Wall Street’s biggest bank just delivered a message that sounds contradictory at first. Chip stocks could fall further from here, and that is not necessarily bad news, according to one of the sector’s most closely watched analysts.
The call came two days before Nvidia reported earnings on Aug. 26. Nvidia beat on every key metric and guided ahead of consensus for Q3. But what Bank of America was telling investors going into that report was more nuanced than a simple read on the earnings print.
Bank of America sees 10% more SOX downside
Bank of America analyst Vivek Arya, who ranks 229 out of more than 12,000 analysts tracked by TipRanks with a 58% success rate, sees roughly 10% further downside risk for the Philadelphia Semiconductor Index (SOX).
A decline of that size would push the SOX back to its valuation discount versus the S&P 500 before ChatGPT’s debut in November 2022, CNBC reported.
Several factors are weighing on the group in the near term. Arya pointed to rising interest rates, public pushback on data center projects, geopolitical tension, worries over circular financing arrangements between AI companies, and heavy investor positioning, with chips now running about 13% overweight against the broader S&P 500.
The data center backlash has moved beyond social media. New York put a moratorium on large-scale projects. Pennsylvania started restricting approvals. Texas launched grid audits.
A year ago, none of that was happening. Now it is a real variable impacting how quickly AI infrastructure actually gets built.
More Bank of America:
- Bank of America sends sharp August stock market warning
- Bank of America CEO warns inflation will back Fed into a corner
- Bank of America just made a strong call on inflation, economy
“Though unjustified on fundamentals, we see another 10% downside risk to the SOX,” Arya wrote, adding that fourth-quarter and first-quarter seasonality has historically been strongly bullish, and that the sector’s 20 times forward earnings multiple looks cheap against a 70% earnings compound annual growth rate through 2028.
Bank of America named eight stocks it considers enhanced buying opportunities if the pullback plays out: Nvidia, Marvell, Micron, Lam Research, AMD, Intel, Analog Devices, and ON Semiconductor.
What Nvidia’s earnings beat means for chip stock investors
Nvidia reported quarterly results on Aug. 26, beating estimates across the board. Revenue came in at $96.22 billion, ahead of the $92.37 billion consensus, and up 106% year over year. Adjusted EPS hit $2.22 against a $2.09 estimate. Q3 guidance of $108 billion also topped the $104.6 billion consensus.
The stock dipped briefly after hours before recovering, as MSN reported. Bank of America’s concern going into the print was that the earnings beat alone would not be enough to resolve the deeper issue around how Nvidia spends its cash.
“The core worry is that NVDA’s seemingly open-ended checkbook to fund customers and suppliers is diluting earnings quality and potentially inflating higher-risk investments,” Arya wrote, while acknowledging the strategy serves both offensive and defensive purposes for the company, according to Investing.com.
His proposed fix borrows from a familiar playbook. Arya thinks Nvidia could follow the same path Apple took after 2012, lifting cash returned to shareholders from around 37% of free cash flow to 75% or more, a move he argues would create a reliable buyer for the stock and ease concerns about its AI ecosystem investments.
That framing lines up with a separate valuation call. Going into earnings, Nvidia shares traded at a roughly 40% to 50% discount to AI compute peers on an enterprise value to free cash flow basis, and a 31% to 36% discount to the S&P 500 on the same measure.
Arya has called this gap overstated, given the risks tied to Nvidia’s investments in companies including OpenAI and Anthropic.
Patrick/Getty Images
Bank of America raises AI data center forecast to $2.2 trillion
Bank of America’s bigger picture view rests on continued strength in AI infrastructure spending for years to come, even if the stocks themselves see a near-term dip.
Following Q2 earnings, the bank lifted its AI data center total addressable market forecast to roughly $2.2 trillion by 2030, with AI accelerators accounting for the largest share at roughly $1.2 trillion, followed by networking at around $316 billion and data center CPUs at approximately $110 billion, Investing.com reported.
That forecast keeps climbing. Bank of America had projected the same market would only reach $1.7 trillion by 2030 as of May, up from an even earlier estimate of $1.4 trillion, according to TheStreet. The bank has raised its own numbers twice in a matter of weeks, reflecting how fast the demand picture is evolving.
Flagship Nvidia GPUs are renting at near all-time highs, a sign that supply still has not caught up with AI-related demand. Memory stocks are also showing what that demand looks like in practice, with Micron among the biggest beneficiaries of the AI memory cycle.
What chip stock investors should watch next
Positioning is the risk most likely to amplify any near-term sell-off. Chip stocks are already running about 13% overweight relative to the S&P 500. A shift in sentiment could trigger outsized selling, even if the underlying fundamentals hold up.
Among the eight names Bank of America highlighted, Micron and Lam Research stand out as the bank’s favorite long-term picks, described as reliable picks-and-shovels plays across various phases of AI and tech inflections.
Equipment suppliers such as Lam Research stand to benefit as memory makers and foundries expand capacity to meet demand, a trend that holds regardless of which chipmaker ultimately wins the AI compute race.
Nvidia’s report delivered the beat the market expected, but the stock’s initial dip and recovery underlined Bank of America’s pre-earnings caution: execution is now a prerequisite, not a catalyst.
The bank’s concern about earnings quality and circular financing did not disappear with the strong quarter. Those questions carry into Q3 as Nvidia guides to $108 billion in revenue and the AI infrastructure buildout continues at full pace.
Related: JPMorgan sends another strong message to stock market investors