Workday to buy HiredScore amid mixed Q4

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Workday's delivered fourth-quarter sales that were slightly below expectations, but handily topped its earnings estimates. Separately, Workday said it will buy HiredScore, which specializes in talent orchestration.

The company reported fourth-quarter earnings of $4.42 a share, which includes a $1.1 billion tax benefit. Non-GAAP earnings for the fourth quarter were $1.57 a share on revenue of $1.9 billion, up 17% from a year ago.

Wall Street was looking for fourth quarter earnings of $1.47 a share on revenue of $1.92 billion.

For fiscal 2024, Workday reported revenue of $7.3 billion, up 17% from a year ago. Earnings were $5.21 a share.

Workday's subscription revenue backlog was $20.9 billion in the fourth quarter, up 27% from a year ago. Carl Eschenbach, CEO of Workday, said the company saw "continued momentum with full platform customer wins and expansions within our base." Workday's Eschenbach becomes sole CEO, Bhusri executive chair

As for the outlook, Workday CFO Zane Rowe said the company is reiterating its fiscal 2025 subscription revenue guidance of $7.725 billion to $7.775 billion. Non-GAAP operating margins will be about 24.5% as the company balances investment in the business with profits.

• Workday Doubles Down on the ‘Hard Stuff’
• Workday Q3 shows strength, raises outlook
• Workday lays out growth plans through fiscal 2027: Here are the takeaways

Rowe said the fourth quarter was solid. On a conference call with analysts, Eschenbach said the company is looking for growth outside of North America.

He said:

"International represents over half of our addressable opportunity yet is 1/4 of our revenues. We're working to change that, and we're seeing early signs of progress. In EMEA, our leadership additions continue to drive improved and more consistent results."

Japan and Australia are also growing markets. 

Regarding HiredScore, Workday said it will combine HiredScore's Talent Orchestration platform with its Talent Management and Skills Cloud. Terms of the deal, which is expected to close in Workday's first fiscal quarter, weren't disclosed.

Eschenbach said the following about the HiredScore deal:

"When you think about the addition of HiredScore, that we just acquired, along with our recruiting platform, our Talent Optimization platform and our candidate engagement platform, it gives us the ability to deliver a full suite for talent acquisition around how we identify talent, how we engage with them, how we source or recruit them and then how we retain them for internal mobility. So I am really confident in our ability to deliver a strong ROI to the recruiting and talent organizations, with us having what I think now is one of the best full platform suites for talent acquisition going forward." 

Mistral AI launches Mistral Large, Mistral Small, Azure distribution

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Mistral AI said it released Mistral Large, a next-gen text generation model that's designed to be on par with OpenAI's ChatGPT, Anthropic's Claude 2 and Google's Gemini Pro. Mistral AI said Mistral Large will be available on Microsoft Azure.

The Mistral Large launch with Azure availability follows AWS' announcement that Mistral AI models would be coming to Amazon Bedrock to join models from AI21 Labs, Anthropic, Cohere, Meta, Stability AI and Amazon. Amazon Bedrock will carry Mistral 7B, Mistral AI's first foundation model, and Mistral 8x7B, which can summarize text, answer questions, and generate code.

Here's why generative AI disillusionment is brewing | Constellation ShortList™ Artificial Intelligence and Machine Learning Cloud Platforms | Get ready for a parade of domain specific LLMs | Trust In The Age of AI | How much generative AI model choice is too much?

For Mistral AI, Azure and AWS can give the company more distribution for its models, which seem to be advancing almost weekly. Enterprises have been evaluating multiple models including smaller ones that are use case specific. Meanwhile, regulators have been looking into generative AI investments and partnerships such as Microsoft's deal with OpenAI.

Mistral Large is fluent in English, French, Spanish, German and Italian and has a 32K token context window. Mistral AI said it will first bring Mistral Large to Azure, which is most closely associated with OpenAI's ChatGPT. Hyperscalers are racing to offer model choices to enterprises. Mistral Large is also available on Mistral infrastructure La Plateforme in Europe.

Although Mistral Large and the Azure deal will garner the headlines. Mistral AI also said it launched Mistral Small, which is optimized for performance, latency and costs.

Also see: How Generative AI Has Supercharged the Future of Work | Generative AI articles | Why you need a Chief AI Officer | Software development becomes generative AI's flagship use case | Enterprises seeing savings, productivity gains from generative AI | Work in a generative AI world will need critical, creative thinking

KKR buys VMware's EUC division from Broadcom for $4 billion

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

KKR said it will buy Broadcom VMware's End-User-Computing (EUC) division in a deal valued at $4 billion. With the move, Broadcom can focus on VMware's business and EUC will become an independent company.

EUC will include Workplace One, a remote access digital workspace, and Horizon, as well as an autonomous workspace emerging roadmap. In a blog post, Shankar Iyer, senior vice president and general manager at Broadcom VMware's EUC division, said KKR will be dedicated to EUC's customer base and keep existing management in place.

During its December earnings conference call, Broadcom said EUC and Carbon Black, which wasn't included in the deal with KKR, has about $2 billion in annual revenue. Broadcom CEO Hock Tan said the EUC and Carbon Black had strong assets. "We’ll find good homes for them because there are a lot of very interested parties who are more than happy to take those assets. And we’ll be very, very thoughtful about where we put those assets eventually, simply because the customers of many of these two assets, many of the customers are also the same customers to the VMware Cloud Foundation," said Tan.

At VMware Explore in August, EUC outlined a series of AI integrations with its Anywhere Workplace platform as well as virtual desktop, endpoint management and security advances. Boeing was also referenced as a VMWare Workspace One customer at scale.

In a statement, KKR Managing Director Bradley Brown said the private equity firm will continue to invest in Workspace One and Horizon as well expand its partner support and go-to-market efforts. The deal is expected to close in 2024.

KKR portfolio companies in technology include BMC, Barracuda aPeoplend Cloudera. EUC rival Citrix was acquired by Vista Equity Partners and Evergreen Coast Capital in 2022 and later merged with Tibco.

With the EUC sale, Broadcom can focus on VMware, which has rattled some customers with its pricing changes.

• VMware mansplains to customers about model changes, subscription pricing, account upheaval
• Dell Technologies to exit commercial agreement with VMware
• Will VMware customers balk as Broadcom transitions them to subscriptions?
• BT150 CXO zeitgeist: AI trust, AI pilots to projects, VMware angst, projects ahead

With Equifax's cloud transformation at finish line, AI scale comes into focus

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

This post first appeared in the Constellation Insight newsletter, which features bespoke content weekly and is brought to you by Hitachi Vantara.

Equifax in 2024 will decommission mainframes and data centers in North America as it largely completes a cloud transformation that dates back to 2018. Now Equifax CEO Mark Begor is looking for more product velocity, artificial intelligence capabilities and competitive advantage.

The Atlanta-based giant is best known as a credit reporting agency but has a portfolio of products via an insights engine powered by multiple data stores financial services, mortgage, commercial and residential real estate, auto, healthcare, government and employer services to name a few.

Begor, speaking on Equifax's fourth quarter earnings call noted that cloud-enabled AI using a combination of proprietary models and Google Cloud's Vertex AI will enable the new Equifax.

BT150 Interview: Equifax's Manish Limaye on data architecture, transformation

"We are convinced that our new EFX cloud single data fabric and AI capabilities are delivering new differentiated products faster with better performance and will provide a competitive advantage to Equifax for years to come," said Begor.

Equifax's journey highlights how cloud migration, digital transformation and data architecture are required to move AI efforts ahead. Equifax has also invested in transformation as its mortgage business has suffered due to rising interest rates. Fewer refinancing and home mortgages mean less Equifax data services.

Among the key milestones for Equifax in 2023:

• 70% of Equifax revenue is in the new Equifax Cloud.
• Equifax decommissioned 7 data centers and migrated 37,000 customers to Equifax Cloud.
• Saved more than the targeted $210 million goal.

And in 2024, Equifax plans on the following:

• Completing the North America move to the cloud and migrate all customers to Equifax Cloud.
• Decommissioning mainframes completely as well as remaining North American data centers.
• Progressing with transformation efforts in Europe and Latin America.
• Saving an incremental $90 million a year in 2024.
• And have about 90% of revenue in Equifax Cloud with the majority of new models and scores being built with Equifax AI.

Begor noted that the cloud enabled Equifax has launched more than 100 new products a year for the last four years. In addition, average revenue per new product is up nearly 50% since 2021.

For 2023, Equifax reported net income of $545.3 million on revenue of $5.26 billion, up from $5.12 billion in 2022. The company's divisions include Workforce Solutions, Online Information Solutions, Mortgage Solutions and Financial Marketing Services.

Equifax cloud's journey

In 2021, Equifax outlined Equifax Cloud, a cloud native architecture designed for highly regulated data workloads. The launch of Equifax Cloud took $1.5 billion in investment.

Equifax Cloud was primarily built on Google Cloud, but Equifax in its annual report does cite Amazon Web Services and other vendors used for outsourcing.

According to case studies from Google Cloud and AWS, Equifax uses Google for its cloud data architecture and AWS to host mission critical applications.

Equifax said it built Equifax Cloud because it needed to build something once and then deploy across its markets with customizations that required little engineering.

Manish Limaye SVP, USIS Chief Architect & Head of Data Engineering Equifax, said in an interview at Constellation Research's Connected Enterprise 2023 the Equifax decision to build on Google Cloud had a lot to do with machine learning and AI expertise.

Limaye, a BT150 member, said:

"Equifax is deep into statistical modeling. AI, you knew was going to be big. We were one of the leaders in the explainable AI space. When we partnered with Google, it was more than just what I'd call run of the mill cloud transformation. We wanted a deep partnership with Google beyond Google Cloud into data engineering. We partner with them. We learn from them. They learn from us because we have the most complex data. There's also the value and security of the data."

The company's cloud transformation also had a lot to do with security. Equifax has argued that a distributed cloud architecture reduces the attack surface.

In Sept. 2017, Equifax announced a data breach that exposed 147 million people. The company settled with various regulators for $425 million to help people affected by the breach.

Limay said Equifax vowed to be a security leader after its data breach.

"That security commitment really meant rethinking and reimagining how we look at securing our data given the sensitivity of it. We came up with a proper security control framework and paired it with our cloud native capability. There is an ability to destroy and rebuild data at any time. When you pair security frameworks and cloud together you walk away with a very comprehensive security framework. That control framework gets translated into a series of technical requirements. It's a very rigorous process."

Equifax CFO John Gamble said the company will gain operating leverage in the second half of 2024 due to the cloud migration. "In 2024, cost savings we will generate from decommissioning of North American infrastructure in the second half of '23 will exceed the redundant system and migration costs we are incurring, generating about 30 basis points of margin benefit," said Gamble.

For Begor, Equifax's cloud transformation is also evaluated based on new product releases. Equifax tracks a Vitality index, which is a percentage of revenue in any given year derived from new product releases over the previous three years and current year.

Equifax's multi-cloud approach reshuffles vendors

As Equifax's cloud transformation evolved so did its contracts and core vendors.

Starting in its 2019 annual report, Equifax began showing cloud payments to Google Cloud and AWS. Its 2018 annual report focused on payments to IBM, which was cited as a key vendor.

Equifax also noted that its cloud transformation was also a risk. The company said in its 2019 annual report filed with the Securities and Exchange Commission: "We are transitioning and migrating our data systems from traditional data centers to cloud-based platforms. This initiative will place significant strain on our management, personnel, operations, systems, technical performance and financial resources and internal financial control and reporting function. In addition, many of our existing personnel do not have experience with native cloud-based technologies and, as a result, we have and will continue to hire personnel with such experience. This effort will be time consuming and costly."

Here's the breakdown of Equifax's vendor roster from 2018 to 2023:

2018: Equifax said it had separate agreements with IBM, Tata Consultancy Services, Fidelity Information Services and others to "to outsource portions of our computer data processing operations, applications development, business continuity and recovery services, help desk service and desktop support functions, operation of our voice and data networks, maintenance and related functions and to provide certain other administrative and operational services." Equifax said it paid IBM $49 million in 2018, $40 million and 2017 and $45 million in 2016.

2019: Equifax added Google and Amazon Web Services as core vendors along with IBM and Tata Consultancy Services. Equifax's future minimum contractual obligation to its technology vendors at this point was $296 million. IBM was paid $52 million in 2019. Equifax noted that its payments to technology vendors could vary based on the volume of data processed and "significant new technologies."

The company said it paid Google $14 million in 2019, up from $7 million in 2018. Payments to AWS weren't disclosed.

2020: Equifax had agreements with Google, AWS, IBM, Tata Consultancy Services and others for portions the functions outlined in previous years. Equifax's future minimum contractual obligation to its technology vendors was about $318 million at the start of 2021.

Equifax outsourced mainframe and midrange operations, help desk service, desktop support and network operations to IBM in various regions. IBM was paid $50 million in 2020. By this time, Google started gaining wallet share. In 2020, Google was paid $29 million, more than double its 2019 payment.

2021: Equifax again cited Google, AWS, IBM and Tata Consultancy Services as core vendors with the aggregate minimum contractual obligation remaining of $902.4 million going into 2022.

Of that sum, the minimum contractual obligation to Google was $520 million. In 2021, Google was paid $62 million. IBM was paid $51 million.

2022: Equifax's vendor lineup changed in 2022. Including Google Cloud, AWS, UST Global, Kyndryl (formerly a part of IBM) and others, Equifax's future minimum contractual obligation at the start of 2023 was $948 million. Google and AWS accounted for the majority of that sum.

In its 2022 annual report, Equifax disclosed that its future minimum contractual obligation to Google Cloud was $440 million with no individual year's minimum to exceed about $120 million. In 2022, Equifax paid Google $152 million, up from $62 million in 2021.

For the first time in a regulatory report, Equifax outlined its obligations to AWS for hosting mission critical applications. Equifax's future minimum contractual obligation to AWS was $222 million with no minimum to exceed about $52 million. Equifax paid AWS $74 million in 2022 and added that it paid AWS $58 million in 2021 and $43 million in 2020.

2023: Equifax cited agreements with Google, AWS, UST Global and Kyndryl with an aggregate minimum contractual obligation of $1.4 billion as of Dec. 31, 2023. The minimum contractual obligation to Google for the remaining term was $1 billion with no individual year to exceed $228 million. Google was paid $171 million in 2023, up from $152 million in 2022. AWS' future minimum contractual obligation was $173 million for the remaining term with no individual year exceeding $52 million. AWS was paid $52 million in 2023, down from $74 million in 2022.

Focusing on Equifax AI

With Equifax's cloud transformation mostly complete, Begor said the company is focusing on Equifax (EFX) AI, which leverages the company's internal data, proprietary data sets and models from its Ignite platform as well as Google Cloud Vertex AI.

"Our proprietary data at scale and our single data fabric leveraging our new EFX Cloud gives us significant advantages in using AI to build more predictive multi-data models, scores and products," said Begor. "Our EFX AI is enabled by our explainable AI solutions that leverage our Ignite platform and our Google Vertex capabilities."

Begor added:

"Our investments in AI are generating results. To date, Equifax has received over 90 approved AI patents supporting areas such as our proprietary AI NeuroDecision Technology, or NDT, an explainable AI with over 130 AI patents pending. We've launched new products developing at EFX AI, including Equifax OneScore for consumers incorporating traditional credit, alternative credit, as well as cell phone utility and pay TV data, which has improved the performance of the solution to score 20% more consumers."

The game plan is to connect more data sets with Equifax AI to create new combinations of data, products and services.

Leveraging Google Cloud's Vertex AI was also an easy leap since Equifax built its data architecture on Google Cloud too. Equifax's data fabric has grown over time, but it's an internal data warehouse with additional capabilities.

Limaye said:

"You have a variety of layers--governance, observability streaming, virtualization, catalog and other things. We built our data platform on top of Google Cloud technology, and we standardized the pipelines at every step. When the data comes, it goes through the initial cleaning and transformation. It also gets entity resolution and linking. There is no differentiation between the operation and the warehouse. Because on the one hand you are getting all this data, you're cleaning it up, and you're making it available for the product. We built our own data hub where we collect data for every platform, and it brings the operational data for different types of uses. We call it purpose views. You could use it for online transactions. There's also typical data warehousing using Google technologies where you can do analytics and marketing on top of Google Cloud."

With cloud savings and new AI-driven products for growth, Equifax is projecting 2024 revenue of $5.72 billion, up 8.6% from 2023. The company expects that its margins will expand due to organic growth and cost savings from its cloud migration. "As we look beyond 2024, the cost benefits of completing our cloud migration as well as accelerating high variable profit revenue growth are expected to drive significant improvement in EBITDA margins," said CFO Gamble.

DOWNLOAD FULL CUSTOMER STORY HERE

BT150 CXO zeitgeist February edition: Low marks for SAP RISE, process automation, change management, AI risk

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Constellation Research held its February call with Business Transformation 150 and SAP's RISE program, process automation and AI were top-of-mind topics.

These gatherings, held under Chatham House rules, are a venue to share information and emerging trends. Here's a look at the topics from our February meetup.

Previously: BT150 CXO zeitgeist: AI trust, AI pilots to projects, VMware angst, projects ahead

SAP's RISE program

• One CIO asked the group for opinions on SAP's RISE program and being forced from on-premises to the cloud. The goal was to have a strategy for SAP in place by the end of the year.
• CXOs weren't thrilled about SAP RISE and items like licensing credits for legacy environments. A CIO wondered what would prevent a customer from moving away from SAP--especially since the enterprise operates in a space that doesn't garner investment from the enterprise software giant.
• One option could be to build a homegrown ERP system to replace SAP. 
• SAP's RISE program is viewed as an exercise in financial engineering more than something that benefits customers. 
• Being on-premises is actually a good thing because the customer isn't beholden to SAP and already paid for the asset. Rimini Street was frequently mentioned as an option for maintenance and support to preserve the initial SAP investment and keep options open down the road.
• For enterprises that don't want to migrate to SAP's cloud, there's an option to build an abstraction layer on top of the transactional system that's eating up budget without value.

Our BT150 CXOs aren't alone. SAP's German speaking user group takes aim at cloud contracts, BTP and more | SAP user group DSAG rips S/4HANA innovation plans, maintenance increases | SAP retools for generative AI, cuts 8,000 jobs, sets 2024, 2025 ambition

Process mining and automation

• Some CXOs were exploring proof of concept projects for process mining and modeling tools. The concerns revolved around process mining turning into more work than it needed to be unless the underlying systems were emitting real signals.
• Process mining was seen as an option to automate repetitive jobs. Celonis and SAP Signavio were mentioned as options.
• Process design is critical to making automation projects work.
• Enterprise automation needs processes that are designed by the bottoms up with employees on the front line. Enterprises should look to redesign processes completely and creatively destroy them to find new ways to work.
• Change management is a required art form to make automation work. In many cases, the objective for employees and departments is to maintain a process that has been running for decades. Resistance to change torpedoes automation projects and CXOs need to evangelize and bring in multiple stakeholders.

Celonis launches Process Intelligence Graph, makes case process enables automation, AI applications | SAP buys LeanIX, aims to couple it with Signavio, system transformation

Management tips

• Use regulatory and security requirements to move projects along and get stakeholders on the same page.
• Sell teams on projects that will last beyond today and influence the future. 
• Enterprise change is more successful if it starts from the inside than outside.
• Plan projects assuming that the CEO has a tenure of 4 years and so does the team attached to that leader. Year one is likely to be about cost cutting. Year two is about innovation funded by the savings of year one. Years three and four are about delivering results that renew the CEO's contract. The goal is to implement technology that can span leadership teams.

Return on Transformation Investments (RTI)

AI risks

• CXOs on our panel all referenced Air Canada's rogue bot that gave a discount that was upheld in court. There are other examples of bots giving deals that no human would approve. CXOs noted that safeguards, process maturity and a kill switch are required to prevent bots from becoming headaches.
• Enterprises need to think through where a human goes into a process to establish trust.
• Enterprises also need to ponder legal liability with generative AI.
• AI applications are becoming a key theme as enterprises look to renew long-term deals with software vendors. Multi-year deals may smooth out costs and contract negotiations.

Reddit's data licensing play: Do you want your LLM trained on Reddit data?

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Reddit has filed to go public in a closely watched initial public offering and one of its potential revenue streams is data licensing. The big question is whether enterprises want that data in a large language model. Perhaps the bigger question is whether enterprises will have a choice.

The social community has made a deal with Google to provide it model training data. Reddit also indicated that it'll provide training data more broadly. Reddit has sentiment, conversational data, and a lot of useful stuff to go along with the ridiculous. In some ways, Reddit data is likely to make LLMs a bit more serendipitous.

Reddit's data licensing plan highlights how training data from media-ish companies is valuable. The plan also highlights Hobson’s Choice facing these companies. Reddit gets revenue by providing training data, but also could get squashed by an LLM in the future.

According to Reddit's S-1 filing with the SEC:

"Redditors may choose to find information using LLMs, which in some cases may have been trained using Reddit data, instead of visiting Reddit directly. We believe that our ability to compete effectively for users depends upon many factors both within and beyond our control."

Not that Reddit has a choice. Data licensing is a high-margin business and Reddit lost $90.8 million in 2023 on revenue of $804.03 million. The company did pare losses from 2022 and may be headed in the right direction if data licensing works out. Advertising is a rough business.

Reddit specifically names ChatGPT (OpenAI CEO Sam Altman is a shareholder of Reddit by the way), Gemini (Google paid Reddit for data) and Anthropic as competition. Reddit's list of competitors for your attention is also extensive: Google, Meta, Wikipedia, X, Snap, Roblox, Discord and any other company that can give you an answer. Marketplaces are also cited as rivals.

In other words, every Reddit competitor will also likely have a data licensing business. It's clear that users have become the product and the data licensing model.

As these data licensing models proliferate, data lineage is going to become critical. Ultimately, I'd want to know what data the LLM was trained on and be able to back out some sources. This ability is especially critical for media sources. Media isn't objective and you need a few outlets (and preferably source documents) to even figure out what the actual truth is. Media has become like navigating a divorce. There's the first spouse's truth. There's the other spouse's truth. And then THE truth.

• Shutterstock will bring its training data to Databricks, Snowflake, Amazon, Google Cloud
• Get ready for the generative AI training data licensing boom

• BT150 CXO zeitgeist: AI trust, AI pilots to projects

For companies like Shutterstock, which is going to provide training data for models on data marketplaces, the business looks more straightforward. Other providers can supply data, but in the end the LLM buyer may not actually want that data set. How do you reverse learning? How do you back out a data set? Do I want my customer service LLM dishing out some snarky comment it learned on Reddit, X or Meta properties? Is Reddit's training data really just a small subset of the company's more than 500 million visitors and 73 million daily active users that actually comment?

We should be asking these questions now because the data licensing parade is just about to leave the station. In its IPO filing, Reddit said it has "one of the internet’s largest corpuses of authentic and constantly updated human-generated experience" and is in the early stages of data licensing. Reddit said its platform is good for real-time perspective on products, market sentiment and signals. It'll provide data API access and model training from Reddit's real-time content.

Ideally, I'd want to see a nutrition label on LLMs, and a list of data sources used for training. I'd also want to check off sources too. It's unlikely we'll get those options, so enterprises are likely to find small language models aimed at specific use cases and industries more valuable. 

• Here's why generative AI disillusionment is brewing | Constellation ShortList™ Artificial Intelligence and Machine Learning Cloud Platforms | Get ready for a parade of domain specific LLMs | Trust In The Age of AI | How much generative AI model choice is too much?

Shutterstock will bring its training data to Databricks, Snowflake, Amazon, Google Cloud

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Shutterstock is best known as a stock image provider and owner of Giphy, but the money and margins may come from training data for model training.

During the company’s fourth quarter earnings call, Shutterstock CEO Paul Hennessy said the company is looking beyond wholesale data deals with the likes of Meta and OpenAI to licensing data across marketplaces. Shutterstock sees the data business as something that can grow at a compound annual growth rate of 22% through 2030.

• GenAI trickledown economics: Where the enterprise stands today

Hennessy said:

"As we look ahead, AI and machine learning model training will continue to be a growth opportunity, especially as we look to diversify our revenue base by targeting new buyers beyond the hyperscalers. In fact, we just won our first seven figure contract involving a venture backed startup in the generative AI ecosystem, and we feel there are much more such opportunities ahead. We'll also be expanding our delivery model by leveraging our cloud marketplace partners. This will allow us to go from being a wholesale provider of data to the likes of Meta and OpenAI to a retail provider of data to the hundreds of companies we believe are going to custom train their own models.

To that end, we are in the process of rolling out Shutterstock's training data onto data marketplaces of DataBricks, Snowflake, Amazon and Google Cloud."

Shutterstock's data business had revenue of $104 million in 2023. Today, Shutterstock has 10 anchor customers for training data, but thinks it can get to 150 over time.

"The data marketplaces of Snowflake, Databricks, AWS and Google are comparatively small but fast-growing businesses for these companies as their customers learn how to enrich and monetize their own data and models," said Doug Henschen, VP and Principal Analyst at Constellation Research. "Interest in GenAI will further accelerate the growth of these data marketplaces as companies start to build their own small and midsize models, including ones harnessing image data."

Giphy has a role in this generative AI play for Shutterstock because the content platform is connected with APIs. That know-how has given Shutterstock "API relationships" with major technology players. Shutterstock will invest in 2024 to build out its training data business.

Hennessey said the plan is to bring Shutterstock data to where the customers are. "Our customers don't naturally think of Shutterstock as a place to go for computer vision training data and training their generative AI models, but they do typically go to a DataBricks or a Snowflake or an AWS or GCS in order to acquire training data. This is also the natural compute environment for these customers," said Hennessey, who added that those partners will boost distribution without hurting margins.

"The way these distribution channels make money is not by taking a cut of the data sales, it's through the compute. And so they're looking forward to having our data on their ecosystems, so they can drive additional compute in the cloud," he said.

It remains to be seen how Shutterstock's training data business develops. Contracts can last years or be shorter depending on volume-based pricing.

Shutterstock reported 2023 revenue of $874.6 million, up 6% from a year ago, with net income of $110.3 million.

Nvidia's GPU boom continues, projects Q1 revenue of $24 billion, up more than $20 billion a year ago

Larry Dignan

Editor in Chief of Constellation Insights
Constellation Research

Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, analysis, interviews, and event coverage designed to help technology buyers and vendors navigate complex markets with clarity and context. ...

Read more

Nvidia's data center business continued to surge and the company continues to raise its revenue guidance at a rapid clip.

The company reported fourth quarter revenue of $22.1 billion, up 265% from a year ago, with earnings of $4.93 a share. Non-GAAP earnings for the fourth quarter were $5.16 a share.

Wall Street was looking for Nvidia to report fourth-quarter non-GAAP earnings of $4.64 a share on revenue of $20.62 billion.

For fiscal 2024, Nvidia reported earnings of $11.93 a share on revenue of $60.9 billion.

• GenAI trickledown economics: Where the enterprise stands today

The fact Nvidia is raking in dough isn't surprising given cloud providers and Meta said they were spending heavily on GPUs to train models. Constellation Research CEO Ray Wang said "we haven't hit peak Nvidia yet." 

By business unit, Nvidia's quarterly revenue trend is almost comical. The company's fourth quarter data center revenue was $18.4 billion, up from $3.83 billion a year ago.

As for Nvidia's outlook, the company projected first quarter revenue of $24 billion with GAAP margins of 76.3%.

To date, most of the spoils from the generative AI boom have gone to Nvidia with SuperMicro being an exception.

• Nvidia's uncanny knack for staying ahead
• Nvidia launches H200 GPU, shipments Q2 2024
• Nvidia sees Q4 sales of $20 billion, up from $6.05 billion a year ago
• Why enterprises will want Nvidia competition soon

CEO Jensen Huang said generative AI and accelerated computing has hit an inflection point with surging demand. "Our Data Center platform is powered by increasingly diverse drivers — demand for data processing, training and inference from large cloud-service providers and GPU-specialized ones, as well as from enterprise software and consumer internet companies," he said.

Nvidia CFO Colette Kress said:

"In the fourth quarter, large cloud providers represented more than half of our Data Center revenue, supporting both internal workloads and external customers. Strong demand was driven by enterprise software and consumer internet applications, and multiple industry verticals including automotive, financial services, and healthcare. Customers across industry verticals access Nvidia AI infrastructure both through the cloud and on-premises."

Kress added that sales to China declined significantly. Gross margins improved due to lower component costs. Nvidia's cash, cash equivalents and marketable securities was $26 billion, nearly double from a year ago. Regarding inventory, Kress added:

"Inventory was $5.3 billion with days sales of inventory (DSI) of 90. Purchase commitments and obligations for inventory and manufacturing capacity were $16.1 billion, down sequentially due to shortening lead times for certain components. Prepaid supply agreements were $5.0 billion. Other non-inventory purchase obligations were $4.6 billion, which includes $3.5 billion of multi-year cloud service agreements, largely to support our research and development efforts."

Kress said data center revenue was driven by the Nvidia Hopper GPU computing platform along with InfiniBand and and networking. Compute revenue grew more than 5x and networking growth revenue tripled from last year, she said. "We are delighted that the supply of Hopper architecture products is improving and demand remains very strong. We expect our next generation products to be supply constrained as demand far exceeds supply," said Kress.

Conference call highlights

Key items from the conference call with analysts:

Nvidia's networking business is at a $13 billion annualized revenue run rate. Spectrum X Ethernet has adaptive routing congestion control, noise isolation and traffic isolation. Huang said Spectrum X Ethernet will be an AI optimized system and InfiniBand will be an AI dedicated networking option.

Huang said there's strong demand across industries but cited healthcare and financial services as well as sovereign AI as key areas.

He said:

"Sovereign AI has to do with the fact that the language, the knowledge, the history, the culture of each region, are different. And they own their own data. They would like to use their data, train it to create their own digital intelligence and provision it to harness that raw material themselves. It belongs to them."

He added that the generative AI boom experienced in the US will be replicated in multiple countries.

Nvidia's supply constraints are improving. "Our supply chain is just doing an incredible job for us," said Huang. "The Nvidia Hopper GPU has 35,000 parts. It weighs 70 pounds. These things are really complicated. We call it an AI supercomputer for good reason."

Demand will outstrip supply. "We expect the demand will continue to be stronger than supply provides, and we'll do our best," said Huang. "The cycle times are improving. Whenever we have new products, it ramps from zero to a very large number. And you can't do that overnight."

Allocation depends on how soon customers can launch services and infrastructure. Kress said allocation is a moving process and Nvidia looks to get inventory to customers right away. Huang said cloud service providers have a clear view of Nvidia's roadmap and the visibility of what products can buy and when. "We allocate fairly and to avoid allocating too early. Why allocate before a data center is ready?" said Huang. "We bring our partners customers. We are looking for opportunities to connect partners and users all the time."

Nvidia will work within government constraints to reset its product offering to China. Nvidia is doing its best to compete. "Hopefully we can go there and compete for business," said Huang, who said the company is sampling with China customers.

Software is as important as hardware. "Software is fundamentally necessary for accelerated computing. If you don't have software, you can't open new markets or enable new applications," said Huang, who said Nvidia has teams working with a bevy of enterprise software companies.

Huang said Nvidia will do the optimization, patching and optimization for enterprise software companies. "Think of Nvidia AI enterprise as a runtime like an operating system. It's an operating system for artificial intelligence," said Huang.

Constellation Research's take

Wang said there's no reason why Nvidia can't keep rolling. He said:

"This is the most important stock in the world right now and a barometer of AI health. Four years ago we talked about Nvidia hitting a $1 trillion market cap and $2 trillion by 2025. We are in the age of AI and Nvidia is king. This is real demand and it's growing. We haven't hit peak Nvidia nor peak AI and we'll see software as the second wave of AI in about 6 months to a year."

Constellation Research analyst Dion Hinchcliffe said Nvidia's moat is also about software. Hinchcliffe said:

"Amid all the hoopla about NVIDIA's chip prowess, what most observers still miss is that Nvidia has carefully cultivated and strategically wielded its CUDA (Computing Unified Device Architecture). CUDA is the unique software stack that has won the hearts and minds of next-gen compute devs in gaming, HPC, and now AI.

CUDA has become the key competitive weapon that connects 3rd party apps to Nvidia GPUs. It’s the magic handshake that enables AI algorithms to work effortlessly with the massive compute power of Nvidia's modern chip architectures. 

But CUDA isn’t just an ordinary piece of software. It's a closed-source, low-level API that optimally wraps a powerful software stack around Nvidia’s GPUs. The result has created an ecosystem for parallel computing that's potent while also carefully keeping devs captive due to their code's dependency on it. Even the most formidable competitors such as AMD and Intel have struggled with only minor success as CUDA has become widely adopted in various industries like HPC, AI, and deep learning. 

I project that we'll continue to see Nvidia's lead grow as it continues to corral devs with the success of CUDA and and keeps competitors at bay as long as the API remains relatively non-portable."