Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

OpenAI launched a new benchmark that grades large language models (LLMs) on real-world work tasks and enterprises need to take note as they ponder AI agents.

OpenAI unveiled GDPval, a system that grades LLMs on tasks that humans currently do. Yes, we know (since vendors tell us repeatedly) that AI is collaborative with humans and not a replacement. But if you were to look to AI as a human labor replacement, OpenAI's GDPval is likely to be handy.

In a blog post, OpenAI noted that GDPval is "a new evaluation designed to help us track how well our models and others perform on economically valuable, real-world tasks."

Why GDPval? OpenAI started with Gross Domestic Product (GDP) as an economic indicator and took tasks from the occupations and sectors that contributed the most to GDP.

Here's a look at the occupations and tasks in GDPval.

The win for enterprises is that CxOs can use GDPval to better align models for use cases. The win for the rest of us is that we can now compare models on real-world tasks instead of math exams and other benchmarks that are abstract for most people.

Based on GDPval's topline figures, Anthropic's Claude Opus 4.1 is the leader for work tasks followed by GPT-5.

OpenAI said:

"People often speculate about AI’s broader impact on society, but the clearest way to understand its potential is by looking at what models are already capable of doing. History shows that major technologies—from the internet to smartphones—took more than a decade to go from invention to widespread adoption. Evaluations like GDPval help ground conversations about future AI improvements in evidence rather than guesswork and can help us track model improvement over time."

As for returns, OpenAI also noted:

"We found that frontier models can complete GDPval tasks roughly 100x faster and 100x cheaper than industry experts. However, these figures reflect pure model inference time and API billing rates, and therefore do not capture the human oversight, iteration, and integration steps required in real workplace settings to use our models. Still, especially on the subset of tasks where models are particularly strong, we expect that giving a task to a model before trying it with a human would save time and money."

A few thoughts on how CxOs may approach GDPval:

• GDPval can make it easier to compare digital and human labor costs. For instance, a model that can deliver good work in one shot is more beneficial than one that requires a lot of back-and-forth since that drives compute costs up.
• OpenAI's GDPval paper also includes failure modes and the reasons why. Failure rates are going to be critical for proper evaluation.

• The benchmark also provides an opportunity to think about workflows and processes before humans work on the task. OpenAI's point about using AI to get a task partly to the finish line is valid. However, it's also worth noting what Harvard Business Review just reported on sloppy AI work.
• Humans in the loop during a process is probably the most important decision to make in using AI to automate processes. GDPval gives you a jumping off point for discussion.

• Accenture: Enterprise AI deployments hit inflection point
• AI agents, automation, process mining starting to converge
• Cognizant CEO Kumar: Agentic AI will create a reinforcing flywheel of ROI
• Pondering the future of enterprise software
• Lessons from early AI agent efforts so far
• Every vendor wants to be your AI agent orchestrator: Here's how you pick

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

Accenture CEO Julie Sweet said the company is seeing an inflection point with companies adopting artificial intelligence enterprise-wide and scaling use cases.

Speaking on the company's fourth quarter earnings call, Sweet said:

"We're also starting to see early signals of an inflection point with more clients looking for true enterprise-wide plans and activation and seeking out our successful experience with scaling in enterprises and at Accenture. Two years into this AI journey, we also are seeing a pattern in how AI can expand our opportunities with our clients."

Sweet said many of these AI projects also require AI readiness, which means more transformational work. Those chores include data modernization to go along with digital operations and cloud deployments. She cited a financial services client that is using Accenture to modernize the data estate, but that also requires retiring legacy systems and more foundational work.

"We're seeing more stories like this across our portfolio, where AI is extending across the enterprise and adjacent work is following," said Sweet. "Building the digital core remains our biggest growth driver."

Banking is one of the main industries revamping the data stack. Sweet added that Ecolab is redesigning processes and then scaling into AI agents. "Ecolab is on a path to deliver an estimated 5% to 7% sales growth and 20% operating income margin without increasing costs at the same pace," said Sweet.

• Accenture's Karthik Narain on human, AI collaboration, trust
• A tour of enterprise tech inflection points
• AI Exponentials Will Rule the World
• Radically Rethink Old Models To Enable the Power of AI
• AI agents, automation, process mining starting to converge
• Agentic AI: Everything that’s still missing to scale

Accenture reported fourth quarter earnings of $2.25 a share on revenue of $17.6 billion, up 7% from a year ago. Accenture recorded new bookings of $21.3 billion in the fourth quarter and $1.8 billion of that sum was generative AI. GenAI bookings doubled for fiscal 2025 to $5.9 billion and revenue tripled to $2.7 billion.

For fiscal 2025, Accenture reported earnings of $12.15 a share on revenue of $69.7 billion.

As for the outlook, Accenture said it expects fiscal 2026 revenue growth of 2% to 5% in local currency. Earnings will grow at a 9% to 12% clip.

Other takeaways from Accenture's fourth quarter:

• Sweet added that many of Accenture's customers are enterprises that tried do-it-yourself AI but ran into roadblocks scaling projects. "We've had lots of clients who have started things on their own and then come to us who've got good proof of concept that their team was able to do but then just can't scale it," said Sweet.
• AI savings are being reinvested. "AI absolutely boosts efficiency in areas like coding or operations. But those savings don't disappear. They're being reinvested into new priorities. The list of what our clients want to do with technology is truly virtually unlimited. And so, when we can save them money by delivering our services with advanced AI, that frees up their budget to do the next things on their list," said Sweet.
• Companies are held back by change management and process reinvention.
• AI strategy includes growth and savings. "Almost every CEO that I've talked to says they pivoted way too far towards productivity and not enough to growth," said Sweet.

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

Databricks said OpenAI's foundational models will be available in the Databricks Data Intelligence Platform and Agent Bricks natively in a partnership worth $100 million.

The deal highlights how OpenAI is expanding the distribution for its ChatGPT family of models beyond Microsoft and direct access.

In recent weeks, OpenAI has become available on Oracle Cloud Infrastructure. The company's open weight models are now available on AWS. Snowflake added OpenAI models via an expanded partnership with Microsoft.

Databricks said OpenAI models, including GPT-5, are available to more than 20,000 customers. GPT-5 will also be the flagship model for all Databricks customers.

• Databricks valued at $100 billion
• Databricks natively integrates Google Cloud Gemini models
• Databricks launches Mosaic Agent Bricks, Lakeflow Designer, Lakehouse

By offering OpenAI models natively, Databricks customers will be able to build AI agents closer to where data lives with no extra movement.

Key points about the Databricks-OpenAI deal:

• Databricks can get OpenAI models via SQL or API.
• Databricks customers have access to high-capacity processing across the latest OpenAI models.
• Agent Bricks will tune and optimize GPT-4 and gpt-oss for accuracy.
• Databricks Unity Catalog will provide governance and observability for OpenAI models.
• The two companies will optimize OpenAI models for enterprise use case.

Databricks said GPT-5, GPT-5 mini and GPT-5 nano will be available natively in Databricks across AWS, Microsoft Azure and Google Cloud in the near future.

Holger Mueller, an analyst at Constellation Research, said the OpenAI-Databricks partnership appears to be a win-win situation. 

"This partnership emancipates Databricks from the cloud providers who have already partnered with OpenAI - and takes a differentiator for the cloud vendors away. Cloud vendors' data lake houses work seamlessly with their AI frameworks. CxOs now have options on how to build their AI powered next-gen apps.

If Databricks plays this well it becomes the multi cloud data and LLM foundation for enterprises. Being multi cloud has ways paid off at any level of the stack - and there is no reason for it not to work for Databricks. What is unique is that this touches more layers of the stack."

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

The European Commission has formally opened an investigation of SAP and its maintenance and support practices for on-premises deployments in Europe.

In a statement, the EC said it has started an investigation into whether SAP "may have distorted competition in the aftermarket for maintenance and support services" for its on-premises ERP applications.

The EC investigation coincides with ongoing SAP lawsuits with Celonis and Teradata.

Specifically, the EC is looking into four areas:

• SAP's requirement that customers seek maintenance and support from the company for on-premises ERP under the same pricing for all software and preventing enterprises from mixing and matching services from other suppliers.
• Preventing customers from terminating maintenance and support for unused software licenses.
• SAP systematically extending the duration of the initial term of on-premises ERP licenses so customers can't terminate maintenance and support.
• SAP charging back-maintenance fees to customers that subscribe to SAP maintenance and support after a period of absence.

For its part, SAP confirmed the investigation and said:

"These proceedings address some areas of our on-premise maintenance and support policies, which are based on long-standing standards that are common across the global software sector. SAP believes that its policies and actions are fully in line with competition rules. However, we take the issues raised seriously and we are working closely with the EU Commission to resolve them.

We do not anticipate the engagement with the European Commission to result in material impacts on our financial performance."

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

Hitachi Vantara is embedding AI agents throughout its storage systems, seeing customers embrace more hybrid cloud and on-premise AI architectures and betting on AI at the edge and sovereign AI as growth markets.

Those are some of the takeaways from Octavian Tanase, Chief Product Officer at Hitachi Vantara.

Constellation Insights caught up with Tanase at Hitachi Vantara's Analyst Live 2025 event in Arlington, VA. At the analyst meeting, Hitachi Vantara highlighted its strategy, customer enterprise AI and industrial AI use cases, hybrid cloud and AI efforts, collaboration with partners including Cisco and Nvidia and how its storage and data platform is leveraging the One Hitachi strategy. Here's a look at the takeaways from the conversation with Tanase.

Enterprise AI adoption. "We see a lot of demand from enterprises looking to get insights out of that data, and use AI to improve productivity," he said. "There's a rush right now to build autonomous modules that will take a business workflow and solve and anticipate problems in an enterprise. Customers are looking to bring in a large language model and train and fine tune with enterprise data before they deploy for inference."

Integrated software with hardware. "We've always seen ourselves as a systems company that builds both software and systems for storage and data management. Storage has been commoditized, and there is more value that can be delivered in software to use not only the data for an application, but to run analytics on that data to get more insights from the data. This is an area of investment for us," said Tanase.

Hitachi Vantara's agentic AI strategy. Tanase said Hitachi Vantara is using AI agents internally for productivity and within the products leveraging it for more autonomy. "We are in the business of providing infrastructure for AI data pipelines that include storage, compute, networking, security and so forth. We are embedding capabilities around the data reduction or data tiering or data classification. These are all areas where one could create an agent and transform a task that required the control and the input of a person into something autonomous," said Tanase.

Customer use cases. "We see a lot of use cases around analytics. A lot of times people will make two or three copies of data. Enterprises are looking to run analytics on data and use AI to do that and coordinate large data sets from heterogeneous data sources," said Tanase.

On-premise and hybrid AI evolution. Tanase said most enterprises started with AI in the cloud because GPUs-as-a-service doesn't require a massive initial investment. What's happening now is enterprises are looking to put AI infrastructure closer to the data. "If the data source is in the traditional data center or being brought from an edge device, enterprises are building AI infrastructure where the data is," said Tanase. "It's too expensive to correlate multiple data silos and move that to the cloud. Customers are sometimes better off building a data lake into their traditional enterprise and then deploying training and inference closer to the data."

Edge computing's importance. "I am a firm believer that there is more data being created at the edge and in the cloud than the data center," said Tanase. "AI will give the power to analyze the data as its created and perhaps enable customers to become more discerning about their data and understand what they need to keep and protect. I'm hoping many of these capabilities in the future are autonomous."

The product roadmap. "Going forward, AI is fundamentally changing everything. The market is moving fast and the standardization of MCP (model context protocol) and other protocols are enabling AI modules to talk to each other," said Tanase. "In order to be relevant in this market, you have to act with agility in a way many companies have not experienced before. Time to market, constant innovation and the reality that no one vendor can do it all are critical."

Tanase added that customers will see integrated systems from Hitachi Vantara and a wide range of natural partners including Nvidia, Cisco, Supermicro, Hammerspace and Commvault to name a few.

Sovereign AI. Tanase said sovereign AI infrastructure is becoming a big market. "AI has become a matter of national security for many countries or provinces, and there is a lot of need to integrate and build sovereign AI," said Tanase. "We live in a very polarized world, and I can see states, governments, and provinces building sovereign AI infrastructure. It's a part of what everybody does in order to compete in the 21st century."

Final word. Tanase said Hitachi Vantara has earned the right to play in the AI space and is being used for critical business applications leveraging structured and unstructured data via VSP 360 and a wide range of systems. "We know our customers. We want to save them time. We invest a lot of tools to enable automation, and we believe that's critical, because people want repeatable results," he said. "Automation is top of mind. We're a leader in infrastructure sustainability and sustainable products save customers money in terms of floor in the data center, cooling power and overall cost of ownership."

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

SAP said it has launched its sovereign cloud offerings on AWS European Sovereign Cloud and inked a deal that brings OpenAI to Germany's public sector customers.

The partnerships revolving around sovereign cloud in Europe include multiple clouds. With AWS, SAP Sovereign Cloud apps with security and regulatory compliance will run on AWS European Sovereign Cloud, a new independent cloud. AWS has said it will invest €7.8 billion in the EU.

As for the OpenAI deal, SAP said OpenAI for Germany will be available on SAP's Delos Cloud, which runs on Microsoft Azure. OpenAI and SAP will combine large language models with SAP apps for the public sector.

Last month, SAP said it was expanding its sovereign cloud efforts in the EU via cloud and on-premises.

Here are the key details for SAP's AWS and OpenAI partnerships.

AWS and SAP

• The AWS European Sovereign Cloud is set to launch its first AWS region in Brandenburg, Germany by the end of 2025.
• The AWS region for public sector customers and highly regulated enterprises will feature data residency, operational autonomy and resilience.
• AWS European Sovereign Cloud is separate and independent from other AWS regions and won't depend on non-EU infrastructure.
• SAP Sovereign Cloud is available on AWS in Australia, New Zealand, UK, Canada and India already.
• SAP Sovereign Cloud on AWS European Sovereign Cloud will initially include SAP Business Technology Platform and SAP Cloud ERP and expand from there.

OpenAI and SAP

• OpenAI for Germany will be supported by Delos Cloud.
• SAP, OpenAI and Microsoft will launch their collaboration in 2026.
• SAP and Microsoft will focus on improving productivity for German government employees with OpenAI for Germany.
• Ultimately, The partners plan to integrate AI agents directly into workflows and automate processes.
• SAP will expand Delos Cloud in Germany to 4,000 GPUs for AI workloads. SAP will invest more on AI infrastructure in Germany based on demand.

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

Microsoft said it will add Anthropic's Claude Sonnet 4 and Claude Opus 4.1 to Microsoft Copilot Studio and Researcher in a move that highlights how the company is diversifying from OpenAI.

Until recently, Anthropic was more of a AWS and Google Cloud play with Microsoft serving as the venue for OpenAI models.

In a blog post, Microsoft said Anthropic models will be available in Microsoft 365 Copilot in the Research agent, which now can be powered by OpenAI or Anthropic. In addition, Claude Sonnet 4 and Claude Opus 4.1 models are available in Copilot Studio.

With the Copilot Studio addition, enterprises will be able to create AI agents using Anthropic models.

OpenAI and Microsoft have been putting a little distance between themselves. For instance, OpenAI launched open weight models that can be used beyond Microsoft Azure. The two companies have also come to an understanding about Microsoft's equity stake and OpenAI's structure.

And OpenAI is also building its own AI infrastructure via a partnership with Oracle.

More on LLMs:

• Stargate ramps as OpenAI, Oracle, Softbank outline 5 new data centers
• Nvidia to invest $100 billion in OpenAI, which will deploy Nvidia next-gen AI infrastructure
• OpenAI, Microsoft reach an understanding of something as AI harmony awaits
• OpenAI vs. Microsoft: Why a breakup could be good
• OpenAI and Microsoft: Symbiotic or future frenemies?
• Anthropic lands $183 billion valuation, hits $5 billion annual revenue run rate

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

Qualcomm laid out a vision where AI workloads are hybrid between the cloud and edge devices such as smart glasses, smartphones and wearable devices. The conduit for these AI workloads will ultimately be 6G.

At the Snapdragon Summit, Qualcomm CEO Cristiano Amon said AI will remake every device you wear, come complete with AI agents and be proactive with you in real time.

For Amon, the Qualcomm vision was partially about talking up Snapdragon and outlining its role going forward. In the big picture, Qualcomm's take on edge devices handling a big chunk of the AI workload is notable. Why? The current AI thinking revolves around brute force compute, billions if not trillions of dollars spent on AI data centers and cloud delivery.

"We envision AI to be both cloud and edge. The edge complements the cloud. It’s immediate. It’s personal. It has context. And there’s one important thing about the edge: that’s where the data originates and then where AI becomes yours," said Amon.

Amon argued there will be a new compute architecture for AI that is cloud and edge. He also noted that foundational models are already designed to be in the cloud and edge and will ultimately be the UI.

• Qualcomm acquires Alphawave for $2.4 billion
• Qualcomm fleshes out Dragonwing with fixed wireless access platform

In addition, the data collected at the edge will be more critical than what was used to train models. This edge AI data will be what personalizes the experience, acts on your behalf and has context.

Amon was talking about personal computing experiences, but it isn't much of a leap to understand how sensors at the edge are going to impact enterprise use cases too. "The importance of edge data is massive. It’s the best kept secret," said Amon.

The conduit for these cloud to edge hybrid AI workloads will be the network in between. Amon touted 6G networks and pre-commercial devices ready as early as 2028.

"6G is designed to be the connection between the cloud and the edge devices. The difference between 5G and 6G is the network of intelligence connecting the edge and the cloud, merging the physical and the digital, providing connected experiences,” said Amon.

A big part of the AI at the edge strategy revolves around Snapdragon and the chips on devices.

Snapdragon 8 Elite Gen 5

Qualcomm announced the 3rd Gen Qualcomm Oryon CPU and Snapdragon 8 Elite Gen 5, which is billed as the fastest mobile system-on-a-chip.

Key points:

• Snapdragon 8 Elite Gen 5 will enable fast multitasking and app switching as well as long-game play.
• The platform enables AI agents to work across apps via continuous on-device learning and real time sensing and multi-modal models.
• Snapdragon 8 Elite Gen 5 has performance gains across the Oryon CPU, a 20% performance boost, Qualcomm Adreno GPU with a 23% boost for gaming, and Qualcomm Hexagon NPU with 37% faster performance.

Snapdragon X2 Elite Extreme and Snapdragon X2 Elite

Qualcomm rolled out new processors for Windows 11 PCs and can deliver 80 TOPS of AI processing. It's also the first Arm processor to run at 5 GHz.

Snapdragon X2 Elite Extreme is designed for premium PCs and is aimed at agentic AI, data analytics and professional media editing.

Devices powered by the Snapdragon X2 Elite family will be available in the first half of 2026.

The launch of Qualcomm’s next-gen processors indicates a move upmarket to target creative pros and data engineers.

Larry Dignan

Editor in Chief of Constellation Insights Constellation Research

About Larry Dignan: Dignan was most recently Celonis Media’s Editor-in-Chief, where he sat at the intersection of media and marketing. He is the former Editor-in-Chief of ZDNet and has covered the technology industry and transformation trends for more than two decades, publishing articles in CNET, Knowledge@Wharton, Wall Street Week, Interactive Week, The New York Times, and Financial Planning. He is also an Adjunct Professor at Temple University and a member of the Advisory Board for The Fox Business School's Institute of Business and Information Technology.&nbsp; <br>Constellation Insights does the following: Cover the buy side and sell side of enterprise tech with news, analysis, profiles, interviews, and event coverage of vendors, as well as Constellation Research's community and… Read more

OpenAI, Oracle and Softbank have announced 5 new data centers under the Stargate project good for 7 gigawatts of capacity. The data centers will be largely powered by Nvidia-based infrastructure.

The companies said the flagship site in Abilene, Texas is operational and can deliver 5.5 gigawatts of capacity built on Oracle Cloud Infrastructure (OCI) and Nvidia's stack. Combined with projects with CoreWeave, Stargate now has 7 gigawatts of planned capacity and more than $400 billion in investment over the next three years.

According to OpenAI, Stargate, announced in January, is well on its way toward securing $500 million and 10-gigawatts committed by the end of 2025.

The AI infrastructure boom is all about a chase for consumer AI with a dash of enterprise, high-performance and scientific workloads. In this week's saga, OpenAI entered a deal with Nvidia for $100 billion in funding where the LLM player builds data centers (on Nvidia GPUs) and then gets paid for every gigawatt deployed. Sam Altman hinted at big plans.

In a nutshell, consumer AI is more like covering a sports story. OpenAI pledges $300 billion to buy AI infrastructure from Oracle. Oracle buys GPUs. Nvidia backstops CoreWeave with purchase guarantees if it has extra capacity. OpenAI also appears to be a big future buyer of Broadcom XPUs (which probably led to the Nvidia deal). Meanwhile, Amazon, Google and AWS all have to buy Nvidia but building their own custom chips for AI workloads.

The Stargate project is the big reason why Oracle's remaining performance obligation surged in its latest quarter. OpenAI and Oracle partnered on up to 4.5 gigawatts of additional Stargate capacity.

Over the next five years, Oracle and OpenAI will develop three sites in Shackelford County, Texas; Doña Ana County, New Mexico; and a site in the Midwest.

In addition to the flagship Stargate site in Abilene, two separate sites will be developed over the next 18 months. Stargate sites in Lordstown, Ohio and another in Milam County, Texas can scale to 1.5 gigawatts of capacity. "The AI race is on and the unit of measure is gigawatts going live. At launch Stargate was questioned as a viable partnership, but it's clear now that they are at the forefront of getting AI capacity online," said Constellation Research analyst Holger Mueller.

According to a report by Bain & Co., $2 trillion in annual global revenue is needed to fund the computing power needed to meet anticipated AI demand by 2030. Bain estimated that even with anticipated AI savings, the world is $800 billion in short to keep pace with demand. 

In an earlier blog post, OpenAI CEO Sam Altman said the time to build is now. He said:

"Our vision is simple: we want to create a factory that can produce a gigawatt of new AI infrastructure every week. The execution of this will be extremely difficult; it will take us years to get to this milestone and it will require innovation at every level of the stack, from chips to power to building to robotics. But we have been hard at work on this and believe it is possible."