Top Technology Themes for 2025

By Daniel Morgan, Synovus Trust Senior Portfolio Manager
Synovus Trust Company, N.A.

• IaaS Cloud Providers Growth and CapEx Spend Remains Vigorous — Given strong demand for artificial intelligence (AI) model training and inference, the largest public cloud providers (AWS, Azure and Alphabet) should continue to experience an acceleration in Year-over-Year (YoY) growth rates. Growth rates for IaaS cloud providers have accelerated in the recent calendar year 3Q24 — Amazon’s AWS being up 19%, Microsoft’s Azure 34% and Google’s GCP also up by 35% — compared to a year ago. Microsoft’s Azure exhibited a 12% YoY growth contribution directly from AI service in the most recent 1Q25, implying that non-AI workloads were up 22% YoY. Amazon has launched several AI products in its cloud and e-commerce businesses, and it’s also expected to announce a new version of its Alexa voice assistant, powered by generative AI. Alphabet reported blowout cloud revenue at $11.35 billion in 3Q24, up nearly 35% from the $8.41 billion a year ago. The company attributed its strong cloud results to its AI offerings, which include subscriptions for enterprise customers. Of the top four hyperscalers – Amazon, Microsoft, Meta and Alphabet – they are collectively estimated to spend $240 billion (up 60% YoY) in CapEx in 2024 to build out its AI presence.


• AI Drives Energy Consumption – AI is projected to significantly drive energy consumption due to the massive computing power required to run its complex algorithms, primarily within data centers. This will result in a substantial increase in electricity demand as AI usage expands across various industries. Hyperscaler data centers are housed in huge physical structures designed to process the vast amounts of data required to support digital technologies, including AI workloads. This will benefit Energy sector companies in the data center build road map, such as Entergy, Southern Company and Duke Energy Corp. Meta recently announced it will build a $10 billion AI data center in northeast Louisiana with Entergy as the sole energy provider. This trend is also being felt in the manufactures that provide systems for the large utility companies. For example, engine provider Cummin’s Power System unit has seen a surge in demand over the past three years and is expected to post revenues of $6.196 billion (up 9.3% YoY) in FY24. Cummins recently said it expects sales from its power-generation business to increase by 10% to 15% this year, up from its previously forecast 5% to 10% growth.


• PC Markets Rebound as AI Server Demand Remains Robust – Overall personal computer (PC) shipments were down 2.4% YoY in 3Q24. However, according to IDC, PC unit growth for 2025 is projected to be around 4.3%, with the primary driver being a significant refresh cycle among commercial users due to the end of Windows 10 support, pushing them to upgrade to Windows 11. Further, it’s an acceleration in demand driven by AI PCs. Traditional server makers (Dell, Hewlett Packard) are riding the AI data center service providers’ expansion wave. In the recent 3Q25, Dell's servers and networking revenue came in at $7.4 billion, up 58% YoY, with demand growth across AI and traditional servers. During HPE’s most recent 4Q25 results, server revenue was $4.7 billion, up 32% from the prior year period, with a 11.6% operating profit margin.


• Semiconductor Growth Likely to Remain Solid – Driven by the ongoing ramp of infrastructure investment for generative AI and a recovery in memory prices, the World Semiconductor Trade Statistics (WSTS) organization projects global sales to reach $697.2 billion (up 11.2%) in 2025. Different sectors within the chip space — cloud/data center/AI, telco/enterprise infrastructure, PC/laptops, smartphones and automotive/industrial — are experiencing different paces of recovery. While consumer-oriented markets — PCs (Intel, Advanced Micro Devices and Micron Technology) and smartphones (Qorvo and Qualcomm) — are weaker in the near term, anticipate a return to growth in the second half of fiscal year. The Automotive and Industrial (Texas Instruments and NXP) segments have weakened. For example, NXP sells a wide portfolio of semiconductors to customers in a host of different industries, including automotive, reported a disappointing 3Q24 as profits fell 8.8% YoY. The AI space has experienced the fastest growth as companies like Advanced Micro Devices, Broadcom, Intel and Marvell race to introduce chips to compete against market leader Nvidia. The chip sector does not appear to be overvalued as the Philadelphia Semiconductor Index (SOX) currently trades at 5.3x Price/Book (P/B). The P/B ratio is nowhere close to previous periods of “irrational enthusiasm” like during the pre “Dot.com bust” summer of 2000 when it traded above 8.0x P/B. It appears that the chip sector is in the beginning stages of a broad-based rebound, led by AI with PC Computing, Smartphones, Auto and Industrial playing catch-up.


• The Hot Spot in Chips is Still AI Accelerators – Nvidia continues to be in the pole position in the AI chip race with an estimated 70% MS. But other traditional chip companies are in the pack: Advanced Micro Devices, Broadcom, Intel and Marvell. Further, data center/search companies like Amazon, Microsoft, Alphabet and Meta are all producing their own proprietary AI chips. Nvidia today accounts for more than 70% of AI semiconductors sales, with Amazon (Trainium 3/ Inferencia 3), Microsoft (Colbalt 100) and Meta (MTIA v1) all producing their own AI chips. Amazon's Trainium and Inferentia chips have seen strong adoption. Amazon has a proven success with custom silicon as its Graviton chips (custom arm-based CPUs) have accounted for more than 50% of new compute instances on AWS over the past two years. Broadcom, which has three existing known ASIC customers (Google’s TensorFlow TPU v6, Meta, ByteDance) and reportedly two more prospects (OpenAI and Apple), closed out FY2024 with AI product revenue at a $14-15 billion annual run-rate, more than doubling YoY. Further, Apple is even rumored to be working with Broadcom on a networking technology chip, code-named Baltra, which is expected to be ready for mass production by 2026 (intended for internal use only). While Advanced Micro Devices upcoming MI325X looks promising for AI inferencing workloads due to large amounts of integrated Memory (192GB), Nvidia management pointed out that the company’s Grace Hopper chip supports nearly 600GB of Memory. Marvell is expected to be major beneficiary of the aggressive spending on generative AI by its cloud customers; Marvell (most well-known for its close partnership with AWS) recently indicated that its AI revenues will likely exceed $1.5 billion this year and $2.5 billion next year – an implied growth rate of 67%. Marvell’s revenues are expected to be driven by strong demand pull for Marvell’s 800G PAM4 DSP chipsets and the 400ZR DCI solutions.