the total by the share count. The discount rate is the weighted average cost of capital , derived through the capital asset pricing model CAPM from the company's beta, the risk-free rate and the debt-versus-equity mix. In the Broadcom example the beta is 1.22 and the discount rate 11.6%, justified by rising risk-free rates as US 10-year yields climbed. The method is transparent, yet the result hangs entirely on these two inputs, the growth path and the discount rate; a small shift in assumptions moves fair value by tens of dollars.
Broadcom: architect of the giants' custom chips
Broadcom's story is helping giants like Alphabet design their own AI accelerators ; in other words, it sits in the engine room of chip demand. According to the host, sales have boomed in the literal sense, profitability is expanding, and the company is sold out for the rest of the year, with next year looking similar. According to SDxCentral reporting from April 2026, Broadcom signed a long-term supply and custom TPU development agreement with Google running through 2031, while Anthropic secured access to about 3.5 gigawatts of computing capacity starting in 2027.
The math goes like this: he projects $50 billion of cash flow in 2026, then $90 billion the next year and $123 billion the year after, followed by a transitional path of $144 billion, $168 billion and $197 billion by 2031. Discounting these flows at 11.6% and dividing by the share count gives a fair value of $523 per share, implying more than 48% upside from the $353 market price. Convinced by this picture, the host says he bought the stock for his portfolio earlier this year and is considering adding more.
TSMC: everyone's factory
He sets the bar high on TSMC: in his view it is the world's best manufacturing company in any industry, with a track record spanning more than two decades to prove it. Broadcom, Marvell, Qualcomm, Nvidia, AMD, Apple; every giant he names outsources at least part of production to TSMC, and that trust is what makes their asset-light business model possible. According to EETimes reporting from July 2026, TSMC raised its 2026 capital budget to a range of $60 to $64 billion and added $100 billion to its Arizona investment, after reporting $40.2 billion of second-quarter revenue.
Most of what the company manufactures is sold out for this year and next, and it is spending tens of billions to expand capacity. He flags tension between China and Taiwan as the biggest risk, and describes geographic diversification as the company's answer: new manufacturing capacity going up in the United States, Europe and Japan. The same method yields a fair value of $647 per share, implying more than 43% upside from the $451 market price. He does not own this one yet but says he is keen to buy.
Nvidia: the showcase of the AI boom
Nvidia is, famously, the company that started the AI wave : its accelerated GPUs paved the way for large language models and created the possibilities on display today. Sales have boomed and keep booming; most products are sold out for this year and next, and selling has already begun for potential 2028 releases. Selling to a small number of large customers gives it visibility: customers put down deposits and sign order agreements years into the future. In results reported by Yahoo Finance, Nvidia posted record revenue of $57 billion and $51.2 billion of data center revenue in the third quarter of 2025, with CEO Jensen Huang saying cloud GPUs were sold out.
The projections are the boldest here: he forecasts $96 billion of free cash flow in 2026, soaring along an explosive path to more than $402 billion by 2030. The result is a fair value of $332 per share, implying more than 47% upside from the $225 market price. He adds that he owns Nvidia in his portfolio and wants to buy more.
Qualcomm: a new door into the data center
Qualcomm is the exception on the list: it is not an AI player yet, but its new data center products are in testing and the customer roster already impresses. Its core smartphone market is struggling, ironically because of AI; soaring memory and storage costs are a headwind for the industry. Against that, the company has expanded relatively successfully into automotive and personal computing. According to Reuters reporting from September 2026, Qualcomm signed a custom-chip development deal with Amazon for AI inference and granted Amazon the right to buy about $4 billion of its stock. According to a Tech Insider analysis from August 2026, Qualcomm unveiled the Dragonfly C1000 server processor on June 24, 2026, with Meta agreeing to deploy the chip in its next-generation server fleet and mass production starting in the second half of 2028.
The math on this name is more cautious: he forecasts 2026 free cash flow of $10 billion, below the $12 billion of 2025, then mediocre growth to $11 billion in 2027 and $11.7 billion in 2028, with the real jump in 2029 from $11.8 billion to $18.5 billion. So the valuation rests not on exponential growth but on a reasonable expansion after 2029; the $268 fair value implies more than 31% upside from the $204 price. The host sees risk to the upside, meaning his own numbers may understate Qualcomm, and he is interested in opening a new position.
Cost pressure on the phone side
Independent data backs up the strain in Qualcomm's core market. According to IsaiahResearch data, smartphone DRAM prices rose roughly 80-90% quarter on quarter in the second quarter of 2026, with serious supply shortages in LPDDR5/X specifications. According to KhaleejTimes reporting from September 2026, the iPhone 17's price rose from 3,399 to 3,799 dirhams within a year, with Samsung and Google raising prices too; Counterpoint Research estimates global phone prices climbed about 15% this year. That picture explains why Qualcomm has to expand beyond phones.
| Stock | Fair value | Market price | Upside |
|---|---|---|---|
| Broadcom (AVGO) | $523 | $353 | Over 48% |
| TSMC (TSM) | $647 | $451 | Over 43% |
| Nvidia (NVDA) | $332 | $225 | Over 47% |
| Qualcomm (QCOM) | $268 | $204 | Over 31% |
Key moments
AI commentary
"The host values four chip stocks with discounted cash flow and finds 31-48% upside. The method is transparent but rests on a single scenario; this piece covers both the math and where the math can break."
AI assessment
The strongest objection is to the method itself: in a single-scenario discounted cash flow model, the answer is a hostage of its assumptions. Broadcom generating $197 billion of cash flow in 2031 and a 11.6% discount rate holding steady for years are bold claims stacked on top of each other, with neither a bear case nor a sensitivity table in sight. Presenting 31-48% upside with a firm 12-to-18-month horizon makes the picture look more deterministic than it is.
There are gaps too. The geopolitical risk at TSMC, the very China-Taiwan tension the host himself calls the biggest risk, never turns into even a single line of discount in the valuation. At Qualcomm, the abrupt 2029 jump in cash flow has no visible foundation; which contracts, which products, which capacity will carry it remains unanswered. On conflicts, the picture is honest: the host discloses owning Broadcom and Nvidia, the video carries a brokerage sponsorship, and he promotes his own book and paid spreadsheets. None of that invalidates the math, but it calls for reading glasses.
The practical takeaway for readers comes in three parts: first, these prices belong to the recording date, so check live quotes and the latest balance sheets before any decision. Second, all four stocks ride the same macro bet, namely that AI capital expenditure keeps flowing; the list feels diversified, but the diversification is thinner than it looks. Third, for the convinced, scaling in gradually and sizing the position to one's own horizon beats going all in on the strength of a single video.
Sources
8 links; no other published story cites them. Stories sharing a link do not confirm each other; a source's origin is not inferred from how often it is cited.
- @youtube.com YouTube — Parkev Tatevosian, CFA
- @sdxcentral.com SDxCentral — Broadcom secures deal with Google for next-gen AI racks
- @eetimes.com EE Times — TSMC boosts 2026 expansion budget, adds $100B to US investment
- @finance.yahoo.com Yahoo Finance — Blackwell sales off the charts, cloud GPUs sold out
- @reuters.com Reuters — Qualcomm strikes AI chip deal with Amazon
- @tech-insider.org Tech Insider — Qualcomm Dragonfly C1000 data center CPU
- @isaiahresearch.com IsaiahResearch — Q2 2026 smartphone memory trends
- @khaleejtimes.com Khaleej Times — phone prices rising on memory costs
semiconductor stocks · equity valuation · broadcom · tsmc · nvidia · qualcomm · cash flow