Why AWS may need a new generative AI narrative

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.

AWS dutifully rolled out a bevy of generative AI tools and model choices for developers. As usual, the AWS cadence is quick. But the bigger question is whether AWS is speaking to the right crowd when it comes to generative AI. The cloud land grab was bottoms up starting with developers. Generative AI is a boardroom issue and Microsoft and Google Cloud appear to have the better CXO narrative relative to AWS.

At AWS' New York Summit, Swami Sivasubramanian, VP of Databases, Analytics, and ML at AWS, outlined a bevy of generative AI updates to Amazon Bedrock, introduced new models and instances and talked vector data stores. "Generative AI has captured our imagination. I believe it'll transform every industry and business," said Sivasubramanian.

From there, AWS launched multiple models for Amazon Bedrock and other generative AI services. The theme: AWS has the infrastructure and the developer base. Give that base the tools and be the generative AI enabler. This approach is straight from the AWS playbook that has worked so well for years.

What's changed? The decision-makers. Microsoft and Google are simply talking a better generative AI game right now and appear to be capturing mindshare as well as workloads. You can debate whether Microsoft picked a fight with Google that it can't win, but what's more interesting is how AWS isn't mentioned much.

To hear Microsoft CEO Satya Nadella tell it, Azure and Microsoft Cloud are in the generative AI pole position. During Microsoft's fourth-quarter earnings conference call, Nadella said:

"We have grown Azure over the years coming from behind. And here we are as a strong #2 in the lead when it comes to these new workloads. So, for example, we are seeing new logos, customers who may have used out of the cloud for most of what they do, or for the first time, sort of starting to use Azure for some of their new AI workloads.

"We are seeing new customers starting to use Azure for some of their new AI workloads. We also have customers who have used multiple clouds start new data and AI projects."

Nadella talks about CoPilot, intelligent data platforms and all the things CXOs talk about: Productivity, efficiency and process. Toss in a few big-name customers like Chevron and boardrooms listen. It doesn't hurt that enterprises are often Microsoft shops.

Google Cloud is playing a different game, but CEO Sundar Pichai is well-versed in AI and sprinkling generative AI throughout its portfolio. Pichai talks about infrastructure but also returns. Google Cloud also courts developers but is using generative AI to upsell its base of customers.

Pichai said:

"We are making it easier for others to innovate using AI. One way is by providing Google Cloud's high-performance infrastructure, optimized for a range of generative AI models. It's being used by thousands of customers and partners to transform their businesses."

Google Cloud, which is playing from behind at No. 3, is betting generative AI will boost its addressable market. Pichai continued:

"Our new generative AI offerings are expanding our total addressable market and winning new customers. We are seeing strong demand for more than 80 models, including third-party and popular open source in our Vertex, Search and Conversational AI platforms with a number of customers growing more than 15x from April to June. Our generative AI capabilities also give us an opportunity to win new customers and upsell into our installed base of 9 million paying Google Workspace customers."

Pichai said Google Cloud is "engaging in more conversations" with enterprise executives.

Even the AWS-obsessed Oracle CTO Larry Ellison is in on the act when it comes to AI workloads.

What's unclear is whether generative AI ultimately means AWS loses a material amount of market share. AWS is the clear cloud leader, but now the company looks like it needs a storyline that appeals to CXOs as much as developers. Stay tuned for AWS' second quarter results Aug. 3. 

Enterprise IT vendors see gradual demand improvement, AI-driven buying

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

Enterprise technology demand has stabilized, and vendors are expecting gradual demand improvement into the second half with pockets of generative AI-driven strength.

Financial reports from key enterprise tech vendors are confirming what a Constellation Research survey has found. CXOs are more optimistic, positioning for growth and looking toward automation and generative AI projects. In fact, there's a decent bit of IT spending optimism--or at least the theory that the worst is over.

Constellation Research's 2023 H1 "CxO Business Confidence Survey" found that 32% of the IT budget is going to investments that add to top-line growth and 28% is going to efficiency. That balance is reflected in comments from enterprise technology vendors. Here's a look at what vendors say on earnings conference calls. 

Francois Locoh-Donou, CEO of F5

"From a demand perspective, we are seeing some early signs of stabilization. Q3 demand played out slightly above our beginning of quarter forecast, which was up from Q1 and Q2 this year, though still off from FY '22 levels."

Ben Suh, SVP of IR at Samsung

"We expect global IT demand in the business environment to recover gradually, driving improved results especially in our component businesses. However, challenges associated with a demand recovery are also likely to persist due to macro risks and other factors."

ServiceNow CEO Bill McDermott

"We see a sustained demand environment and pipeline for all of our product businesses, geographic regions and industry verticals. We're set up very well for a strong second-half.

This is a dynamic period for the IT industry. Think about it this way. Every leader in every department, in every business, in every industry is writing a new playbook for the AI world. CEO's are sponsoring them. The C-suite across all functions is funding them." 

ServiceNow launches two generative AI use cases with proprietary LLM, outlines commercialization plans

Seagate CEO Dave Mosley

"In today’s earliest stages of gen AI development, you’re seeing the necessary first steps of building and training of AI models. These efforts require significant investment in compute architecture, and we’re seeing that investment ramp today. The next stage of development will yield enterprise-specific use cases that leverage trained AI models to convert data into value-enhancing applications."

Microsoft CEO Satya Nadella

"Every customer I speak with is asking not only how, but how fast they can apply next-generation AI to address the biggest opportunities and challenges they face, and to do so safely and responsibly."

Microsoft Q4 better than expected with Azure revenue gains of 26%

PROS Holdings CEO Andres Reiner

"Across the B2B industries we serve, sales cycles with new customers in 2023 are 30% faster year-over-year. We're also driving rapid time to value, leading to rapid expansions."

Alphabet CFO Ruth Porat

"Google Cloud Platform revenue growth remained strong across geographies, industries and products. That being said, we saw a continued moderation in the rate of consumption growth as consumers optimize their spend."

Google Cloud revenue in Q2 tops YouTube ads

Related:

• A look at the surprises of 2023 so far...

Intel's Q2 better than expected, issues remain

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

Intel's second quarter results weren't as bad as feared as the company saw an uptick in its Intel Foundry Services business and declines elsewhere.

The company, which is lagging Nvidia in processors used for AI workloads, reported second quarter revenue of $12.9 billion, down 15% from a year ago. Intel delivered net income of $1.5 billion, or 35 cents a share, in the second quarter. The company reported second quarter non-GAAP earnings of 13 cents a share.

Wall Street was expecting Intel to report a loss of 3 cents a share on revenue of $12.13 billion, down 21% from a year ago. At the end of the first quarter, Intel forecasted second quarter revenue of $11.5 billion to $12.5 billion with a non-GAAP loss of 4 cents a share.

As for the outlook, Intel projected non-GAAP third quarter earnings of 20 cents a share on revenue of $12.9 billion to $13.9 billion.

• AMD makes its case for generative AI workloads vs. Nvidia

While the second quarter results were better than expected, Intel is still facing multiple issues. Its PC chip business was down 12% from a year ago to $6.8 billion and its data center and AI business had revenue of $4 billion, down 15% from a year ago. Network and edge saw revenue fall to $1.4 billion, down 38% from a year ago.

Intel noted the following:

• In the client computing unit, revenue fell as PC makers worked through excess inventory.
• For the data center and AI group, Intel said it saw lower revenue due to competitive pressure (AMD and Nvidia) as well as a shrinking total addressable market for CPUs.  
• In network and edge, revenue ws lower on weak telecom demand. 

On the bright side, Intel Foundry Services had second quarter revenue of $232 million, up 307% from a year ago.

Intel CEO Pat Gelsinger said the company was "well-positioned to capitalize on the significant growth across the AI continuum by championing an open ecosystem and silicon solutions that optimize performance, cost and security to democratize AI from cloud to enterprise, edge and client."

The company also said it's on track to deliver $3 billion in cost savings in 2023.

Speaking on a conference call with analysts, Gelsinger said the following:

• Intel said the company is regaining customer trust with expectations of a "modest second half 2023 recovery."
• The client business performed better than expected with share gains in consumer and education. Gelsinger added that the AI PC will drive demand as well now healthy inventory levels. 
• The CPU server market remains soft in enterprise and internationally. Gelsinger said focus has been on AI accelerators. He expects the server market to recover in 2024. 
• "Our strategy is to democratize AI across the continuum," said Gelsinger, adding that Intel will bring AI to cloud to edge to client. 
• Foundry services is "a significant accelerant to our Intel 2.0 strategy" and diversifies the semiconductor supply chain outside of Asia. Gelsinger said it is investing in additional capacity. 
• "I'm pleased to report all of our programs or on track or ahead of schedule," said Gelsinger, noting that Intel plans to close the power and performance gap by 2025. 

ServiceNow launches two generative AI use cases with proprietary LLM, outlines commercialization plans

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

ServiceNow launched case summarization and text-to-code generative AI tools built on its ServiceNow LLM, which is optimized for its platform. On the commercialization front, ServiceNow said it will introduce new premium generative AI offerings across IT service management, customer service management and HR service delivery in September with its Vancouver platform release. 

Large language models (LLMs) are a big focus for enterprises looking to add generative AI capabilities to improve productivity. ServiceNow's approach is to create purpose-built generative AI capabilities that can be deployed quickly across IT, HR and customer service cases. The two new use cases are part of ServiceNow's Now Assist family of generative AI features. ServiceNow's Generative AI Controller enables enterprise to connect to ServiceNow instances, Microsoft Azure OpenAI Service and OpenAI API LLMs. 

Premium SKUs for generative AI features are also a big topic since it can fuel ServiceNow's growth. ServiceNow saw strong growth for its current premium offerings, ITSM Pro and CSM Pro and it's likely that technology buyers will be kicking the tires on the company's generative AI "Pro Plus" offerings. ServiceNow is lining up partners such as KPMG to reinvent finance, supply chain and logistics using its platform. ServiceNow also launched AI Lighthouse, a program with Nvidia and Accenture to speed up enterprise generative AI deployments. 

ServiceNow CEO Bill McDermott said the company "is already seeing our own significant productivity increases with the generative AI solutions we're releasing to the market." 

In the second quarter, ServiceNow topped estimates on multiple fronts. 

The company saw subscription revenue growth of 25% compared to a year ago. ServiceNow had 70 transactions worth more than $1 million in net new ACV and 45 customers with more than $20 million in ACV. ServiceNow reported second quarter revenue of $2.15 billion, up 23% from a year ago, with net income of $1.04 billion, or $5.08 a share. Non-GAAP earnings for the second quarter were $2.37 a share. 

"We have more than 1 trillion workflows running through ServiceNow each year. Our trajectory is being supercharged by generative AI," said McDermott. 

McDermott said large deals for ServiceNow were evenly distributed across products in the second quarter and the company saw sustained demand.

Previously:

• ServiceNow outlines generative AI roadmap, bring your own LLMs

As for the outlook, ServiceNow projected third quarter subscription revenue growth of $2.18 billion to $2.19 billion, or about 26% growth from a year ago. For 2023, ServiceNow is projecting subscription revenue growth of $8.58 billion to $8.6 billion, or about 25% growth.

Constellation Research analyst Andy Thurai said:

"It is no surprise that ServiceNow is doing some generative AI use cases. What is good is that ServiceNow trained its own LLM instead of using one of the existing LLMs which would have raised a lot of security, privacy, tenancy, and data sharing questions. However, this will obviously bring the question of the accuracy of these models compared to the other existing high-performance models such as LLAMA and also the question of how often it will be retrained, etc. Assuming those issues are addressed and mitigated, I really like the case summarization use case. Summarization from documents is one of those cool low risk use cases where Generative AI can perform really well."

ServiceNow LLM is engineered to save time and speed up resolution time. Here are the key points:

• ServiceNow LLM is proprietary and part of the Now Assist family of generative AI features.
• Text-to-code converts natural language text prompts to executable code for the Now Platform.
• Case summarization streamlines essential information to simplify resolution processes and automate manual data entry.
• ServiceNow LLM was built on a specialized version of the 15 billion parameter StarCoder LLM.
• ServiceNow LLM was trained and tuned with Nvidia accelerated computing and DGX Cloud.

According to ServiceNow, case summarization and text-to-code are available to a limited set of customers and will be generally available in the ServiceNow Vancouver release in September.

AWS launches generative AI tools for developers

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

Amazon Web Services made its case for generative AI workloads starting with developers.

At AWS' New York Summit, Swami Sivasubramanian, VP of Databases, Analytics, and ML at AWS, outlined a bevy of generative AI updates to Amazon Bedrock, introduced new models and instances and talked vector data stores.

"Generative AI has captured our imagination. I believe it'll transform every industry and business," said Sivasubramanian, who stayed on point despite having his keynote interrupted a handful of times by protesters.

AWS New York Summit came hours after both Microsoft and Google reported earnings and executives talked about generative AI driving cloud demand. AWS has been relatively quiet about its generative AI story but has the infrastructure and enterprise heft to run those workloads.

"AWS is always about providing Lego building blocks for developers. AWS Bedrock offers model selection between multiple LLMs to developers. They are about developers and always have been. Always will be," said Constellation Research analyst Andy Thurai.

Here's a look at some of the AWS updates from New York.

• Amazon Bedrock gets new foundation models and fully managed agents. Amazon Bedrock, which is in preview, is a managed service for foundation models from the likes of Amazon, AI21 Labs, Anthropic, Cohere, and Stability AI for developers to use via API. Sivasubramanian also outlined a preview of agents for Amazon Bedrock so developers can automate the prompt engineering and orchestration of user requested tasks.
• Amazon EC2 P5 Instances with Nvidia H100 Tensor Core GPUs. AWS said Amazon EC2 P5 instances with Nvidia's H100 GPUs are generally available.
• AWS announced vector engine for Amazon OpenSearch Serverless, which provides an API for storing and querying billions of embeddings. AWS said all databases will have vector engines over time.
• AWS launched AWS HealthScribe, which is a HIPAA-compliant service that automatically generates clinical notes by transcribing and summarizing physician-patient conversations.
• Amazon QuickSight gets generative BI that allows users to ask questions about their data using natural language tools to create visuals, answers and calculations.

Tech M&A may perk up again

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

The market for tech mergers and acquisitions appears to be opening a bit—especially for smaller deals that round out capabilities for larger vendors.

Teradata said it has acquired Stemma, which provides a cloud data catalog platform. Stemma, founded in 2020, is best known for its machine learning that surfaces data and metadata for customers. Teradata said Stemma will broaden Teradata’s analytics engine for artificial intelligence and machine learning. Terms of the deal weren’t disclosed.

Meanwhile, Haveli Investments, a private equity firm, said it will acquire Certinia, which was previously known as FinancialForce. Haveli Investments bought Certinia from Advent International and Technology Crossover Ventures.

These smaller deals land as small- and mid-cap tech companies get scooped up. Maybe IBM's purchase of Apptio for $4.6 billion and Databricks $1.3 billion acquisition of MosaicML are just appetizers.

It’s wait-and-see for big deals, but things are starting to look up.

A US federal court shot down a request by the Federal Trade Commission to block Microsoft's deal to acquire Activision Blizzard. Meanwhile, the EU granted conditional approval to Broadcom's VMware purchase for $61 billion. These deals aren’t completed but headed in the right direction.

The jury is still out on M&A, but the music has stopped for startups looking for funding and those companies could be bargains for the right players.

Google Cloud revenue in Q2 tops YouTube ads

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

Google Cloud revenue in the second quarter topped $8 billion as parent Alphabet reported better-than-expected results and said CFO Ruth Porat will take on a new President and Chief Investment Officer role Sept. 1.

Porat was the longest serving CFO at Google. Alphabet's move to put Porat in charge of investments is notable. She will be responsible for Alphabet's "Other Bets" portfolio. Porat will also work with policymakers and regulators.

Alphabet reported second quarter revenue of $74.6 billion, up 7% from a year ago, with net income of $18.4 billion, or $1.44 a share.

Wall Street was expecting Google to report second quarter adjusted earnings of $1.34 a share on revenue of $72.82 billion, according to Refinitiv. Google Cloud was expected to deliver revenue of $7.87 billion, according to StreetAccount.

CEO Sundar Pichai said the company is “driving the next evolution of search and improving all of our services” with AI. Porat added that there was an acceleration of revenue growth in search and YouTube and momentum with Google Cloud.

Indeed, Google Cloud revenue in the second quarter topped YouTube ad sales of $7.66 billion. Google Cloud operating income was $395 million in the second quarter, compared to a loss of $590 million a year ago. 

Google has been busy this quarter adding capabilities to Bard, its generative AI technology, and sprinkling generative AI throughout its portfolio. Google Cloud also launched a series of AI-driven enhancements to its platform. While there will be multiple generative AI winners in the tech sector, Google vs. Microsoft, which is partnered with OpenAI, is the main card. It’s a battle that will take years to play out.

Here’s a look at some of the notable enterprise technology headlines from Google this quarter.

• Google Cloud lays out generative AI partnerships, Mayo Clinic deal, customer wins
• Google I/O 2023: Google makes its generative AI case, Bard becomes ingredient brand
• At Google I/O 2023, Google Cloud launches Duet AI, Vertex AI enhancements
• Google’s generative AI strategy from Google I/O 2023: Hits and misses

Alphabet ended the quarter with 181,798 employees, up from 174,014 a year ago. Alphabet said the majority of employees affected by workforce reductins are no longer included in the headcount totals.  

Speaking on a conference call with analysts, Pichai made the following points:

• User feedback to generative AI in search has been strong. "We see this new experience as another jump-off point for users to go deeper," he said.
• Google has sped up response time of Bard in search. Bard is also improving coding productivity.
• Ads will continue to play a role in the generative AI evolution of search. "We are testing and evolving placements and formats," said Pichai, who added AI is also being used in Google's ad products.Generative AI is being used to create ad formats. 
• Google Cloud is being optimized to work with multiple generative AI models. "We see continued growth," said Pichai. "Our AI infrastructure is optimized for training and serving generative AI models."
• Machine learning and AI are making data centers more efficient.
• Generative AI is driving Google Workspace demand and upsell opportunities to existing customers.

As for the outlook, Porat said the company sees stabilization in advertising revenue for YouTube, YouTube Music, YouTube Premium, search and hardware, which was fueled by the Pixel 7a launch in the second quarter.

Regarding cloud, Porat said the company is excited about demand for its AI offerings, but customers continue to optimize their spending.

Porat added that Google is optimizing its cost base through slower organic hiring, redeploying workers on more high value projects and improving efficiency. The largest capital expenditure was servers in the second quarter, but the rest of 2023 will feature more spending on data centers, GPUs and TPUs.