The current stock price movement of NVIDIA (NVDA) completely fails to match the market's expectation of an 83% revenue surge this year. Instead, the stock is pricing in the realization of every conceivable risk over the next two years. This dynamic creates a potential opportunity for investors with a long-term horizon.
Most semiconductor investors chase stocks with the fastest expected growth over the next several years. Year-to-date, NVIDIA shares have risen only 10%. In contrast, its long-time rival Advanced Micro Devices (AMD) has surged 142%, while memory chip maker Micron has gained 213% this year, fueled by the explosion in AI data center demand. The broader Philadelphia Semiconductor Index has risen 71% overall.
Data from S&P Global Market Intelligence shows that AMD trades at a forward price-to-earnings (P/E) ratio of 53 times, while NVIDIA's forward P/E has fallen to its lowest level since July 2021. The company's current forward enterprise value to EBITDA (EV/EBITDA) multiple is below 17 times, significantly lower than its five-year average of 36 times.
The bearish thesis for NVIDIA is now clearly defined. Since OpenAI launched its first chatbot, ChatGPT, NVIDIA's AI chip business has boomed, but a wave of competitors has entered the market to develop their own AI chips. These include startups like SambaNova, Cerebras, and Groq (which ultimately chose to license its technology to NVIDIA). Meanwhile, Google has been developing its own AI chips for years, now selling them externally and also leasing them through its cloud business. Amazon has also developed its own AI chips and is pushing toward commercialization. Meta, Microsoft, OpenAI, and Ansopar have all begun developing their own AI chips.
Furthermore, AMD is set to launch its first complete AI server system later this year. This solution highly integrates AI chips with supporting hardware to ensure stable soft and hardware co-performance. The product, named Helios, aims to deliver capabilities similar to NVIDIA's recent Grace Blackwell and Vera Rubin AI chip platforms.
On an objective level, NVIDIA's revenue base is already enormous, making it naturally more difficult to sustain high growth rates compared to smaller players.
The core bullish argument for NVIDIA is that the company's growth potential is far from being as pessimistic as the stock price suggests. Morningstar analyst Brian Colello states that with NVIDIA's current stock price around $212, the market is implicitly pricing in growth that essentially stalls after 2027. He believes the fair value should be closer to $280, which corresponds to a valuation of about 16 times the company's projected sales for fiscal year 2029 (ending January of that year). According to his model, if NVIDIA maintains its recent high-growth trajectory, with revenue and adjusted earnings per share each growing over 45% annually through fiscal 2029, this valuation is reasonable.
"Looking out two to three years, NVIDIA is very cheap. The core question is: Can the capital expenditure of major cloud companies and enterprises remain high in two years? Can NVIDIA hold onto its dominant market share? Our answer to both is yes, which is why we believe the stock is undervalued," Colello said.
Analysts forecast that NVIDIA's revenue for fiscal year 2028 (ending January) will grow 42% to $560 billion, followed by another 23% increase in the subsequent fiscal year. AMD's expected growth rates are not significantly higher. AMD's revenue grew 34% last year to $34.6 billion. Institutions estimate its revenue for the fiscal year ending December 2027 will grow 57% to $78 billion, with another 36% growth in the following year. Such growth rates do not justify the substantial valuation premium that AMD currently enjoys.
Despite this, investors have stopped treating NVIDIA as a growth stock. John Belton, a portfolio manager at Gabelli Funds, says that semiconductor capital is flowing into areas where the supply-demand gap is most pronounced and where there is significant unrealized growth potential. "Right now, NVIDIA doesn't fit that investment thesis," Belton noted.
Belton has been increasing his positions in AMD and Micron, while holding his NVIDIA position steady, but has not added to it in recent months.
"AMD has a stronger speculative element," he said, adding that AMD's products are still in their early stages, and if their technology roadmap is successful, the upside potential for the stock is greater.
However, the market may be underestimating the high barriers to entry that NVIDIA has built over years of dedication to AI chips. Should technology companies like OpenAI scale back their AI capital expenditure, many of the chip-making newcomers that jumped into the fray might abandon their in-house efforts and return to buying chips from NVIDIA. In simple terms, if the AI industry hits a downturn, the impact on NVIDIA is likely to be less severe than on newer chip players, whose products have not yet been tested through a full market cycle.
Despite the risks, NVIDIA continues to hold a commanding lead in the AI chip market. A recent report from The Information indicates that in the AI inference chip market (used for running models, as opposed to training chips), NVIDIA's market share has actually increased.
There is no sign of a slowdown in the industry's current push to develop new chips. Alphabet, the parent company of Google, on Wednesday raised its capital expenditure forecast for 2026 and plans to expand its capital spending further next year.
Additionally, NVIDIA's strategy of investing in potential customers, including emerging cloud providers like Nebius and CoreWeave, has recently drawn criticism. But if the industry enters a downturn, this strategy could effectively hedge against business risks.
Comments