Focus on custom AI chips and cloud scale
On Thursday’s earnings call, Jassy highlighted Trainium chips as central to lowering the cost of AI workloads on AWS. The company has nearly sold out its latest version, Trainium3, through mid-year and is already fielding “strong interest” in Trainium4, which is still in development. Executives said the stepped-up spending will allow AWS to meet rising demand for generative AI services and maintain competitive performance.
Fourth-quarter operating results offered support for the expansion plan. Amazon reported revenue and operating income that exceeded analyst forecasts, while AWS growth accelerated to 24 percent year over year, the fastest rate in 13 quarters. The cloud unit’s backlog reached $244 billion, a 40 percent increase from a year earlier and 22 percent sequentially. Management added that AWS margins remain healthy, suggesting existing capacity is being used efficiently.
Cash flow squeeze raises flags
Despite stronger top-line trends, the new capital plan sharply compresses expected liquidity. FactSet previously pegged Amazon’s 2026 free cash flow at about $37 billion. With projected capex now at $200 billion—roughly $50 billion above prior assumptions—that surplus essentially disappears. Free cash flow is calculated as operating cash flow less capital expenditures, a metric closely watched for companies balancing growth initiatives with shareholder returns.
Amazon is not alone in boosting outlays. Alphabet and Meta Platforms have also signaled above-consensus budgets tied to AI infrastructure. However, several analysts argue investors can more readily quantify near-term monetization at those firms. The perceived opacity around Amazon’s payoff timeline prompted multiple target reductions on Friday:
- Wedbush: target trimmed to $300 from $340
- Cantor Fitzgerald: target cut to $250 from $260
- D.A. Davidson: target slashed to $175 from $300 and rating lowered to neutral
D.A. Davidson analysts wrote that AWS appears to be trailing rivals after Google Cloud posted 48 percent growth and Microsoft’s Azure rose 39 percent in their latest quarters. They also warned Amazon’s retail unit could face a “structural disadvantage” without deeper integration of third-party AI platforms such as ChatGPT or Gemini. Management instead promoted an in-house shopping assistant called Rufus and left broader partnership timelines open-ended.
Imagem: Internet
Cramer’s Investing Club adjusts, but stays positive
Cramer’s CNBC Investing Club maintained its buy-equivalent “1” rating on Amazon after the report but lowered its internal price objective to $250 from $275. The stock has fallen 12 percent over the past 12 months, and Cramer cautioned that a slide into the $190s cannot be ruled out given investor unease.
Nvidia Chief Executive Jensen Huang defended the record-level industry spending during a separate CNBC interview Friday, calling the budgets “appropriate and sustainable.” Nvidia, whose GPUs dominate the AI training market, stands to benefit from orders placed by Amazon, Alphabet and other hyperscale customers. Cramer’s club simultaneously upgraded Broadcom, another chip supplier, to a buy-equivalent rating while keeping Nvidia as a hold. Huang said he is not concerned about custom chips such as Trainium, asserting that no single company can match Nvidia’s scale and pace of innovation.
Competitive landscape in the cloud
While AWS remains the largest commercial cloud provider, the sequential pickup to 24 percent growth comes against faster expansion at Google Cloud and Azure. Analysts attribute part of the divergence to AWS’s larger revenue base, noting that the “law of large numbers” naturally depresses percentage gains. Even so, some firms view relative momentum as an early indicator that competitors are narrowing the capabilities gap.
The question now facing investors is whether Amazon’s unprecedented capital plan can translate into outsized returns before debt or dilution pressures emerge. Management argues that building enough capacity—both in data centers and specialized silicon—will safeguard AWS’s market share and position the company to supply enterprise AI demand well into the next decade.
For the broader market, Amazon’s move reinforces a theme of escalating tech infrastructure bets, underscoring how central AI workloads have become to the strategies of the largest U.S. corporations. The company’s next earnings cycle will offer the first data point on how quickly that spending begins to flow through to revenue and margins.
Crédito da imagem: CNBC
