September looms as a risk month for stocks, Yardeni says
Investing.com -- Here are the biggest analyst moves in the area of artificial intelligence (AI) for this week.
InvestingPro subscribers always get first dibs on market-moving AI analyst comments. Upgrade today!
AMD upgraded at Truist to Buy on AI traction
Advanced Micro Devices (NASDAQ:AMD) received an upgrade at Truist Securities this week, to Buy from Hold, with the broker also lifting its price target to $213 from $173. The firm pointed to stronger industry feedback on the company’s data center and AI momentum.
Truist analysts said “industry contact feedback turns constructive on AMD’s DC/AI traction,” noting that hyperscale customers are increasingly treating AMD as a true partner “expressing true interest in deploying AMD at scale.”
Analysts said this marks a shift from past years, when AMD was seen mainly as a “price check” against Nvidia (NASDAQ:NVDA).
The bank acknowledged its long-standing view that Nvidia’s GPU dominance, anchored by CUDA, left little room for a second supplier. But sentiment has turned.
“Recently, they have told us that hyperscale customers are working with AMD as a potential partner rather than simply as a ‘price check’ to NVDA. This change in messaging from the field is the basis of our upgrade,” Truist’s team wrote.
Analysts compared the development to AMD’s rise in server CPUs, where its share climbed from less than 1% in 2018 to about 21% after the launch of its “Rome” product as Intel faltered. For GPUs, they don’t expect Nvidia to repeat Intel’s mistakes but now see AMD achieving a sustainable 10% market share.
Truist raised its earnings forecasts, putting calendar year 2027 (CY27) EPS at $7.89, and highlighted AMD’s MI355 chip, introduced in June, as a catalyst for growth in the coming quarters.
AI stocks’ valuations getting closer to dotcom levels
UBS is warning that valuations in artificial intelligence stocks are approaching levels last seen during the dotcom boom, raising concerns over sustainability despite record investment by major U.S. technology firms.
The bank noted that the U.S. tech sector is trading at an aggregate HOLT Economic price-to-earnings (P/E) above 35 times, a level comparable to the post-dotcom peak.
HOLT Economic is UBS’s proprietary valuation and performance model, which suggests that much of today’s market value is tied to expectations of future cash flows rather than current earnings.
That leaves “little room for cash flow disappointments,” said Michel Lerner, head of the HOLT analytical service at UBS, citing uncertainties around massive capex returns, data center energy limits, and intensifying competition from China.
AI has become a dominant theme in corporate earnings, with one in four company releases now mentioning the technology, Lerner said.
Spending has surged, with the largest U.S. tech companies expected to commit $350 billion to capex this year—exceeding the combined annual capex of all listed energy and utilities companies in the U.S. and Europe.
Lerner added that Apple (NASDAQ:AAPL), Nvidia and Broadcom (NASDAQ:AVGO), together with hyperscalers, spent more on R&D in 2024 than all listed European equities combined. Collectively, these firms are projected to generate 37% of U.S. economic profit in 2025, more than six times Europe’s total.
Still, he cautioned that “many of the use cases are premised on future, rather than current revenue opportunities,” highlighting comments from OpenAI CEO Sam Altman, who has acknowledged the sector may be in a bubble.
Also, a recent MIT study found 95% of generative AI pilots are failing to deliver immediate revenue growth, underscoring the gap between hype and realization.
Lerner warned that cash flow resilience for Big Tech could weaken, with consensus pointing to declines in Cash Flow Return on Investment for Amazon (NASDAQ:AMZN), Meta (NASDAQ:META), Microsoft (NASDAQ:MSFT) and Alphabet (NASDAQ:GOOGL) over the next two years.
Power supply constraints also present risks, as hyperscaler investment far outpaces utilities’ capacity to deliver.
In light of these risks, UBS advised investors to diversify exposure, pointing to opportunities in non-U.S. quality growth stocks, global names within secular growth ETFs outside U.S. AI, and sectors like real estate, utilities, energy, communication services and staples, which historically show low correlation to U.S. tech performance.
BofA cuts Marvell to Neutral on softer AI growth outlook
Bank of America downgraded Marvell Technology Inc (NASDAQ:MRVL) to Neutral from Buy on Friday, citing softer confidence in the chipmaker’s AI growth outlook through 2026. The price target was also cut to $78 from $90.
The bank now values the stock at 24 times CY26 earnings, down slightly from 25 times but still consistent with Marvell’s historical multiples.
BofA analysts said they “did not hear the same level of confidence/visibility about MRVL’s AI growth prospects in the near/medium term,” pointing to “incrementally higher uncertainty” around Microsoft’s Maia project and Marvell’s role in Amazon’s next-generation 3nm chip program.
BofA’s CY26 data center growth forecast was cut to mid-teens from a prior 23-25%. That translates into “a $100mn lower quarterly run-rate, and hence a $400mn drag on projected FY27/CY26 sales.”
The downgrade follows weaker recent performance. Marvell slightly missed second-quarter data center guidance, delivered 3% sequential growth, and projected flat third-quarter revenue compared with consensus for a 5% increase.
Even so, BofA highlighted the chipmaker’s valuation support and financial flexibility. “
At after-hours price, MRVL is trading at 20x-21x PE, which tends to provide a floor to compute stocks with strong breadth of IP,” the analysts said in a note earlier this week, noting $2.5 billion in proceeds from its auto unit sale could be used for buybacks or acquisitions.
Wall Street starts coverage on ’key Edge AI beneficiary’
Ambiq Micro (NYSE:AMBQ) has drawn a mix of optimism and caution from Wall Street as analysts initiated coverage on the low-power chipmaker following its July IPO.
Stifel started with a Buy rating and a $45 target, calling Ambiq’s proprietary “SPOT” platform a key differentiator.
“Based on SPOT’s unique ability to deliver materially better compute performance/watt at the transistor level, we believe Ambiq is positioned to become a key beneficiary of the rapidly-emerging Edge AI market opportunity,” analysts led by Tore Svanberg wrote.
The brokerage expects a revenue inflection in 2026, supported by the Atomiq platform and a broader customer base.
Ambiq shares jumped more than 60% in their debut, closing at $38.53 and valuing the company around $657 million. After some early volatility, the stock has steadied near those levels, last trading at $39.71.
Other banks were more cautious. UBS began coverage at Neutral with a $40 target, pointing to Ambiq’s strong positioning in wearables but warning that “profitability will take time with Ambiq only turning profitable in 2028E and even this keyed to the launch of Atomiq.”
UBS highlighted execution risks and customer concentration, though it noted relationships with Google, Garmin, and WHOOP.
Bank of America also initiated at Neutral with a $42 target, describing Ambiq as a “low-power edge-AI specialist 3+ years from profit.”
The bank said the company has shipped over 270 million devices but remains reliant on a few key customers. It projects gross margins climbing to about 54% by 2028, the same year Ambiq is expected to achieve profitability.
AI will not be the ’death of software’, RBC says
RBC Capital Markets pushed back against the notion that AI will make traditional software obsolete, calling the “death of software” narrative overstated.
Software stocks have faced pressure recently, though RBC argued much of it reflects misplaced concerns. The IGV software ETF (NYSE:IGV) is up around 8% this year, but gains have been driven largely by Microsoft, Oracle (NYSE:ORCL) and Palantir.
Excluding those names, the ETF would be down by double-digits, the bank noted.
RBC analysts said the market is split between two views: one claiming “all software will be replaced by agents and multi-agentic systems” and another that believes incumbents’ “valuable data and distribution” will protect them and enable AI monetization.
Analysts led by Rishi Jaluria outlined a middle ground, writing that “AI will benefit some, but not all incumbents, while also creating net-new scaled companies and accelerating AI-focused M&A.”
The team questioned whether incumbents’ data advantage is as strong as often claimed, pointing to ownership uncertainty, data commoditization, and the growing importance of real-time data.
They also flagged disintermediation risks as AI-native vendors augment existing platforms before competing directly.
Lower entry barriers, such as “vibe coding,” could intensify competition, but analysts said this may also expand software budgets by fostering more innovation.
M&A is expected to be a key strategy for traditional vendors, though analysts cautioned that monetization of AI may take longer, with broad adoption across enterprises potentially not arriving until 2028 or later.
“In the interim, we may start to see indirect monetization of AI, whether that shows up in greater consumption, higher engagement, improved win rates, or better gross retention rates,” they said.
The bank identified Microsoft, Intuit (NASDAQ:INTU), HubSpot (NYSE:HUBS), MongoDB (NASDAQ:MDB) and Pegasystems (NASDAQ:PEGA) as best positioned to “cross the chasm” in a post-AI world, while being more cautious on Salesforce Inc (NYSE:CRM) and ZoomInfo (NASDAQ:GTM).
It added that recent pullbacks in names such as Dynatrace (NYSE:DT), HubSpot, MongoDB, ServiceNow (NYSE:NOW) and Snowflake (NYSE:SNOW) may be overdone.