Monetizing AI in the Semiconductor Industry

There has been a lot of hype around AI, yet questions persist as to how AI will translate into ROI. I’ve been asked for my perspective, and those inquiries have prompted me to discuss the topic with some in my network, and to think about it in some detail. This blog is a pass at sharing the perspective I’m developing, and to provide a thoughtful and consistent reply to the inquiries I receive.

In November 2022 the ChatGPT AI platform was released by OpenAI, and a fervor of interest and enthusiasm ensued. There would certainly be application of AI in ways that would drive new levels of productivity and profitability.  Yet even today, various industries where AI is being exploited seem to be in agreement that a sustained (and legitimate) path to monetization and growth has yet to emerge.

My perspective on AI is primarily through a semiconductor industry lens, so I’ll be keeping my comments here narrowed accordingly.

Companies are already making money with AI, as evidenced by quarterly earnings reports and forecasts from NVIDIA, Microsoft, Alphabet and IBM, the semiconductor supply chain, notably TSMC, as well as several LLM (large language model) platform developers. The companies with highest AI earnings that are also on the list of 2023 Top 10 US Semiconductor Companies by Revenue (see the Table in my June 15, 2024 blog) are NVIDIA, Intel, Broadcom and AMD.

NVIDIA: In 2H’23, nearly 80% of revenue was from their Data Center Division, which delivers AI client and cloud infrastructure (microprocessors, networking equipment, and AI enabling software) to their customers.

Intel: Intel®CoreTM Ultra processors (aka, Xeon 6) are design-tailored for AI user applications. Intel lagged NVIDIA and AMD in chip sales at the outset, as users such as Meta, Microsoft and Google bought up as many NVIDIA and AMD chips as they could. Intel is closing that revenue gap.

Broadcom: Provider of chips and network tools, Broadcom indisputably leads the industry in offering highly complex, customized SoC ASICs and ASSPs for deeply differentiated computing and networking, as well as application specific, systems. This expertise is being applied to design of AI processors and infrastructure products. 

AMD: Microsoft recently announced it will begin to offer AMD AI chips to its cloud computing customers. Meta and OpenAI already use AMD processors.

These 4 semiconductor companies are leading the way to monetizing AI in the Semiconductor industry by creating the front end infrastructure for the highly computational and memory intensive AI workloads. The wafer foundry TSMC is also benefiting and plans to report a 29% increase in Q2’24 net income, fueled by strong AI chip demand. The company's valuation has increased to a US$1 trillion market cap, with promise of further growth as it plans to raise wafer prices for their latest technology chips. A predicted 3% to 6% price increase could lead to a gross profit margin increase of 58% next year, reports Bloomberg.

AI offers more to semiconductor companies than productivity improvements across their workforce. It also is enabling a restructuring business cycle that reminds me of the offshoring of clerical and so-called overhead jobs that the semiconductor industry exploited in the ‘90’s. As AI reduces the need for these kinds of back end jobs, those positions can/will be expunged.

And finally, AI as a product feature add-on:

This means of monetizing AI came up several times in discussion with some in my network, so I thought it made the cut for general interest.

Some companies will likely (and should) integrate AI capability into their existing software products without charging for it (initially, anyway), therewith enhancing the value of the product with AI as an add-on feature. The company could later raise the price, justified by the value-add from that feature.  The company could also be expected to enjoy sales growth from new business by marketing such feature enhancements.

Companies in the semiconductor industry use myriad software tools, for finance, business development, QMS, supply chain management, as well as manufacturing and customer data mining and compiling, and charting tools (known as business intelligence and analytics tools). An example of such a tool is Tableau (acquired in 2019 by Salesforce). Tableau added AI to its platforms in 2023, delivering insights and contextual reporting from AI learning to each report. Tableau continues to add new features and links.