JPMorgan has lifted its forecasts for Google Cloud and Amazon Web Services while cutting its earnings outlook for Meta, saying AI monetization remains at an early stage but is beginning to show up in financial models through pricing power and partnership expansion at major cloud providers.
Google Cloud forecast raised on TPU contribution
In its latest report, JPMorgan raised its forecast for Google Cloud full-year revenue growth in 2027 to 107% from 70%. It also increased its 2028 forecast to 63% from 55%.
Nearer-term estimates were revised higher as well. The bank lifted its forecast for the third quarter of 2026 to 100% from 88%, and for the fourth quarter of 2026 to 105% from 90%.
JPMorgan said the main reason for the change was the inclusion of a larger contribution from Tensor Processing Unit, or TPU, sales in its model. TPUs are custom chips designed by Google for machine learning workloads, and the bank said rising TPU sales are becoming an important variable in assessing Google Cloud revenue.
The report said the 107% growth forecast reflects more TPU sales and should not be interpreted entirely as faster consumption of traditional cloud services. It also noted that differences in margins, delivery timing, and cash collection across business lines will affect how much of that revenue growth ultimately turns into profit.
At the group level, JPMorgan raised its 2027 U.S. GAAP earnings-per-share forecast for Alphabet by about 3% to $14.05, and its 2028 estimate by about 4% to $17.29. It lifted its December 2027 price target to $425 from $420 and kept an "Overweight" rating, based on roughly 25x expected 2028 earnings per share.
AWS seen reaching 40% growth in 2027
JPMorgan now expects AWS revenue growth to accelerate to 40% in 2027 and said that number could still prove conservative. The bank cited two main supports: a broader partnership with OpenAI and an additional deployment of 2 million GPUs across 2027 and 2028, along with stronger pricing power for AI compute.
According to the report, AWS increased reserved GPU capacity prices by 15% starting Oct. 7. JPMorgan said that marked the fourth consecutive quarter of price increases for related AI reserved capacity. It also said prices tied to H100/H200 and B200/B300 products have risen 40% to 75% since the beginning of the year.
The report described reserved capacity as a service arrangement in which customers book a set amount of compute capacity in advance.
JPMorgan also said AWS demand appears broad-based. Compute allocation is relatively diversified, with the largest customer accounting for only a single-digit percentage share. On that basis, the bank’s growth case for AWS combines incremental demand from large AI partnerships with support from a wider customer base.
It raised its forecasts for Amazon in 2027 and 2028 accordingly, increasing group sales estimates by about 1% to 2% and operating profit estimates by about 3% to 5%.
Meta earnings forecast cut on legal costs and capex pressure
Unlike the upward revisions for the two cloud businesses, JPMorgan’s changes for Meta were driven mainly by rising costs. The bank added a $10 billion legal charge to its model for the third quarter of 2026 and linked it to Meta’s settlement with 52 state attorneys general.
JPMorgan said that expense was not included in the company’s full-year expense guidance of $165 billion to $169 billion provided in its second-quarter filing.
As a result, the bank lowered its Meta GAAP earnings-per-share forecast for 2026 by about 10% to $27.99, and cut its 2027 estimate by about 2% to $31.08.
It also raised its 2027 GAAP total expense forecast for Meta to about $210 billion, up about 18% year over year. Excluding the $10 billion legal charge from the 2026 base, the report said 2027 expense growth would be about 25%.
Capital spending and cash flow were another focus. Under JPMorgan’s model, Meta’s 2027 capital expenditure is projected at about $242.7 billion, while free cash flow is estimated at negative $72.8 billion. The report said the negative figure indicates expected operating cash generation would not be enough to cover capital spending in the same period.
Fiverr faces weaker traffic and a long transition
JPMorgan took a more cautious view on Fiverr as well, drawing a contrast with cloud providers. While cloud companies are trying to absorb rising AI compute demand, services marketplaces are dealing with the effect of AI on traffic and transactions.
Citing SimilarWeb data, the bank said the year-over-year decline in visits to Fiverr’s U.S. website widened from 10% in the second quarter to 22% during the portion of third-quarter data already collected.
JPMorgan said changes such as AI summaries in Google Search are reducing the chances that users click on paid or organic search links. Even as Fiverr invests in generative engine optimization to improve its visibility in AI-generated answers, click-through rates from large language models remain below those of traditional search, putting pressure on lead generation.
Supply-side changes are also in progress. Fiverr is shifting from a marketplace centered on lower-value projects toward a platform for higher-end, higher-value work. JPMorgan said those higher-value services currently account for only 15% of gross merchandise value, and management expects the transition to take at least six quarters. Based on those factors, the bank forecast that Fiverr revenue in 2027 will decline by about 15%, the same drop it expects for 2026.