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!
Get premium news and insight, AI stock picks, and deep research tools by upgrading to InvestingPro - get 55% off today
Morgan Stanley lifts Nvidia, Broadcom targets as AI demand tightens supply
Earlier this week, Morgan Stanley hiked its price targets on Nvidia and Broadcom after meetings in Asia and the U.S. pointed to broad-based AI strength and tightening supply across key parts of the semiconductor chain.
The bank said both companies look positioned for a meaningful ramp next year as customers struggle to secure enough hardware to support accelerating workloads.
Analysts led by Joseph Moore said they “continue to see Nvidia maintaining dominant market share,” arguing that competitive concerns remain “overstated” and that customers’ primary worry heading into next year is securing supply, especially for the upcoming Vera Rubin platform.
Nvidia’s data-center revenue outlook remains constrained by limited availability through 2026, and Morgan Stanley now sees the company tracking closer to the revenue path management outlined earlier this year. The price target rises to $250 from $235.
“We still are below the "$500 bn in 5 quarters" voiced by the CEO at the GTC event, but clearly the situation is strong,” Moore wrote.
The firm’s checks also support stronger-than-expected demand for Google’s TPU supply chain, which Broadcom designs and sells. Suppliers across analog, memory and ODMs pointed to upward revisions in TPU builds, with the largest increases projected for 2027.
“We remain Overweight (OW) AVGO and we are encouraged to see this - and raising numbers - but we would introduce a few caveats,” Moore said. He added that Google is advancing a homegrown TPU variant with Mediatek—potentially a long-term risk but not one that materially shifts the base case.
Reflecting higher ASIC expectations for 2026 and 2027, Morgan Stanley raised its Broadcom target to $443 from $409.
Moore said the surge in AI demand is straining back-end capacity such as CoWoS and high-bandwidth memory, while front-end wafers across 3-, 4- and 5-nanometer nodes are also tight—a setup that reinforces a bullish backdrop as the industry requires more capacity additions.
Memory markets are firming rapidly as well, with cloud buyers showing what the analysts describe as a “gold rush purchasing mentality.”
AI winter risk grows as valuations stretch and adoption cools, strategist warns
An “AI winter” is becoming a more plausible outcome for the industry over the next several years, BCA Research warned in a note this week, as the sector’s breakneck investment cycle enters a more fragile stage and valuation assumptions run far ahead of what is likely to be delivered.
“Over a three-to-five-year time horizon, the balance of probabilities points to the emergence of an ‘AI Winter,’ likely beginning over the next one-to-three years,” BCA chief strategist Jonathan LaBerge wrote.
Such a phase would entail a cooling in AI-related capex and data-center buildouts, alongside “a meaningful correction in tech/growth stock prices.”
BCA lays out three structural paths for AI, assigning only a 5% probability to a genuine artificial general intelligence breakthrough that would validate the market’s most bullish expectations.
Instead, it sees an 80% chance that AI ultimately generates “moderate macro-level productivity gains,” lifting output by roughly 0.4–0.5% a year—positive for the economy, but too modest to support the profit and cash-flow trajectory embedded in leading tech names.
A more adverse scenario, involving a major data-center investment misfire and a productivity bust, carries a 15% probability.
LaBerge estimates that U.S. equities have already priced in far greater AI-driven upside. He calculates that $9–$12 trillion in market gains since late 2022 cannot be explained by earnings trends or interest rates alone.
"It is not enough for artificial intelligence to be productivity enhancing: It needs to boost productivity growth / corporate profits very significantly for current pricing of the overall equity market to be justified," he wrote.
The strategist also flags early signs of stress in adoption trends, noting that business uptake of AI tools has cooled while capital spending continues to accelerate. At the same time, the industry’s debt dependence is deepening as hyperscalers and infrastructure providers issue large amounts of bonds to fund data-center construction.
The capital-intensive nature of today’s AI wave—demanding constant investment in chips, energy, networking and cooling—adds another layer of risk.
BCA advises investors to prepare for a slower reset in expectations, recommending an underweight stance on tech-linked names until excesses unwind. It sees adopters better positioned than enablers and monetizers, with financials, pharmaceuticals, biotechnology, life sciences and defense likely to benefit most from practical AI deployment.
BofA names ASML a top semis pick for 2026, raises target
Bank of America has named ASML one of its top semiconductor picks for 2026, saying the company is entering a multi-year upswing supported by rising lithography intensity, stronger earnings momentum and a substantial improvement in free cash flow.
The bank reaffirmed its Buy rating and raised its price objective to €1,158 from €986, with analysts led by Didier Scemama pointing to 2027 as a clear turning point as ASML captures a greater share of customer spending and benefits from a more favorable product mix.
ASML features on BofA’s “25 stocks for 2026” roster and is included in its Europe 1 list of top ideas.
The Wall Street firm’s expected re-rating is anchored in three forces: higher lithography intensity in memory as DRAM makers add EUV layers, gross-margin expansion of roughly 150 basis points that supports about 30% earnings growth, and free cash flow doubling to €14 billion.
Lithography intensity remains resilient and is projected to rise toward an estimated 26% by 2028 as ASML gains a growing share of spending.
BofA also sees several long-standing investor worries receding. The analysts argue that customer concentration risk is easing as Samsung regains competitiveness, Micron accelerates EUV adoption, Intel stabilizes and AI chipmakers move deeper into advanced nodes.
They expect China’s revenue contribution to settle in the low-to-mid-20% range, helping shift sentiment from a “WFE minus” to a “WFE plus” narrative. This transition is set to be reinforced by a projected 5-percentage-point expansion in gross margins by 2030 and an estimated 18% earnings CAGR over the next five years.
TD Cowen puts AMD on ’Best Ideas 2026’ list
TD Cowen has placed AMD on its “Best Ideas 2026” list, saying the company is approaching a major inflection as its Helios rack-scale platform prepares to debut.
The broker argues the launch will “ignite AMD’s AI business” and set up a stronger multi-year trajectory for the chipmaker. Analyst Joshua Buchalter reiterated a Buy rating and a $290 target, calling the recent pullback an “attractive entry point.”
Buchalter says “AI compute spending will prove durable and AMD has cemented itself as a winner,” despite growing competitive scrutiny and debate around whether the industry is entering an “AI bubble.”
He believes that AMD “has garnered more bearish sentiment than deserved and than peers,” a dynamic the firm expects to reverse as Helios and the MI450 accelerator ramp from mid-2026.
The rollout is central to their bullish stance. TD Cowen says the platform “will mark a key inflection in AMD’s story,” with fourth-quarter 2026 EPS expected to reach “a >$10 run-rate, or up ~2x… Y/Y and Q/Q.”
TD Cowen models $89 billion in Instinct sales by 2030, implying a 67% compound annual growth rate, while noting that even this estimate “falls below AMD’s guided >80% CAGR.”
On concerns around AMD’s exposure to OpenAI, Buchalter says the stock is being “unfairly punished,” noting the company’s expected revenue tied to OpenAI in 2027 is “roughly the same” as peers.
He also highlights the chipmaker’s overlooked server and PC segments, arguing they “remain an important part of AMD’s core business” and stand to benefit as AI adoption increases the need for real-time inference.
Summit lifts Marvell to Buy, sees data-center growth inflection ahead
Marvell received an upgrade on Tuesday at Summit Insights. The brokerage said a clear turning point is approaching for the company’s data-center business, setting the stage for a sharp acceleration in growth over the next two fiscal years.
Summit raised its rating to Buy, arguing that the coming topline expansion should more than offset near-term softness and ongoing concerns around product-mix pressure.
Marvell expects its data-center segment to grow about 25% year over year in fiscal 2027, followed by roughly 40% in fiscal 2028. Summit analyst Kinngai Chan said the company is positioned to benefit from higher port counts, rising bandwidth needs and a strengthening electro-optical portfolio.
He also noted that he “continues to see risk of lumpy shipments for its AI ASIC programs.” Even so, Chan added that “our industry checks, however, lead us to believe that AI capex will only accelerate in the next 2-years and both AI accelerators and interconnectivity will be the largest beneficiary of this capex spend.
For the January quarter, Marvell forecasts 6% sequential revenue growth to $2.20 billion. Data-center revenue is expected to increase about 9%, driven by recovering custom ASIC demand and ongoing strength in electro-optics.
Gross margin is projected to hold in the 58.5% to 59.5% range amid a less favorable mix.
