NVDA at $196.93 (+4.41% YTD) and GOOGL at $367.03 (+16.61% YTD) are both flashing value signals that the broader market is underappreciating — NVDA on a relative basis given its earnings power, GOOGL on an absolute basis with a credible 15% undervaluation case. The divergence in their paths matters: GOOGL has delivered a staggering +111.10% over 52 weeks and is a fundamentals story with regulatory overhang, while NVDA is a sentiment-and-cycle story where the DeepSeek competitive threat narrative is keeping a lid on what should be a higher multiple. Both deserve closer attention ahead of the second half.
Let's start with NVDA, because the setup here is genuinely interesting. At $196.93 with a TTM P/E of 30.2x and a market cap of $4.77T, the chip giant is trading at a multiple that looks almost pedestrian relative to where it sat a year ago — and relative to the earnings trajectory the company has been on. The semiconductor sell-off that Ross Gerber flagged as a buying opportunity has brought NVDA into a range where long-term buyers are re-emerging, evidenced by today's +0.71% move on 122.7 million shares. With the S&P 500 up roughly 9% in H1 2026 and Argus forecasting an additional 5-7% in H2, the macro backdrop is not hostile to risk assets — and NVDA is one of the highest-beta names in the cohort at 2.21x. A market that continues to grind higher should disproportionately reward that beta.
The DeepSeek angle is the key bear case overhang I want to address directly. The narrative that DeepSeek's own chip development reduces NVIDIA's TAM is real, but it's also premature. Developing competitive AI silicon at scale — with the yields, software stack maturity, and supply chain depth that NVIDIA has built — is a multi-year effort. In the near term, more compute-efficient models actually increase demand for inference infrastructure, which still runs on NVIDIA hardware. This isn't a new argument, but the market keeps discounting it, which means the buying opportunity persists. At 30x TTM earnings on a company with 62.9% profit margins and a dominant position in AI training and inference, NVDA is not expensive. It's cautiously priced.
GOOGL is the more compelling short-term fundamental case. The +111.10% 52-week return is eye-catching, and one might assume the stock is stretched — but the fundamentals tell a different story. At 28.0x TTM P/E and an EV/EBITDA of 19.8x, Alphabet is priced like a steady compounder, not a hypergrowth AI platform. Revenue TTM of $422.5B and net income of $160.2B in Q1 2026 alone suggest the earnings engine is running hot. The 15% undervaluation thesis from recent analysis is not outlandish when you model Google Cloud's growth trajectory against current multiples — cloud is compounding at rates that should command a premium, and the market is still applying a blended discount for regulatory risk that may be excessive.
The regulatory overhang is real and I won't dismiss it. DOJ antitrust proceedings remain a structural risk that could affect the search monopoly, and any forced structural separation of Google Search or Chrome would be a material negative. But Alphabet's AI integration across Search, Workspace, Cloud, and Waymo creates a diversification of earnings streams that the market has historically underpriced. GOOGL's beta of 1.25 means it participates meaningfully in bull market conditions without the extreme volatility of NVDA — that's a favorable risk-adjusted entry point for investors who want AI exposure without the 2.2x beta ride.
Comparing the two within the broader Mag 7 context: NVDA at $4.77T market cap has now overtaken AAPL at $4.56T market cap as the largest company in the cohort by this metric, which is a significant narrative shift. GOOGL at $4.48T is essentially neck-and-neck with Apple. The three largest companies in the world are now all AI-infrastructure-exposed names, and all three are trading at multiples that — while not cheap in absolute terms — are reasonable relative to their growth vectors. Against MSFT at 23.0x with declining revenue narratives, or TSLA at 383x on paper-thin margins, NVDA at 30x and GOOGL at 28x look like the rational value plays within the Mag 7 right now.