The Picks And Shovels Of The AI Gold Rush

As artificial intelligence captures the public zeitgeist, companies such as Google, Microsoft and NVIDIA have become leading players in the race to develop AI. In 1848, during the California gold rush, many people focused on mining for gold. Some people recognized that providing crucial tools and equipment for prospectors was equally important. These enterprising individuals became known as "picks and shovels" providers of the gold rush; they supplied picks, shovels, pans, and other mining implements which were crucial for prospectors in their search for gold. By catering to the needs of miners, pick-and-shovel companies played a vital role in supporting the rush to mine gold and in Levi Straus' case, built businesses that endure to this day. It is easy to draw parallels between the 1848 gold rush and 2023's AI technology rush.

Three companies are capturing a disproportionate portion of the attention: Microsoft with its ChatGPT investments, Google with its Bard offering, and NVIDIA as the arms supplier with its H100 and A100 GPUs. However, these vendors are only part of the generative AI landscape. Arguably the largest addressable market is Enterprise AI where enterprises train their AI model against private in-house data sets. Below are five vendors well-placed to provide the underlying infrastructure for enterprise AI.

IBM

IBM has been investing in artificial intelligence for decades, showcasing its capabilities through high-profile projects like computer systems beating grandmasters at chess and Jeopardy champions. Recently, IBM introduced watsonX at its annual Think conference. The platform aims to bring advanced AI capabilities to enterprise businesses, enabling them to scale their efforts quickly. WatsonX consists of three components: WatsonX.ai, a design studio for base models and generative AI; watsonX.data, an open and hybrid data store for analytics and AI workloads; and watsonX governance which focuses on responsible and transparent AI. The platform provides a comprehensive tech stack that can train deploy and support AI capabilities across various cloud environments. IBM's emphasis on enterprise-centric AI and its collaboration with ecosystem partners demonstrate its commitment to making AI accessible and impactful for businesses. The company's announcement positions IBM as an early leader in the Enterprise AI space and will be foundational for the company’s fortunes going forward.

Elastic

Elasticsearch is an open-source project that has gained immense popularity among developers worldwide over the last few years. The company's revenue model is based on developers' commercial usage of the free version of Elasticsearch and eventually transitioning to commercial relationships. The connection between open-source adoption and commercial usage is intrinsic in Elastic's business model, enabling its expansion into various industries such as log analytics and security threat hunting.

I recently had the chance to chat with Ashutosh Kulkarni the CEO about the trajectory the company is on and our discussion zeroed in on the challenges and opportunities in leveraging generative AI for enterprise applications, highlighting the importance of combining proprietary data with public language models. Elastic's role as a bridge between these two realms is starting to come to the fore, as is the company’s ability to enable enterprises to use their specific data to enhance the context and relevance of large language model outputs. The key component in the company’s approach is the Elasticsearch Relevance Engine as a crucial tool in optimizing infrastructure and improving the accuracy of AI-generated answers by providing contextual information to the models which will be crucial as Enterprise AI matures.

Oracle

Oracle has been a key provider of enterprise software for decades and is seen by many enterprises as a key provider for the storage of mission-critical data in its database solutions. The other key component of Oracle’s portfolio is for enterprise applications that amongst other things run the CRM and key back-office functions. One example is that with the Cerner acquisition, Oracle has access to a huge corpus of data in the healthcare vertical.

Oracle has recently announced its partnership with Cohere to develop generative AI services for organizations globally. The collaboration aims to automate business processes, improve decision-making, and enhance customer experiences. Oracle's generative AI services, built on Oracle Cloud Infrastructure (OCI) and utilizing Oracle's Supercluster capabilities, will provide high levels of security, performance, and value to enterprises deploying Enterprise AI. Cohere will train and deploy its generative AI models on OCI, taking advantage of the platform’s powerful GPU cluster technology. The integration of Cohere's models into Oracle's cloud applications, such as Fusion Cloud Applications and NetSuite, will enable customers to deploy generative AI to solve business challenges. Oracle's comprehensive portfolio of cloud applications, coupled with Cohere's large language models, will deliver the data security, powerful models, embedded AI services, and generative AI availability that enterprises will need to scale their Enterprise AI projects.

Lenovo

Many will see Lenovo as a provider of desktops and laptops, and they would not be wrong as the company has market leading share in these sectors. However, this is only part of the story. Lenovo is also the 4th largest provider of storage to enterprises and as such is well-placed to provide the foundational technology to corporations deploying Enterprise AI.

I recently had dinner with Kirk Skaugen the SVP of Lenovo’s Infrastructure Solutions Group and we spent our time talking about the significant investments Lenovo is making, which were highlighted the next day in the announcements the company made concerning AI. I went deep on those announcements here. Suffice to say Lenovo will be a player in providing the underlying infrastructure for Enterprise AI through demonstrating a commitment to simplifying AI deployment and providing end-to-end infrastructure solutions, and ultimately providing the tools to empower organizations of all sizes to harness the transformative power of AI across industries.

Hewlett Packard Enterprise

Hewlett Packard Enterprise (HPE) has recently introduced a new cloud service called HPE GreenLake for Large Language Models (LLMs) to enter the AI market. This service allows enterprises to access NVIDIA H100 GPUs and data science tools curated by HPE's Cray Super Computing division, enabling them to train, tune, and deploy AI at scale in a sustainable manner. Why is this important? Access to the NVIDIA kit is a key constraint on an enterprise's AI ambitions right now and HPE has just made this issue largely go away with the launch of this new service.

Based on a recent 1-2-1 briefing from HPE the company plans to release more industry-specific AI applications in the future, focusing on areas such as climate modeling, healthcare, finance, manufacturing, and transportation. According to the company, the service differentiates itself by addressing performance, data management, security, and reliability concerns while providing a trusted environment through single-tenant nodes. HPE's partnership with Aleph Alpha, a European AI company, will help streamline model development and cater to European cultures and more importantly regulations. Sustainability is also a key consideration, with initial deployment taking place in Quebec, Canada, and comprehensive efforts are being made to address the power consumption of GPUs and the resulting environmental impact from the get-go.

Early details look impressive but we will need to see this service in real deployment scenarios and at scale for a fuller picture. Watch this space.

Looking Ahead

With thousands of start-ups vying for VC capital to enable them to grow businesses in the AI space and with many of them looking at Enterprise AI uses cases the sector is certainly dynamic and evolving. However, busy executives should fight the urge to chase the latest bright shiny solutions. They should look to enduring companies such as those highlighted above and work with them to provide robust Enterprise AI platforms that will underpin their AI ambitions in the years ahead.

By offering advanced AI capabilities, bridging proprietary and public models, providing secure storage solutions and delivering generative AI services, these providers of "picks and shovels" stand to play instrumental roles in shaping the future of enterprise AI. These companies and the collaborations and commitments they are making to addressing challenges will drive innovation, and the widespread adoption of AI technologies.