Bitcoin 
Ethereum 
Tether 
BNB 
USDC 
XRP 
Solana 
TRON 
Lido Staked Ether 
Cardano 
Gram (prev. Toncoin) 
Avalanche 
The Nvidia share price drop of 17% from its record high has brought the AI chip trade back to a question investors thought they had already answered: is the pullback an entry point, or the first leg of something worse? At £197, Nvidia (NVDA) sits 20% above where it was a year ago, yet the pace that made it a five-year compounder of 882% has clearly moderated.
To put that run in concrete terms: a £10,000 position five years ago would be worth £98,200 today. The problem is that most investors who were going to buy did so much earlier, and those who didn’t are now staring at a stock that is both cheaper than it was and still carrying real risk.
Where the Nvidia Share Price Drop Leaves the Valuation
The valuation case for NVDA at a trailing price-to-earnings (P/E) multiple of 40 times is, by the standards of this cycle, almost restrained. First-quarter revenue grew 85% year-on-year. Free cash flow rose by a similar margin to $48.6bn and is forecast to exceed $200bn across the full year. Those numbers justify a premium; the question is how large a premium.
Compare that with its two most popular peer plays. Micron Technology, up 667% over the past twelve months before its recent slide of around 13%, trades at 122 times earnings. Marvell Technology, which fell 18% in the same week that the Nvidia share price drop accelerated, carries a stated P/E of 172 times. Both figures demand scrutiny.
| Company | P/E Ratio | Recent move |
|---|---|---|
| Nvidia (NVDA) | 40x | -17% from all-time high |
| Micron Technology | 122x | -~13% over last week |
| Marvell Technology (MRVL) | 172x | -18% over last week |
Marvell’s Earnings Complexity Clouds the Simple P/E Read
The 172 times multiple attributed to Marvell Technology deserves a closer look. On a GAAP basis, Marvell reported a full-year net loss of $(885.0) million for fiscal year 2025, or $(1.02) per diluted share. A positive P/E derived from that figure is undefined. The 172 times figure almost certainly uses non-GAAP or forward earnings: non-GAAP net income for the same fiscal year was $1.377 billion, or $1.57 per diluted share. Investors comparing Marvell’s multiple against Nvidia’s 40 times need to be certain they are comparing like with like.
The more recent picture is mixed. For the fourth quarter of fiscal 2025, GAAP net income recovered to $200.2 million. And in the most recent quarter (Q1 fiscal 2027, reported 27 May 2026), Marvell posted a non-GAAP EPS of $0.80, beating the analyst consensus of $0.75 by 6.67%. On 8 July 2026, Marvell’s closing price on Nasdaq was $231.71.
Strategically, Marvell has been moving quickly. Marvell’s investor relations page shows the company joined the NVIDIA AI Ecosystem through NVLink Fusion, launched what it describes as the industry’s first 102.4 Tbps switch designed for AI and cloud data centres, and completed the acquisition of XConn Technologies in February 2026. The strategic direction is coherent. The question is whether the current price already reflects it.
The Risks That Could Unravel the AI Trade
The macro backdrop adds another layer of uncertainty. The snippet’s concern about the Federal Reserve tightening policy has not gone away: any rate increase aimed at cooling the US economy would almost certainly compress multiples across high-growth tech, and the stocks with the most stretched valuations would take the largest hit.
The deeper structural worry is whether hyperscalers will generate adequate returns on the hundreds of billions being committed to AI infrastructure. The dotcom comparison is worth acknowledging but also worth qualifying: unlike most dotcom-era plays, the largest AI infrastructure companies are generating real profits today. That does not make them immune to a valuation reset, but it changes the character of the risk.
For investors weighing the Nvidia share price drop against its peers, the ranking by risk is relatively clear. At 40 times earnings with 85% revenue growth and $48.6bn in free cash flow, NVDA presents the most legible risk-reward of the three. Micron at 122 times and Marvell at a non-GAAP-dependent 172 times leave far less margin for disappointment. Drip-feeding into NVDA on further weakness is a defensible approach; building a full position in either peer at current valuations asks for a great deal of confidence in a still-uncertain AI spending cycle.
The next test arrives with Nvidia’s second-quarter results: if revenue growth continues above 80% and free cash flow tracks toward that $200bn annual forecast, the thesis holds. A miss on either would justify the caution the market is currently pricing in.
