Stifel downgrades Microsoft to hold and says it's time for a break.
Microsoft's (MSFT) stock saw a rare downgrade on Wall Street this week. Analyst Brad Reback of Stifel downgraded the stock to hold from buy, warning that market expectations for fiscal and calendar year 2027 are overly optimistic amid cloud supply constraints, increased investment, and intensifying competition in artificial intelligence.
Rebak said it was time to take a break, lowering the company's target price from $540 to $392.
He pointed to Azure's ongoing capacity constraints as a major downside factor. The analyst stated that, given Azure's well-documented supply issues, coupled with Google's strong performance in GCP/Gemini and Anthropic's growing momentum, they believe near-term acceleration for Azure is unlikely. Competitive pressure from rivals, including Google (GOOGL) and Anthropic, is reshaping the dynamics of cloud growth.
Rebak also expects revenue recognition to return to normal after fiscal year 2026 (FY26) benefited from several overlapping product cycles, reducing near-term upside.
Meanwhile, spending is set to rise sharply. Stifel raised its capital expenditure estimate for fiscal year 2027 to around $200 billion (approximately 40% growth), significantly higher than the market expectation of roughly $160 billion. Rebak said the increased investment will squeeze profitability, lowering the company's forecast for its fiscal year 2027 gross profit margin to around 63%, compared to a consensus estimate of around 67%.
Operationally, the analyst believes that Microsoft is entering a new, albeit still effective, phase of spending as it builds and markets its own AI platforms, which are likely to be a constraint on raising the operating margin (OM).
While Stifel remains positive about Microsoft's long-term position, Reback cautioned that the near-term outlook looks more uncertain, arguing that the stock is unlikely to be revalued until capital spending for Azure growth slows or the cloud business achieves significant acceleration.
DA Davidson downgrades Amazon as it loses its cloud leadership
DA Davidson downgraded Amazon (AMZN) to neutral from buy on Friday, with analyst Gil Luria saying the company is losing leadership in cloud computing and showing early signs of a strategic disadvantage in an increasingly AI-driven retail environment.
The broker also lowered the target price to $175, arguing that Amazon is now scrambling to catch up by ramping up investment.
Luria said that Amazon Web Services (AWS) still lags behind Microsoft Azure and Google Cloud. While AWS saw 24% year-over-year growth, the company noted that Google Cloud accelerated to 48% growth and Azure grew by 39%, limited by customization.
The analyst pointed to Amazon's lack of a sophisticated in-house AI lab, unlike Google, and the absence of a core partnership with OpenAI similar to Microsoft's, as factors driving a shift in customer preferences.
Luria warned that falling behind would necessitate even greater spending. He pointed to over $200 billion in capital expenditures and said Amazon might have no choice but to proceed with its $50 billion investment in OpenAI to remain competitive in advanced models.
He also raised concerns about Amazon's ability to adapt to a new conversational internet dominated by Gemini and ChatGPT. Without direct integrations, Amazon faces a structural vulnerability, as merchants integrated with leading AI models could gain a significant advantage in traffic and advertising, Luria said.
Wolf: Tesla's autonomous taxi potential is enormous, but the pressure is closer now.
Earlier this week, Wolf Research said Tesla (TSLA) could turn its autonomous taxi platform into a massive long-term revenue engine, predicting the business could reach $250 billion in annual revenue by 2035 as self-driving adoption accelerates.
Analyst Emanuel Rosner said 2026 is shaping up to be a catalyst-rich year for the stock, as investors track progress through the expansion of robotic taxis, the production of Optimus and the rollout of its fully autonomous, unsupervised driving software.
If successful, the long-term return on investment (ROI) could be very attractive, Rosner wrote.
Wolf’s top-down model assumes a 30% penetration rate for self-driving vehicles, a 50% market share for Tesla, and pricing at $1 per mile, which the company says could underpin a share value of around $2.75 trillion — or nearly $900 billion at a discount, which equates to more than $250 per share.
Optimus and the FSD license should support further gains, Rosner added.
In the near term, however, the analyst remains cautious about the fundamentals and places his forecasts below the consensus earnings forecasts for 2026 and 2027. Rosner pointed to margin pressure from higher input costs, pricing dynamics and shifts in Tesla's FSD monetization model.
The large investment in AI initiatives – including robot taxis and Optimus – is also expected to impact profits next year as expansion costs accumulate.
Beyond self-driving, Rosner pointed to the strong momentum in Tesla's energy storage sector, predicting a sharp rise in deployments as new capacity comes online, even with competition and tariffs squeezing margins in the short term.
Overall, while we have concerns about near-term earnings, we remain tactically positive, with a steady stream of catalysts ahead, Rosner said.
Analyst urges investors to buy the dip in AMD
Advanced Micro Devices (AMD) remains a long-term winner despite a turbulent fourth quarter, according to Trust Securities, which told clients to buy double as earnings strength continues to build.
The chip company has lost more than 14% over the past five days, reaching its lowest level since October 2025.
Analyst William Stein said Advanced Micro Devices is doubling earnings at a compound annual growth rate of ~45% through calendar year 2030 and is trading at ~11 times its earnings per share for calendar year 2030, adding that it is still a buy.
Advanced Micro Devices' fourth-quarter results exceeded expectations, and its first-quarter guidance was higher, although much of the gains stemmed from unusual dynamics related to China. Stein noted that 65% of revenue and all of the superior earnings per share came from the sale of a discounted Chinese SKU.
Even with this distortion, he said Advanced Micro Devices has confirmed a 60% compound annual growth rate for its data center and a 35% compound annual growth rate for total sales, which he believes translates to more than $20 in earnings per share in 2030.
Stein said the core of the positive case remains strong customer engagement, with industry checks bolstering momentum across Advanced Micro Devices' data center and AI roadmap.
Revenue for the quarter reached approximately $10.3 billion, exceeding the consensus estimate by about 6%, driven by a nearly 9% increase in both the data center and customer segments. Stein cited the accelerated deployment of Instinct MI350 GPUs, the strong adoption of 5th-generation Epyc processors, and management's view that data center revenue could grow by more than 60% annually over the next 3-5 years.
Trust raised its 2027 earnings forecast to $10.11 and increased its price target to $283.
Buy the weak as the long-term growth message outweighs the flaws in the fourth quarter, Stein said.
Jefferies: The next stop for Palantir stock is a decline.
Palantir Technologies (PLTR) faces further downside risks despite this year's sharp decline, Jefferies analyst Brent Thiel said in a note on Friday, adding that pressure on the stock's valuation is likely to continue.
The stock is down about 27% year-to-date, but Thiel argues that it remains expensive relative to the broader software sector even after a significant contraction in trading multiples. He notes that Palantir traded at as much as 73 times future revenue in November before falling to around 31 times—still roughly double the price of its most expensive large-cap software peer.
We are offering a recommendation on the assessment, not on the fundamentals, Thiel wrote in a note on Friday.
We acknowledge that PLTR's fundamentals have improved rapidly, and we believe its competitive differentiation has only strengthened as the overall available market has expanded. However, the downside risks to trading valuation outweigh the upside potential from improved fundamentals.
He said this premium makes the stock highly sensitive to shifts in investor sentiment, particularly regarding AI enthusiasm and broader software sector trends. Thiel cautioned that the current multiple makes PLTR especially vulnerable to narrative changes, including concerns about slowing growth or waning optimism about AI.
We believe that declining sentiment in the software industry could accelerate PLTR's return to more sustainable valuation levels, he added.
This caution comes despite a strong fourth-quarter performance, with Palantir reporting accelerated revenue growth, rising U.S. commercial demand, and expanding operating margins, along with initial 2026 guidance indicating continued momentum.
However, the results failed to justify the high valuation, with the stock falling nearly 21% after the earnings announcement. Jefferies reaffirmed its underperformance rating and maintained its $70 price target per share.
