A 7-person startup just hit $2.8M in ARR — and it's hunting for the next Bill McDermott

The McDermott Signal: Why a Former SAP CEO Is Selling AI to Data Center Operators

Entangl is hunting for a salesperson who can "get a VP of Data Center Operations to reply to a cold email on Tuesday, win over a room of facilities veterans who've never trusted software on Wednesday." The Y Combinator-backed data center AI startup, which lists its team size at 11, posted a job listing for an Account Executive with a title that sounds like a dare: "The Next Bill McDermott." The listing describes someone who "bought a deli at 16 because nobody would give them the job they wanted" and "thinks a discount is the last tool in the box, not the first." The role pays $100,000 to $150,000 in base with 0.1% to 1% equity, per the listing on Standout.

The name isn't accidental. Bill McDermott is one of the most aggressive enterprise sales operators in modern software history. He started at Xerox, became CEO of SAP America in 2002, rose to co-CEO of SAP AG in 2010, and then became the first American CEO of SAP SE in 2014. After leaving SAP, he took over as CEO of ServiceNow and drove the company from roughly $3.5 billion to $10 billion in revenue, according to the Acquired podcast. Entangl is explicitly invoking that playbook. The startup wants to hire someone who operates the way McDermott does, and it's naming the archetype outright.

The hire signals something specific about where infrastructure AI sits right now. Enterprise data center operators, the people who keep power, cooling, and compute running at hyperscale facilities, are not early adopters. They're risk-averse buyers managing uptime requirements measured in nines. Selling them software requires the kind of relationship-driven, multi-quarter enterprise motion that McDermott perfected at SAP and ServiceNow. Entangl's bet is that AI-driven operations tools have crossed the threshold from technical experiment to enterprise-ready, and that the bottleneck is no longer the model — it's the go-to-market.

That bet is backed by serious investors. Entangl is funded by Y Combinator, Tekedia Capital, and others, according to its company description. The company claims AGI Labs already use its platform to keep data centers online. McDermott's involvement, even as a recruiting archetype at this stage, gives Entangl something no technical founder can manufacture: instant credibility in the C-suites of the largest infrastructure operators in the world.

Entangl's Trajectory: $2.8M ARR and $120M Valuation in Under Two Years

Entangl hit $2.8 million in annual recurring revenue in February 2025, barely a year after founding, and carries a $120 million valuation, all with a team that numbered three people at the start of 2024. The company's revenue curve is stark: zero at launch in 2024, $2.8 million by early 2025. That trajectory, documented by GetLatka, puts Entangl in a rare class of infrastructure AI startups scaling faster than most SaaS companies with ten times the headcount.

The numbers tell a specific story about leverage. Three employees generating $2.8M ARR means roughly $933,000 in revenue per person, a figure that rivals or exceeds many mature enterprise software companies. PitchBook lists Entangl at seven total employees as of 2026, meaning the company added roughly four people while presumably growing revenue beyond the earlier mark. The company raised $500K total across three rounds, per CB Insights, with investors including Y Combinator, Transpose Platform, Tekedia Capital, Evolution VC Partners, and Eight Capital Group. Its latest funding round, a Seed VC, closed December 9, 2024.

That ratio (half a million in capital raised against a nine-figure post-money) is the part that should make every ML engineer and data center operator pay attention. Entangl didn't need to burn through a Series B to command that valuation. The company's Y Combinator pedigree gave it distribution and credibility, but the revenue did the heavy lifting.

The talent implication is direct: when a seven-person company generates $2.8M ARR, every hire is a revenue-critical decision. Entangl isn't looking for generalists. It needs engineers who can build ML models that understand physical infrastructure (cooling systems, power distribution, network topology) and operators who can sell that capability to hyperscale data center owners who don't trust black-box automation with billions in assets. That combination of skills is scarce, and Entangl's valuation ramp is proof that the market will pay a premium for it.

What Entangl Actually Builds

Entangl's product rests on a premise most enterprise software founders wouldn't touch: that data center outages are primarily a human coordination problem, not a hardware one. The company's platform uses deterministic AI, not generative models, to crawl across a facility's entire design, flag errors autonomously, and route fixes to the right engineer before anything breaks.

The distinction matters. Generative AI produces plausible output that can be wrong. In a facility running the compute for global derivatives trading (Entangl says over 90% of the world's derivatives rely on the data centers it helps operate), a wrong answer is an outage. The company's system integrates with building monitoring systems and knowledge bases like GitHub and Google Drive, tracking design changes in real time. When it detects a conflict, it pushes a targeted fix to the specific engineer responsible and delivers daily project reviews to keep teams aligned.

The founders came out of aerospace. Shapol M and Antanas Zilinskas led a reusable rocket program and oversaw four launches before starting Entangl in 2024. Their Y Combinator launch post is explicit about the origin story: cross-team design reviews in aerospace grind for weeks, and errors still reach manufacturing. They cite the Challenger O-ring failure as the archetype, a decision made in one siloed team that destroyed the system at low temperature. Entangl's pitch is that machines can crawl across an organization's fragmented knowledge bases and catch those interdependencies before they compound. The company claims the platform saves each engineer roughly two months of review work per year.

The market sizing backs the ambition, even if the numbers vary by source. Roots Analysis puts the global autonomous data center market at USD 48.6 billion in 2026, growing to USD 286.4 billion by 2040. MarketsandMarkets projects the broader AI data center market hitting USD 2,023.52 billion by 2032. The compute buildout itself (Entangl's website says more compute is coming online this decade than in all of history) means the operational complexity is scaling faster than the human workforce managing it.

The product runs on a SaaS subscription model aimed at hyperscalers, AGI labs, and enterprise engineering teams. Beyond outage prevention, the platform simulates work procedures for junior engineers before they touch live infrastructure, and it evaluates the system-wide impact of every change. SOC 2 compliance is built in for enterprise deployment.

Entangl is hiring for roles including a Greg Brockman position and a Jony Ive position, titles that signal the company is recruiting senior operating and design talent, at salaries ranging from $120K to $180K with equity, per its Y Combinator job listings.

Why Infrastructure AI Is the Next Talent War

Entangl is not an outlier. It is the leading edge of a hiring collision that Deloitte quantified in March 2026: data center operators and power companies are now fighting over the same pool of engineers, technicians, computer specialists, and plant operators. Between 2023 and 2025, data center job postings for those core roles surged 64%, while power sector postings rose 20%, both far outpacing the 4% growth in the broader economy. More than one-third of new postings in both sectors targeted the same workers.

The numbers behind that demand are stark. Deloitte estimates US data center power demand will jump from 47 gigawatts in 2025 to more than 176 gigawatts by 2035. Data centers already account for roughly 4.4% of US electricity consumption, a share projected to reach between 6.7% and 12% by 2028. Each new gigawatt requires electrical technicians, power plant operators, mechanical engineers, and the ML systems that keep those facilities from burning money during downtime.

That is where the talent bottleneck gets specific. The roles growing fastest are hybrid positions that did not exist at scale five years ago. Postings for electrical technicians at data centers climbed more than 180% from 2023 to 2025. Nuclear power plant operator postings increased nearly tenfold. Computer specialist roles at utilities grew quickly as grid operators adopted more automation and data tools. These are not pure software jobs. They require people who understand both model training pipelines and the physical systems those models control.

The February 2025 labor data reinforces the split. Demand for AI/ML engineers, data scientists, and cybersecurity specialists rose 18%, with NVIDIA, Microsoft, and Oracle leading recruitment. But the same month saw 172,000 total US job cuts, including 16,084 in tech, as companies like Cisco, Salesforce, and Intel eliminated entry-level positions and redirected budgets toward AI and cloud. Junior coding and data entry roles shrank. Hybrid roles that combine infrastructure operations with AI integration grew.

For engineers, the implication is concrete. Pure ML research roles remain concentrated at a handful of labs and hyperscalers. The faster-growing demand sits at the intersection of machine learning and physical infrastructure, MLOps engineers who can deploy models that manage cooling systems, power distribution, and fault detection in facilities where downtime triggers six- and seven-figure penalties. Deloitte's 2025 AI Infrastructure Survey found that 63% of data center executives cited skilled labor shortage as their top obstacle to securing talent, while power sector leaders ranked workforce competition as their number one challenge.

The salary data reflects the squeeze. Zero G Talent's board shows Databricks listing 61 roles added in the past week alone, with staff research engineer positions paying between $190,000 and $270,000 a year. Anthropic posted 30 roles in the same window, with account executive compensation reaching $450,000 to $550,000. These are not outliers; they are the market signaling where scarce talent commands the highest premium.

Entangl's $120M valuation and Bill McDermott's recruitment make sense inside this pressure. The company is not just selling AI software. It is selling the ability to do more with a workforce that both sectors cannot hire fast enough. The startups that win the next phase of infrastructure AI will be the ones that figure out how to build systems requiring fewer scarce operators, not the ones that need more of them.

Who Else Is Building Autonomous Data Center Workflows

The capital flowing into AI data center infrastructure is staggering. U.S. data center investment alone represents over $1.1 trillion across 604 documented projects totaling 131.7 GW of capacity, according to an analysis by Michael Bommarito. The AI infrastructure segment, companies like Stargate, CoreWeave, Crusoe, and Applied Digital, is growing at 50-100% annually, the fastest of any category. Goldman Sachs Research projects global data center power demand will rise 50% by 2027 and as much as 165% by 2030 compared with 2023 levels.

But the competitive landscape breaks down by layer, and most of the money is going into concrete and GPUs, not into the software that runs the facility itself.

The hyperscalers are building the most capacity, Meta alone leads at 6,054 MW of projects, but they're building it for their own workloads. Google's DeepMind team deployed an AI-driven cooling optimization system that cut its data center cooling energy by 40%, a result published by Richard Evans and Jim Gao. That's the closest direct competitor to Entangl's value proposition, and it's an internal project at a hyperscaler, not a product other data center operators can buy.

CoreWeave, which went public in 2025, is the most visible pure-play AI infrastructure company. It orchestrates GPU workloads at scale and has raised billions in venture and debt funding. But CoreWeave's business is compute provisioning, selling GPU hours, not autonomous facility operations. Crusoe is taking a different angle, using flared natural gas to power modular data centers, solving an energy sourcing problem rather than an operations one.

Then there's ScaleOps, which raised $130 million in Series C funding for autonomous infrastructure resource allocation in Kubernetes environments. It's the closest startup analog to Entangl in terms of using AI to autonomously manage infrastructure, but its scope is container orchestration, not the physical layer of power, cooling, and facility engineering that Entangl targets.

This is the gap McDermott's background matters for. The data center industry has no shortage of companies building GPU clusters or sustainable power solutions. What it lacks is a company selling AI-driven operations as an enterprise product to the operators who already run the world's data centers, the mechanical engineers, facility managers, and site reliability teams who make decisions about cooling set points, maintenance schedules, and failure response every shift.

The colocation and hyperscale operators spending $200 billion-plus annually on expansion need a software layer that reduces the operational complexity of those facilities. Entangl is pitching itself as that layer, and McDermott's SAP-era playbook, selling mission-critical software to conservative enterprise buyers through relationships, references, and risk-reduction framing, is the go-to-market motion that matches the buyer.

Nobody else is combining AI-native issue detection and auto-resolution with an enterprise sales engine built for data center operators. That's the opening.

What McDermott's Playbook Means for the Next Wave of AI Startups

McDermott's move to Entangl is not an isolated case. It is a pattern, and the pattern tells you where AI is actually going.

In November 2025, Intuit appointed both McDermott (still ServiceNow's chairman and CEO) and Nasdaq CEO Adena Friedman to its board, effective August 2026. The press release called McDermott an "enterprise AI leader" and cited his "deep expertise in AI-powered transformation and scaling platform businesses." Sasan Goodarzi, Intuit's CEO, said McDermott's experience driving "complex, global enterprise sales" would be "invaluable" as Intuit pursues larger business customers. A financial technology company with roughly 100 million users is bringing in a former SAP chief specifically to sell AI to bigger, harder, more regulated buyers.

McDermott himself has been blunt about what that requires. Speaking at Sierra Ventures' CXO Summit in November 2025, he said executives are "done with pilots that never scale" and want "fewer platforms, deeper partners, and tangible ROI." His advice to startups: "Pilots are not a business model. Proof is." He described how ServiceNow deployed a live system for the NHL in two weeks when the league hesitated to start another proof of concept. The point was not the technology. The point was the speed from demo to production.

That philosophy, integrate first, prove value in weeks, skip the pilot purgatory, is exactly what McDermott is now selling at Entangl to data center operators who cannot afford downtime. And it is the same playbook other former enterprise CEOs are running at AI startups. Vishal Sikka, former CEO of Infosys, launched Hang Ten Systems, an enterprise AI startup that raised $32 million in seed funding led by Mayfield, with Aramco Ventures participating. The pitch: help large organizations continuously build, modify, and operate enterprise software using AI. Former big tech leaders are starting or joining AI ventures at a clip, and they are not building research projects. They are building revenue machines.

The signal here is structural, not anecdotal. AI is graduating from the lab phase. The startups that will win the next round of infrastructure deals, in data centers, energy, defense, logistics, are the ones that can close a Fortune 500 contract, not just publish a benchmark. That requires a different kind of founder and a different kind of hire. It requires people who know how to navigate procurement at a hyperscaler, how to structure a deal with a utility, how to sell to a CFO who has seen ten AI demos and signed zero contracts.

For engineers and operators reading this, the implication is concrete. If you have ML skills and you understand physical systems (power, cooling, compute, uptime), you are now competing for roles against people who have run billion-dollar enterprise sales organizations. The job postings reflect it. That same hiring board shows Databricks adding 61 roles in the past week alone, including staff research engineers in data agents and director-level enterprise positions. Anthropic added 30 roles in the same window, with applied AI architects and account executives for global system integrators. These are not research lab hires. These are go-to-market and deployment roles at scale.

McDermott's career arc, from SAP to ServiceNow to Entangl to the Intuit board, maps the maturation of enterprise AI itself. Each move is further from pure R&D and closer to revenue, operations, and regulated infrastructure. The next wave of AI startups will be built by people who have already sold through a procurement cycle, not just trained a model. If you want to work on the infrastructure side of AI, learn to speak both languages: the model and the contract.

Working in AI? Zero G Talent tracks the openings: browse AI jobs, openings at Databricks and Anthropic, and the people building the field.