Bernstein is cautious about strong demand for many ultra-micro computers (SMCI.US) AI servers, sending a positive signal to Dell and HPE

The Zhitong Finance App learned that after AI server manufacturer Ultra Micro Computer (SMCI.US) released results and guidelines that exceeded expectations, Bernstein raised the target price of the stock from $37 to $42, maintaining the “same as the general market” rating.

Bernstein analyst Mark Newman pointed out in the report that ultra-microcomputers are expected to have revenue of 65 billion to 72 billion US dollars for the 2027 fiscal year starting in July, which is far higher than the general market forecast of 54.4 billion US dollars. Meanwhile, revenue for the first quarter of fiscal year 2027 is estimated at US$14.5 billion to US$15.5 billion, which also exceeds market expectations of US$12 billion; earnings per share (EPS) guidance is US$1.01 to 1.10, higher than market expectations of US$0.74; gross margin guidance is 10.4% to 10.8%, which is also better than market expectations of 9.0%.

“Overall, we raised our expectations slightly, mainly because revenue growth and product portfolio optimization will drive up gross margins, but equity dilution over the next few quarters will partially offset this impact,” Newman said. Although Ultramicrocomputer has failed to meet earnings expectations per share 7 times in the past 10 quarters, and revenue falls short of expectations from time to time, “we still have some confidence in it.”

Newman further analyzed that the strong performance of ultra-microcomputers provided a “favorable reference” for Dell Technologies (DELL.US) and HPE (HPE.US). But at the same time, he pointed out that HPE Technology's situation is more complicated.

“The excessive growth of ultra-microcomputers may reflect a further increase in their market share, which may have come at the expense of HPE Technology,” Newman added. He noted that compared to Dell, HPE Technology's growth rate in the traditional server sector was slow in the last quarter, and the AI server business was relatively weak.

The target price is as high as $46 and as low as $15. Wall Street analysts are polarized in their views on ultra-microcomputers

Overall, Wall Street analysts are still divided on ultra-microcomputers. Some analysts are optimistic about the growth potential brought about by AI infrastructure demand, but others warn that the high share of the GPU server business may continue to suppress profit margins and free cash flow.

Needham maintains the ultra-micro computer's “buy” rating and a target price of $46. The bank pointed out that the company's revenue for the fourth fiscal quarter fell short of expectations, mainly due to short-term customer delays in electricity, cooling and network infrastructure, rather than a marked deterioration in demand for AI server terminals. The overall gross margin of the company is expected to gradually improve from the level of about 10.6% in the first quarter of fiscal year 2027.

Wedbush raised its target price for ultra-microcomputers from $34 to $40, while maintaining a “neutral” rating. Expectations for the future growth of ultra-microcomputers currently seem reasonable, and management expects gross margin to fall back to around 10%. This target is also relatively easy to achieve. At the same time, the bank believes that the strong backlog of orders and prospects for ultra-micro computers indicates that demand for AI servers and related infrastructure is still strong, which is a positive sign for Dell Technology and HPE Technology as well.

Susquehanna, on the other hand, maintained a “negative” rating for the ultra-micro computer and a target price of $15. The agency believes that the increased contribution of CPU servers and traditional enterprise systems to revenue is beneficial to improving profitability to a certain extent, because the gross margin of these products is generally higher than that of GPU servers. However, to maintain the current rapid growth rate, ultra-micro computers still need to maintain a high level of exposure to GPU-driven AI infrastructure deployment, which will continue to depress the company's overall gross profit margin and may lead to continued free cash outflows.

According to Tipranks data, out of 14 analysts covering ultra-microcomputers, 3 gave a “buy” rating, 9 gave a “hold” rating, 2 gave a “sell” rating, and the consensus rating was “hold”. The average target price was 40.5 US dollars, which is 3.4% higher than the latest closing price.