Energy storage firm Base Power commenced sales of its large home battery systems to residents of Illinois yesterday, according to Canary Media. Significantly, this marks the startup’s initial venture into the grid zone managed by PJM Interconnection, the largest grid operator in the U.S. by area, which has notably had challenges adapting to the surge in new data centers.

In addition to Illinois, PJM’s jurisdiction includes Northern Virginia, one of the world’s most concentrated data center hubs. This concentration, combined with a lack of new generating sources, has nearly doubled wholesale electricity prices in PJM over the last year. The situation has deteriorated so severely that AEP, one of the largest utilities in the area, has threatened to exit the market.

Base Power was established two years ago in Texas to develop a virtual power plant centered around residential batteries. Base’s batteries, beginning at 25 kilowatt-hours, surpass those of many competitors, and instead of selling the batteries, it requires clients to purchase electricity from it. In Illinois, its prices are 25% lower than those of utility ComEd.

The timing of the startup’s initiatives has been impeccable. Base is presently operating over 500 megawatt-hours of battery storage in Texas, charging when electricity prices are low and disbursing them when the grid is under the most strain. 

Its entry into the PJM grid occurs at a moment when the operator has faced criticism for mishandling the increase in electricity demand. PJM halted applications for new generating sources starting in 2022, only re-opening the queue in April. Unlike Base, its timing couldn’t have been worse — electricity demand has surged in the past four years.

Base’s expansion has accelerated since October, when it publicized a $1 billion funding round led by Addition. This round closely followed a $200 million round spearheaded by Andreessen Horowitz, Lightspeed Venture Partners, and Valor Equity Partners in April 2025.

Traditionally, PJM has been sluggish to embrace new technologies like distributed energy storage, but Base’s focus on residential customers allows it to circumvent the sluggish grid operator.

“We are deploying capacity behind the meter at the residential home, where an interconnection already exists, so we don’t wait in the interconnection queue,” Zach Dell, founder and CEO of Base Power, stated to Canary Media.

Data from Indagari, a credit card transaction analysis firm, indicates that more consumers opting for AI services are selecting Anthropic’s Claude.

This suggests that the AI lab boasts a more extensive and robust customer base than the specific demographic it is typically assumed to cater to — developers from enterprises and startups utilizing Claude Code.

Indagari reviews billions of anonymized credit card transactions from approximately 28 million consumers in the U.S. While this data may not offer precise figures regarding Anthropic’s revenue or customer count, it is sufficiently large to reveal trends.

And in the case of Anthropic, the trend is clearly positive.

The examined data encompasses weekly transactions from 2025 to May 10, 2026, covering payments for various items, including subscriptions and API tokens. It reveals a month-over-month increase in the number of paying customers and revenue for Claude, now showing approximately a 75% growth since January 2026 in this sector.

Notably, growth persisted even after a surge in consumer interest in March, when the company declared that it would not permit its models to be used by the Trump administration for mass surveillance or autonomous weaponry.

An additional sign of Claude’s rising consumer appeal stems from DataCamp, an online education platform that imparts AI skills to users and business personnel, claiming around 20 million users.

Since the beginning of the year, consumer interest in Claude on DataCamp has surged. The term “Claude” is now the most frequently searched on its platform, surpassing even “AI,” as reported by DataCamp to TechCrunch.

Although ChatGPT courses continue to dominate corporate training, self-guided consumers are showing a threefold greater demand for Claude courses compared to ChatGPT, according to the company. Demand for Claude courses has skyrocketed 18 times within the last 30 days alone, it states.

Nevertheless, despite Claude’s remarkable progress, ChatGPT remains the far more popular AI option among consumers in every aspect.

For example, recent findings from market intelligence firm Sensor Tower indicate that Claude is experiencing growth this year across various platforms, but still lags significantly behind ChatGPT.

Although the data suggests that ChatGPT’s growth has been somewhat slower recently (mainly due to its expansive existing user base), it still boasts a significantly larger number of paying users, according to Indagari’s analysis.

However, it is evident that Claude has begun to close the gap with ChatGPT this year in terms of revenue generated from consumers, as well as overall consumer recognition and enthusiasm.

As OpenAI and Anthropic approach the possibility of going public, there’s a strong interest in understanding the foundations supporting their business models.

It remains especially unclear how Anthropic’s recent confrontations with the U.S. government will affect its operations. Earlier this month, the government prohibited Anthropic from deploying its most advanced cybersecurity-oriented models, Mythos 5 and Fable 5, for use by non-U.S. citizens, prompting the AI lab to withdraw these models from the market for the time being.

Nevertheless, every indication we can gather suggests that Anthropic continues to expand its user base, both among consumers and in the business/enterprise sectors.

Anthropic has chosen not to provide a comment.

Upon stepping into the R&D area of General Intuition’s New York office, the company’s 31-year-old co-founder and CEO Pim de Witte pointed out a monitor positioned on a standing desk. It seemed as though someone was engaged in playing a game similar to Fortnite. However, it wasn’t a human.

“Our agent has been gaming for 100 consecutive hours,” Kent Rollins, the chief product officer, stated with a smile.

Before I could become captivated by the sight of an AI maneuvering through the game’s digital landscape, I heard the mechanical footsteps of a substantial quadrupedal robot approaching.

“The same intelligence that powers the agent playing the game is also driving the robot,” de Witte informed me.

Josh Duplantis, a data analyst equipped with a laptop that was streaming a live feed from the robot’s sole camera, interjected to clarify that the robot’s primary operation mode was “exploration.”

Utilizing that camera, its singular vision, the enormous insect-like robot approached me, orbited around me, and ventured further into the office. It occasionally bumped into chair legs or collided with an errant wastebasket, reminiscent of a toddler still discovering how to navigate their surroundings. Duplantis mentioned that merely eight minutes of physical robotics data were required to optimize an AI model for the quadruped. Additionally, that data had been collected outdoors, rather than within the office where the robot was autonomously exploring.

An agentic model capable of generalizing from gaming to simulation to embodiment is the central mission of General Intuition. The competence of that model in discerning its position in the environment has attracted the interest of significant investors.

On Thursday, General Intuition announced it had secured $320 million at a $2.3 billion valuation, validating prior reports from TechCrunch. This funding round brings the startup’s total disclosed capital to $454 million, following the $134 million raised at its inception last October.

The startup originated from de Witte’s previous venture, Medal, which permits gamers to upload and share video game clips. The hundreds of millions of hours of gameplay uploaded formed the foundational dataset used to educate General Intuition’s model in spatial-temporal reasoning — or comprehending movement through both space and time.

However, the pivotal element wasn’t the gameplay footage itself; it was the action labels integrated within those videos: detailed records of the exact buttons a player pressed and at what times. According to de Witte, most competitors strive to deduce actions solely from video data, which he contends is inadequate.

“We see this as merely a transition to the next phase of future pre-training,” de Witte remarked. “We possess a singular model that can react to Fortnite visuals on the screen and take action, in addition to real-world dynamics in a manner that an LLM could never achieve.”

At one point, de Witte connected me to a laptop running General Intuition’s world model, a simulation built frame-by-frame rather than rendered using a conventional game engine. As is usual for me when experimenting with world models, I walked directly into a series of walls. In other demonstrations I’ve encountered, the agents often pass through barriers, but this particular model did not. From the millions of hours of gameplay, it had somehow discerned that walls are solid, ladders serve a climbing purpose, and that shadows lengthen as the sun progresses.

For General Intuition, this world model is not the product; it serves as the training space (internally known as “the gym”). The company ultimately seeks to market the agentic model itself, and de Witte maintains that the action data embedded in gaming aids the model in distinguishing the “self” from the “environment” in a way that enriches its understanding of causality.

Although General Intuition’s technology appears impressive in demonstrations, it is not the sole entity attempting to solve this challenge. Furthermore, achieving such a model that performs reliably in the physical world at scale has yet to be accomplished. Most methods of this nature demand vast quantities of real-world data, which are typically gathered gradually and at high costs. General Intuition’s premise is that gameplay provides a scalable shortcut.

Investors are supportive of this premise as well. The latest funding round was spearheaded by Khosla Ventures, with backing from General Catalyst, Jeff Bezos, Eric Schmidt, Nico Rosberg, as well as researchers from Google DeepMind and MIT.

The majority of the funds from this round will be allocated to enhancing computational capabilities. General Intuition has a partnership with CoreWeave and aims to concentrate on pre-training the next iteration of its model. A portion has been set aside to broaden access to its API by the end of summer.

Vinod Khosla, whose firm led the funding round, expressed that he was attracted to de Witte’s vision and the company’s proprietary data advantages.

“If one examines LLMs, the onset of reasoning was a significant breakthrough,” Khosla shared with me during a phone conversation. “In the realm of world models, I believe the major breakthrough is the emergence of intuition within AI, akin to human intuition. The human action and reaction data present in games is crucial for this intuitiveness to develop.”

The vision is a generational company

General Intuition is not the only enterprise to recognize that Medal’s human action data is crucial in constructing dynamic world models and general agents. Brianna Martin, the startup’s chief of staff, mentioned that the company was partially founded after Medal declined an acquisition proposal from a prominent laboratory. Additionally, there have been other offers following that.

De Witte and his co-founders, Eloi Alonso, Adam Jelley, and Vincent Micheli, are not inclined towards being acquired, nor are the startup’s investors seeking an exit at this moment. The volume and quality of proprietary data General Intuition possesses through Medal is among the reasons Khosla believes the startup represents a generational investment rather than an M&A prospect; that it could evolve into the foundational platform for generalized agents and world models in both simulation and real-world applications.

“At this stage, it would merely constitute a data acquisition, which lacks excitement,” Khosla stated.

Part of that belief also entails faith in de Witte’s principles.

The entrepreneur dedicated three years working in the humanitarian field, including time with Doctors Without Borders. Consequently, he has established a clear boundary regarding the application of General Intuition’s technology: No agents will be utilized in any harmful way towards humans.

“We don’t intend to be a part of an escalatory system,” de Witte mentioned. “Hypothetically, if I were to announce, ‘We’re developing lethal autonomy,’ what do you suppose would be the reaction from other nations?”

This restriction on military applications comes at a time when Silicon Valley is increasingly enthusiastic about warfare, although de Witte states he is open to his models being employed for search and rescue operations.

De Witte hails from the Netherlands, and a significant portion of his team is European, which influences the identity of the company. He noted that he recruited Martin partly due to her choice to publicly resign from Palantir because of its collaboration with the United States Immigration and Customs Enforcement.

“I can’t comprehend the motivations behind Silicon Valley’s actions,” he remarked. “There’s a reason I’m not situated there.”

De Witte’s ethical stance doesn’t solely restrict the actions his models will avoid. As a gamer who earned $1.5 million by creating and managing a private RuneScape server during his adolescence, de Witte also considers the fate of individuals who might be marginalized by the advancements AI models can achieve.

Recently, General Intuition introduced a platform named Nerve, a job marketplace that allows gamers to earn income using their existing setups. Those who register start with data labeling tasks and can eventually progress toward robot teleoperation and other responsibilities. De Witte pointed out that Medal’s user community comprises the very generation most vulnerable to AI-induced job displacement, and he aims for them to have a stake in what the future holds.

A data flywheel

De Witte aspires for General Intuition to function as an ecosystem facilitator, akin to Anthropic or OpenAI — a model provider that empowers others to build upon its technology. As of now, the startup has several clients in gaming, simulation, and robotics.

“We’re not planning to establish a self-driving car company,” de Witte stated. “Our goal is to simplify the process for the next entity to create a self-driving car company significantly.”

The business asserts that once it makes its API accessible to a wider array of clients, it will be able to test its capabilities across various use cases — such as trialing a robot in a digital twin of a factory floor, powering a lifelike bot within a gaming studio, or deploying a quadruped to navigate perilous environments.

Though the quadruped represents the first tangible embodiment that General Intuition has tested in the real world, it has also experimented with drones and other devices, including assessing the model in driving simulations.

“It functions on any platform controlled via a game controller or a mouse and keyboard,” de Witte remarked.

The ambition to create a data flywheel is one of the objectives.

“We’ll select clients where we can diversify the embodiments that this generalized foundation model serves as the core for,” de Witte stated. “We’re going to prioritize choosing clients based on their ability to provide real-world data that will be intriguing and beneficial for advancing research. Moreover, we want to work with agile internal teams where we can function as genuine embedded partners and learn from one another.”

Khosla stated that General Intuition’s proprietary data has facilitated its progress up to this point, and its capability to continue gathering unique data will be crucial. Especially since, despite remarkable demonstrations, whether the transfer from simulation to the real world can sustain itself at scale remains a question that has not yet been fully addressed.

Correction: The headline previously misrepresented the amount raised by General Intuition in this round. The mistake has been amended.

The pursuit of unearthing the next breakthrough in AI has funded numerous ambitious initiatives — yet one company views this as an opportunity to reconstruct computing architecture from scratch.

Under the leadership of Naveen Rao, who previously headed AI at Databricks, Unconventional AI aims to enhance inference processing to be significantly more energy-efficient. The unique feature: a novel oscillator-based computer architecture.

On Thursday, the firm unveiled its initial AI model — named Un-0 — a tool for image generation that demonstrates for the first time how the company’s technology can emulate traditional AI systems. In a correlating new publication, the firm’s research team elaborates on how they developed a fully operational image-generation model utilizing a software simulation of the new architecture — one that performs comparably to cutting-edge diffusion models.

“This is the ‘hello world’ of a new type of computer,” Rao stated to TechCrunch. “Over the next year, you’re going to start witnessing some fascinating news around this.”

The production from the fresh Un-0 model resembles that of image-generation frameworks like Stable Diffusion or OpenAI’s GPT Image 1. What stands out is how it achieves that level of performance. The model relies on an oscillator-based architecture that diverges entirely from the chips that drive conventional computing and traditional LLMs. The advantages of the oscillator-based computing are intricate, but Rao is convinced it will ultimately cut energy consumption by as much as 1,000-fold.

Much of the infrastructure required for this goal is still under development. The existing version of Un-0 operates on a software simulation of Unconventional’s oscillator chips, but the company intends to share designs for a tangible chip soon. From that point, the strategy is to construct a complete inference stack from the ground up, with Unconventional AI eventually providing computing capacity just like any other supplier.

“We will create a new type of system made up of our chips,” says Rao. “We will execute AI models there, and we will have a network cable where prompts are inputted and inferences outputted, but it’ll be accomplished at 1/1000 of the power.”

It’s an extraordinarily ambitious objective, especially for a company with fewer than 50 employees. However, considering the magnitude of the AI expansion and the expected costs of satisfying the rising demand for inference, it may represent one of the few initiatives capable of addressing the scale of the challenge. As Rao perceives it, the available energy supply will be one of the crucial limits for AI in the coming years — and Unconventional is among the rare projects that can tackle it.

“AI scaling is challenging due to energy. It will be the fundamental constraint in the next several years. You simply can’t surpass it. Ultimately, it’s going to be an energy-constrained issue,” he remarks.

Market research firm Klue, which experienced a hack this month resulting in the theft of significant data belonging to various customers, has stated that it is in contact with the cybercriminals. The company has indicated its belief that the group is in the process of erasing the stolen information, as reported by TechCrunch. 

“We are still in talks with the threat actor we have been engaging with (‘Icarus’),” the organization mentioned in a privately shared update with its clients on Thursday evening, which TechCrunch has accessed. “Icarus informed us that they are initiating actions to erase the data obtained from Klue clients. The Icarus website is still offline, and we have signs that Icarus is genuinely proceeding to eliminate data taken from Klue customers.”

On Monday, Klue verified that intruders accessed its systems on June 12, leading to the theft of an undisclosed quantity of data from an unspecified number of customers. Since then, multiple Klue clients have confirmed they were impacted by the breach, including Gong, Jamf, HackerOne, Huntress, Insurity, LastPass, OneTrust, Recorded Future, ReliaQuest, Snyk, Sprout Social, and Tanium. 

At that time, the hacking group Icarus was threatening Klue with the release of the stolen customer data as a means to extort the company. 

As of Thursday morning, when TechCrunch checked, the Icarus site seems to be non-operational, which aligns with what Klue had informed its clients privately. 

While all signs indicate progress towards a resolution, the situation has become more complicated in recent days. Klue has stated that Icarus has warned the firm of a second group of hackers attempting to directly extort its customers. 

This unnamed group published a list of purportedly affected companies on its own website, which TechCrunch has examined, claiming to have acquired Klue’s customer data directly from Icarus. The hackers also asserted that Klue made a payment to an “Icarus operator who is a teenager residing in the UK or nearby countries.” TechCrunch has not independently verified that Klue compensated Icarus, nor could it ascertain the reasons behind the Icarus website being down. A spokesperson for Klue did not reply promptly to a request for a statement. 

According to the hackers, this individual erred, which allowed them to gain access to the server housing the stolen Klue customer data.

“Pay the ransom or we will expose everything if you don’t pay us,” the cybercriminals stated in a message on the site, where they claimed that 195 Klue customers are affected in total. 

In its Thursday briefing to clients, Klue stated: “Icarus informed us that the other group only possesses samples of data for a limited number of customers, not all the information. Icarus has requested us to advise Klue customers against making any payments to this other party.” 

Klue recommended that customers who are in communication with this second group of hackers should request a random sample of data as proof of their claims regarding data possession. 

The company has previously disclosed that the hackers acquired customer data by exploiting a third-party credential from 2022 that was part of a limited pilot program. The hackers then leveraged their access to Klue’s systems to steal authentication keys from customers — known as OAuth tokens — and gain entry to their clouds and databases. Klue has not provided further details about this stolen credential, such as its assigned user or the reasons it remained active for the last four years.

Parker Conrad wants you to think that a significant portion of data analytics should be integrated within human capital management systems — a statement that conveniently positions Rippling, initially an HR software firm, to directly compete with specialized business intelligence solutions.

The proposition is that the contemporary data stack — the array of tools that organizations currently patch together from several vendors — can be unified into a single entity. Transferring data from your various business systems to a warehouse itself constitutes a substantial industry; that’s the function performed by companies like Fivetran and Airbyte. After that, you require a platform to store and query the data, such as Snowflake; then a solution to transform and cleanse it, like dbt Labs; and finally, a visualization layer like Tableau on top.

Conrad’s stance is that Rippling integrates all of these elements into a cohesive system and wraps it in something that others lack: an innate comprehension of your organization, its constantly changing reporting structure, and everything affected when any metric fluctuates. This is the objective of the Rippling Data Cloud, which is set to officially launch Thursday morning.

To illustrate, Conrad shares his screen from his San Francisco office and then provides a glimpse into what Rippling discovered when it activated the product on its own team.

“There were employees mentioning things like, ‘Claude is incredibly helpful for me — it assesses my calendar and my emails and devises a plan for me,’” he notes. “That person was incurring a cost of $30,000 a year for this.”

No one was at fault, he quickly clarifies, but the return on investment simply wasn’t justifiable. It’s the type of insight that most organizations presently lack the means to uncover.

He then shows me a real-time dashboard he constructed by merely asking Rippling AI to evaluate his company’s latest compensation review cycle — distributions of performance ratings, promotion rates by department, salary ratios, all of which can be drilled down to the individual level. He then brings up another dashboard, this one cross-referencing support ticket volume from Salesforce with employee scheduling data — enough to instantly reveal which teams are overwhelmed and which are not. The enrollments team, he points out, is critically understaffed. The travel team has more than double the unresolved tickets compared to the platform team.

However, the example that seems to excite Conrad the most is one related to a concern that many executives currently share: AI token expenditure. He displays a dashboard that merges data from Anthropic’s usage logs, GitHub pull request information, and Rippling’s own performance ratings to scrutinize which engineers are genuinely benefiting from their AI tools and which are wasting money without significant results.

“The top performers are spending the most, which you might expect,” Conrad observes. Yet, the dashboard also highlights engineers with high spending and elevated peer rejection rates on code reviews — these individuals are frequently being asked by their colleagues to redo their work. “If your peers constantly tell you to revisit this, perhaps you’re just producing a lot of subpar work,” he explains.

This analysis has already prompted Rippling to lower spending caps for particular employees. The product can also be set up to notify managers — or automatically revoke access — when an employee exceeds a spending limit.

Regarding the effect on Rippling’s own margins when clients exceed their token limits, Conrad remains vague — “it’s somewhat early,” he states — but dismisses the notion that Rippling is subsidizing customer usage. “We’re not incurring losses,” he asserts, adding that the aim is to maintain it “as affordable as feasible for clients.” The baseline SKU, bundled with Rippling AI, is approximately $20 a month, with usage-based fees applicable for higher users. Currently, about 560 companies utilize it, generating new revenue for the product at around $5 million to $7 million a month.

As for which AI models underpin Rippling’s expanding AI suite, Conrad mentions that the company has a new preferred option currently. “We’ve shifted a substantial amount from Anthropic to OpenAI recently,” he reveals, labeling OpenAI’s 5.5 model as “both superior and more cost-efficient” for Rippling’s objectives. He’s also mindful to state that the balance is constantly evolving, and the company employs different models for varying tasks.

Rippling Data Cloud is the headline launch this week, but it’s not the sole one. Earlier this week, the firm also unveiled Business Banking, which provides a high-yield checking account and same-day payroll processing, a feature Conrad describes as alleviating the mental strain of managing two timelines concurrently. Most payroll systems necessitate processing two to four days in advance; Rippling’s banking offering allows companies to execute payroll on the actual day employees are compensated, with modifications accepted as late as 1 p.m. on payday.

It’s a strategic move into territory dominated by fintechs like Ramp, which recently secured $750 million at a $44 billion valuation — nearly thrice the $16.8 billion valuation assigned to Rippling by its investors last year — and which has been establishing itself as the financial operating system for companies dealing with AI expenses. Conrad welcomes the comparison, noting that Rippling’s banking venture is currently much smaller than Ramp’s but is “growing rapidly and performing exceptionally well,” and that “there are some benefits to centralizing everything.”

Conrad indicates that overall, Rippling is still about two years from becoming cash-flow positive, allocating 45% to 50% of its revenue to research and development compared to the approximately 8% to 9% that public-market HR companies like Paylocity and Paycom allocate. The rationale for building everything in-house is the key point, meaning the reward is a system that can readily address inquiries without the need to extract data from four distinct vendor stacks to do so.

As for an IPO, Conrad is quite clear that he’s not rushing, even though the opportunity is currently favorable. “The public markets have become rather stagnant, favoring slow-growth firms,” he remarks, adding that he’s “not rigid in either direction,” even as it seems quite the contrary. For the time being, he states plainly: “We are not pursuing a public offering. Not even with a ‘wink, wink,’” he emphasizes.

On Thursday, Google unveiled a dedicated mobile application for Google Finance that consolidates users’ watchlists and delivers real-time market information, live financial updates, and Google’s AI-driven “Key Moments” feature, which elucidates stock price movements.

Initially, the app is being released on Android, with Google promising an iOS version will follow in the upcoming months. Additional functionalities, such as the capability to listen to live earnings calls, are also on the horizon.

The introduction of a standalone finance app by Google is likely more focused on positioning itself in the densely populated financial information app market rather than simply providing investors with another avenue to monitor stock prices.

This initiative puts Google in direct rivalry with consumer finance services like Yahoo Finance and trading platforms such as Robinhood.

Additionally, Google announced that its AI-enhanced Google Finance web experience, which was revealed last year, is exiting beta with new offerings.

Google is also implementing portfolios worldwide within the revamped Google Finance web experience, enabling users to monitor their investments through a centralized dashboard that tracks assets and their performance. Existing portfolios in Google Finance will be automatically accessible, and users can establish new portfolios by uploading documents or detailing their investments to the chatbot.

After the portfolios are arranged, users can utilize Google Finance’s AI research feature to inquire, like “which sectors are lacking in my portfolio right now?”

Google has implemented an AI function that enables users to establish tasks using natural-language requests, such as timely updates analyzing market fluctuations, or summaries regarding the performance of their assets. Users can instruct the AI assistant to utilize their watchlist or portfolio for insights tailored to their investments, and once a task is initiated, Google Finance will manage it in the background.

According to Google, these new portfolio and task functionalities are available on the web beginning today, with plans to integrate them into the Google Finance app in the forthcoming months.

The surge in AI has inspired virtually everyone to start their own data center venture. However, establishing a data center is no simple task.

Even after overcoming the challenge of securing GPUs, network switches, and storage solutions, you must also ensure everything is adequately set up, operational, and able to meet the diverse requirements of customers. Preparing a data center to deliver cloud-computing services specifically for AI inference and training can require months of effort. Moreover, the longer it takes to enter the market, the greater the expenses incurred from having those valuable GPUs idle.

Netris, a network automation startup, asserts that it can eliminate this challenge for neoclouds. The firm offers software that operates on network switches and also provides a platform that connects to these switches, enabling neocloud operators to streamline the time it takes to become operational by automating setup, configuration, and management. This platform facilitates network abstraction, allowing for hardware configurations to be adjusted as needed, and it isolates servers and resources at the hardware level, enabling neoclouds to serve multiple clients (multi-tenancy).

If this appears to address a clear issue, you’re absolutely right. Until recently, data centers were predominantly within the realm of large infrastructure entities like Equinix, NTT, Digital Realty, Oracle, Microsoft, AWS, or Google. These companies effectively resolved network setup, configuration, and multi-tenancy for themselves by employing numerous engineers or developing automation in-house. Smaller neocloud enterprises seldom possess such resources.

“As an operator of GPU clusters, modifications must be made to every link daily. Traditional data centers utilized a method known as SDN [software-defined networking] for this, but SDN is inadequate because it is software-based,” stated Netris’ CEO Alex Saroyan to TechCrunch. “For AI, software alone is insufficient due to the high volume of traffic; everything necessitates hardware acceleration. Therefore, you require a solution akin to SDN, yet fully hardware accelerated. This is our expertise, and we have been engaged in this for eight years.”

Saroyan indicated that Netris’ platform is independent of vendors, compatible with networking hardware and standards employed in data centers, accommodating both Nvidia and AMD servers.

The ambitions of the startup have garnered substantial support, including accolades from Nvidia. Two years prior, the prominent chipmaker was so taken by a demonstration of Netris’ technology that it referred the company to several clients. Presently, Netris operates in over 35 GPU clusters globally (approximately a million GPUs), managed by entities such as Lightning AI, Foxconn, Visionbay, Hewlett Packard Enterprise, TensorWave, Telus, and others.

To capitalize on this momentum, Netris has raised $15 million in a Series A funding round, as exclusively reported by TechCrunch.

Importantly, there is no AI involvement in this process. Saroyan explained that the company relies solely on previously developed algorithms for managing and configuring automation and operations.

“Our journey began long before AI. We recognized the challenges early on and started crafting this algorithm promptly. AI isn’t deterministic, right? Sometimes it operates independently. While it excels in creative endeavors, for adjusting thousands of switch configurations, creativity isn’t necessary. What you need is persistence and repeatability.”

a16z partner Guido Appenzeller is now joining the company’s board. Looking ahead, Netris intends to utilize this funding to recruit more engineers and sales personnel, expand support for additional hardware vendors, and enhance its algorithm’s functionality.