Why Generative AI Could Grow After Bad Nvidia Earnings – Forbes

AUSTIN, TEXAS – MARCH 12: Kevin Kelly attends “The Universal Intern and Partner: How Generative AI … [+] is Changing How we Work” during the 2023 SXSW Conference and Festivals at Austin Convention Center on March 12, 2023 in Austin, Texas. (Photo by Jason Bollenbacher/Getty Images for SXSW)
Does the rush of capital into Generative Artificial Intelligence depend too heavily on one company?
After all, the Generative AI shot heard round the world was fired May 24 when chip designer Nvidia forecast much faster-than-expected second quarter growth — sending its market capitalization up $250 billion.
That left the investment world on edge in the days leading up to its second quarter results. But the chip designer did not disappoint — instead it grew faster than expected and raised its guidance.
How so? On August 24, Nvidia reported much better than expected second quarter results and raised its forecast. The company’s revenue up 88% from the previous quarter — $2.51 billion more than forecast while Nvidia guided investors to expect 170% revenue growth in the third quarter — $3.4 billion above estimates.
Can Nvidia keep blasting through investor expectations? What would happen to capital flows into the Generative AI ecosystem if Nvidia were to deliver disappointing earnings or a more conservative forecast?
I do not know the answer to the first question, but it partially depends on the answer to the second one — which is if companies continue to find high payoffs from using the technology to solve pressing business problems, then capital will keep flowing into Generative AI.
Based on my interviews with five experts — from KMPG, EY, Forrester Research, PwC, and MIT’s Sloan School — I see strong evidence of business demand for Generative AI. Here are the reasons why:

Companies are adopting Generative AI because they fear missing out on its benefits. As Sreekar Krishna, KPMG’s National Leader Artificial Intelligence & Head Of Data Engineering, told me in an August 21 interview, “People are seeing value across the board. KPMG started experimenting five months back with 15 to 20 people. A couple weeks later thousands of people were using it. Now there are 15,000 to 16,000 people at KPMG whose are using it.”
To be sure, some executives are skeptical. As Joe Atkinson, PwC’s Vice Chair, U.S. Chief Products and Technology Officer said in an August 25 interview,” I have been on a roadshow meeting with clients and partners. I went to 15 markets and met with 100 clients. There was peer-to-peer sharing. Overall, there is a sense of excitement and some skepticism. While people see some hype, we think Generative AI will have a massive impact.”
Yet virtually all companies are either using Generative AI or experimenting with it. According to my August 22 interview with Rowan Curran, Forrester Research FORR Senior Analyst, “There is a huge amount of excitement and hype regarding Generative AI. Companies are building and using large language models and generative adversarial networks. They are experimenting with full production Generative AI — building and buying the needed capabilities. Others are doing small explorations. A vanishingly small number of companies are doing nothing with Generative AI.”
Companies are in a hurry to capture what they see as Generative AI’s productivity and efficiency improvements. According to my August 23, 2023 interview with Dan Diasio, EY’s Global Consulting AI And Automation Leader, “In July, I interviewed 1,300 CEOs. 65% were clear AI would have a significant role in their business. Everybody is dipping their toe in the water. What is different about Generative AI is they don’t just want to learn about it; they want to push it into production. It is already creating value. It is augmenting work — resulting in more productivity and greater efficiency.”
Generative AI has been around for decades, however, experts agree its ease of use compared to other technologies — such as blockchain, the metaverse, and NFTs — resulted in very rapid scaling.
Unlike these other technologies, companies are afraid of being left behind. Krishna sees value in blockchain. As he said, “Across the board there is interest. Nobody wants to be late the way they were with mobile. Why is Generative AI taking off faster than blockchain technology did? Blockchain is an engineering achievement. But the realization of blockchain was cryptocurrency and other applications. Blockchain has made improvements in cybersecurity — enabling the sharing of data across industries.”
Yet he sees Generative AI as different. “It enables the democratization of access to data — you no longer need a data scientist to get access. Companies see Generative AI as easy to test and worth putting the energy into it. When something becomes adopted, if you delay your customers will push you into adopting it,” he said.
PwC sees Generative AI as a user-friendly front end to a powerful technology— analogous to the effect of the Web browser for the Internet in the 1990s. “Most people are not engaging with blockchain and augmented reality. By contrast, it is easy for most people to see how Generative AI applies to what they do everyday. It is having a similar effect as the graphical user interface on web browsers did. For decades, a small number of people in academic and military settings used the Internet. [When the Netscape (and other) browsers were launched], internet business cases emerged. Generative AI makes AI accessible. People don’t understand a black box,” Atkinson explained.
Forrester sees Generative AI as far more appealing to society. As Curran said, “The rate of adoption is extremely high. It is different from previous technologies like NFTs and the metaverse. Generative AI models are very visceral with real world applications. There is a huge explosion of ideation.”
EY sees Generative AI as highly democratized. “Unlike many previous technological waves, Generative AI is widely democratized. It is accessible in every application. The challenge is to find the areas with the most value balanced against the risk,” said Diasio.
Companies are finding valuable applications of Generative AI for specific business functions and to solve industry-specific business challenges.
EY’s Diasio’s provided the following list of common Generative AI functional solutions:

He also provided examples of industry-specific Generative AI applications:

Many KPMG clients are using Generative AI. “Between 200 and 300 of our clients are investing in Generative AI. Companies investing the most are Microsoft, Google, Meta — they are investing billions in the technology,” said Krishna.
Financial services and manufacturing firms are also investing. “These are industries driven by manual processes. Generative AI can reduce the workload on everybody’s hands. For example, 30 days is the average time it takes to close a mortgage. People don’t want to wait that long. Manufacturers are starting to use Generative AI because it can interpret images and videos,” he said.
KPMG is using the technology internally. As Krishna explained. “We have four use cases that cut the time for complying with documentation and work paper requirements. For example, every three months a chief information security officer must certify the company is in compliance, that everyone has been trained, and so one. We have mini-chatbots enabling users to ask a few questions to automate a service.”
PwC has an ambitious forecast for how much Generative AI will add to global growth. As Atkinson said, “Generative AI will add $15.7 trillion to the global economy by 2030. This will come from a productivity gain of at least 25% — unlocking capacity to build value. Generative AI will create new business models and make existing models better.”
PwC is among many firms offers consulting services to help companies with Generative AI. “Companies are seeking consulting help on everything from Generative AI strategy to implementation. We set up an AI factory with use cases, AI specialists and data scientists. We are helping with specific value chain applications of Generative AI such as customer care, financial reporting, and training employees,” Atkinson said.
Forrester sees many companies using Generative AI for marketing and communications. As Curran said, “People are using it for marketing and design — coming up with ideas, drafting, and developing content for tag lines, white papers, blogs, and social media posts. People are also using if for inter and intra-office communications — including sales prospecting letters and memos.”
Forrester also envisions Generative AI saving time in customer-focused business processes and observes some companies struggling to turn prototype solutions into widely used applications.
As Curran said, “Generative AI can take information and summarize and translate it in comprehensible ways. It can speed up existing processes such as call centers by retrieving knowledge and summarizing transcripts of customer calls — saving time and money. People are not having much trouble getting started, but they do have challenges in doing from a prototype to enterprise-wide deployment with low latency and tight access controls,”
Startups are developing industry-specific Generative AI solutions. As George Westerman, MIT Sloan School senior lecturer, told me in an August 23 interview, “Some startups are selling Generative AI services for law, healthcare, and finance. Now Generative AI is part of what we do.”
Westerman urges business leaders to get started with Generative AI. “You should get started doing something, put your toe in the water. There is fear of being left behind. Customer facing messaging is taking off fast. 60% of companies are using Generative AI — sending emails to customers. For customer service, summarizing everything we have done before takes it to a new level,” he told me.
Companies are managing Generative AI’s risks in different ways. They are adopting policies to control how their employees use ChatGPT, they are piloting potential high value applications before rolling them out companywide, they are adopting frameworks for responsible AI, and protecting proprietary information and limiting hallucinations by building company-specific Generative AI applications.
KPMG observes some using Generative AI tools that give companies greater control. enable. As Krishna said, “Microsoft’s MSFT OpenAI Studios — which has been operating for the last six or seven months — provides companies a semi-controlled environment. They can use OpenAI without giving it proprietary information, they can choose what data they will not share with OpenAI, retain full rights to the data, and require OpenAI to delete their data.”
PwC is reducing risk by testing applications on a small scale before rolling them out. As Atkinson said, “PwC is piloting Generative AI across multiple teams. PwC will make it available more widely following the pilots. We want to deploy it in a responsible way in a safe environment with help from our employees. It will be used for customer care, help desk, analysis of data, reading and summarizing lots of data. We will train people firm-wide in responsible uses of Generative AI.”
PwC observes companies taking different approaches to balancing Generative AI’s opportunities against its risks. “Some are shutting down employee access to third party chatbots; some are allowing the use of chatbots with guardrails; others — like PwC are deploying Generative AI chatbots trained with company-only data,” Atkinson said.
PwC sees an opportunity to limit the likelihood of hallucinations and protecting proprietary data by following Responsible AI principles — such as infrastructure, governance, training, and human review. Atkinson said companies can use specific data sets — he calls micro-models — so “companies can secure proprietary data.”
Forrester advises companies to avoid free, online chatbots. As Curran said, “Using free, online chatbots is risky and dangerous because you don’t have control of inputs or outputs.” To limit such risks, he advises people to use applications such as Writer, Jasper.ai, Grammerly, and Microsoft 365 CoPilot.
Since it is early in Generative AI’s evolution, it is unclear how it will evolve. In the next few years, I am guessing companies will implement many of the Generative AI applications discussed above that save cost, boost efficiency, and raise productivity fairly quickly.
After that, companies may invest more heavily in Generative AI applications that create more customer value than competing ones and are difficult for rivals to copy. While it could be many years before all such applications are built, at some point —- unless it reinvents itself — Generative AI will become a mature industry.
Until then, capital will pour into the Generative AI ecoystem — benefiting suppliers from chip makers to management consultants.