NVIDIA Is Just The Beginning: AI And The Exponential Age
NVIDIA just posted a monster quarter, and it’s important for a few reasons. NVIDIA was a gaming company just a few years ago. Its leadership saw AI as the biggest opportunity for revenue growth in the long term, and they pivoted. Over the last 2 years, NVIDIA’s assessment of where AI is heading has been very accurate. The company’s results offer insights about what’s REALLY happening in AI today and what comes next.
Competitors like AMD looked at more traditional CPU and PC markets as the primary growth drivers. AMD has had some success there, but its strategy missed the massive opportunity that AI could have delivered. Most companies are facing a similar decision. Do they stay with traditional growth drivers or make the transition to AI?
I wrote much of ‘From Data to Profit’ to support companies struggling with that strategic decision. NVIDIA is a case study of what happens when CxOs make a successful decision about their companies’ technical strategy. We are also seeing what happens when that decision is made poorly. Many companies are chasing after false opportunities or missing out altogether.
In this article, I go deeper into NVIDIA’s success and the potential risks to it continuing. I use NVIDIA’s roadmap to reveal the primary directions that AI will take next. Seeing around corners will help business leaders and data professionals position themselves for success in the next 5 years.
Breaking Down NVIDIA’s Results And Why Most Didn’t See It Coming
Some analysts call NVIDIA’s growth ‘parabolic’ and that’s uncomfortable territory for investors rooted in the linear growth paradigm. The legacy mindset is deeply entrenched in the C-Suite, too. Many will remain on the sidelines while opportunities pass by. NVIDIA delivers a case study to help data professionals and CDOs make the case for changing how other business leaders look at AI.
Investment analysts believe NVIDIA’s stock could double over the next 1 or 2 years. It’s already up 240%, and the consensus is that it could double again to nearly $1000 per share. Forecasts are backed up by the company increasing its data center business revenue from $4.2 billion to $10.3 billion. Those numbers are not year over year. The increase happened in 3 months.
For the last 3 quarters, NVIDIA has raised its estimates of the next quarter’s earnings, or guidance, above what analysts expected. All signs point to some of the most intelligent people in the business and investor analyst communities having no handle on how AI companies grow. When you cannot see where AI goes next, it’s difficult to see the opportunities and who is best positioned to seize them.
On the other hand, NVIDIA is an AI leader because they have an evidence-based approach to predicting future AI trends. NVIDIA released Megatron in early 2021, well before the Generative AI hype cycle started. The best way to predict the future has always been to build it. With NVIDIA, they have a world-class data science organization building solutions with leading-edge research. The goal isn’t to productize the solutions but to learn what’s viable.
NVIDIA gains insights from that work to power leadership’s decisions about what types of products will be necessary to support the next AI wave. It’s not enough for the technology to be viable; it must support high-value use cases and products that can be brought to market before NVIDIA invests. The company has developed a very forward-looking, prescriptive strategy out of necessity. Staying ahead of competitors demands it because chips take years to move from R&D to customers’ hands.
NVIDIA must think 3 years ahead of everyone else to be ready when the opportunities arrive. For most companies, their forward-looking technology view and strategy only need to see 12-18 months ahead. NVIDIA’s vision is a unique opportunity to gather insights and build to where the opportunities will be.
Why It’s Critical To Be In Front Of The Opportunities
The focus today is on NVIDIA’s continuation of the current trend. It can feel forward-looking, but in reality, this cycle is already in the books. Businesses must prepare strategically for the next cycle. Current strategy frameworks don’t support executing on one technology wave while preparing for the next one.
Traditional strategy is built with the assumption that technology waves come every 7-10 years or more. I built the continuous transformation framework to support a different paradigm of strategic decision-making. Technology waves break every 2-4 years, and that distance is decreasing rapidly. For those of us in the data science field, this is obvious. For people outside the domain, it sounds ridiculous. That’s why we see such slow reaction times and massive pushback to the new paradigm.
We are firmly in exponential territory for current AI workloads. The trend is playing out with hyperscalers like Azure and GCP showing accelerating AI workload growth. Even smaller cloud players are quickly ramping up to take advantage of the opportunity. Generative AI had its ChatGPT moment just over a year ago. Had you bought NVIDIA’s stock at the beginning of OpenAI’s hype cycle, you would have picked it up for $120 a share and made 4X in roughly 9 months.
It took 6 months for the workload increase to hit the major cloud providers, but the rewards aren’t evenly distributed across the board. Early indicators are that Salesforce, Microsoft, and Google are positioned to be winners.
Snowflake is struggling with large customers looking for ways to reduce overall cloud spending. AI workloads are rising. However, businesses not viewed as AI leaders don’t get those workloads, even if their platforms can handle them.
Other cloud providers will show up to the party late and get a tiny share of the gains. The Snowflakes and latecomers are much like AMD. They will see linear benefits, not hyperbolic gains. It’s critical to be ready when customers are. Forward-looking businesses will take most of the opportunity before competitors can react.
In 3-6 months, downstream businesses that provide software to enable LLM product development will see their lift. If that trend continues, 9-12 months from now, businesses will begin to report earnings from Generative AI products. Again, there will be an uneven distribution of revenue gains. It’s not just about being first to market but being first to the opportunity and meeting customers when they get there.
Where Does NVIDIA See The Field Going Next?
The attention is on data centers and AI workloads in the cloud, but that’s just the tip of the iceberg. AI has several waves that will break over the next 5 years. I don’t believe anyone can predict the exact timing, but NVIDIA has several projects and products lying in wait for customers.