Hypersonix Inc.

Somewhere in a grocery chain's pricing department, a spreadsheet is open. A category manager has been staring at it for two hours, cross-referencing competitor prices, last week's sell-through data, and a promo calendar that was already outdated on the day it was printed. She has 40,000 SKUs to reprice before the weekend.

At the same moment, inside Hypersonix's AI platform, an autonomous agent has already repriced those SKUs - 42,000 of them, to be exact - factoring in real-time competitor moves, demand signals, margin thresholds, and promotional overlap. That process took 11 minutes. The category manager is still on row 47.

This is what Hypersonix does every day, at scale, for more than 100 retail and eCommerce brands across 12 industry verticals. The company's flagship product, ProfitGPT, makes over 100,000 autonomous pricing decisions daily across 5 million+ SKUs. Its customers include names like LVMH, Woodman's Market, WildBerries, and Rogers Sporting Goods. Gartner calls them a "Cool Vendor." Their customers call them something more practical: essential.

With around 86 people in San Francisco and an estimated $27.6M in annual revenue, Hypersonix isn't the biggest company in enterprise software. But in the narrow, ruthlessly competitive corridor of retail AI, they are building something that looks increasingly difficult to displace.

Retail margins are not generous. The average grocery margin hovers around 2-3%. eCommerce, for all its hype, routinely burns through investor capital trying to compete on price while keeping inventory costs from turning a warehouse into a graveyard. The difference between a profitable quarter and a write-off is often one decision made two weeks too late.

The problem isn't that retail teams are bad at their jobs. It's that the problem is too big for a team. A mid-size eCommerce retailer might carry 100,000 SKUs. Each one has a price, a competitor price (usually several), a demand curve, a margin structure, a promotional history, and an inventory position. The interactions between these variables change constantly. Pricing something correctly on Monday can mean something entirely different by Thursday if a competitor runs a flash sale.

Before Hypersonix, this problem was managed through a mix of manual analysis, periodic pricing reviews, and whatever intuition a seasoned merchandiser had developed over 20 years. It worked, more or less. Until it didn't - until competition got faster, supply chains got messier, tariff policy became a weekly surprise, and the window for a good pricing decision shrank from weeks to hours.

Prem Kiran, Rama Rao, and Gina Becchetti founded Hypersonix in 2017 with a specific belief: that the mechanics of retail decision-making - pricing, promotions, inventory, competitive positioning - were not fundamentally human problems. They were data problems. And data problems, at sufficient scale and with the right architecture, were AI problems.

The bet was that merchants would eventually pay for a system that didn't just surface insights but acted on them. Not a dashboard. Not a report. An agent. Something that could look at 5 million SKUs, understand the margin implications of a competitor's overnight price drop, and recommend (or execute) a repricing strategy before the first analyst arrived at their desk.

That was 2017. "Agentic AI" wasn't a buzzword yet. The founders were early on the thesis and had to build much of the infrastructure from scratch - training models on retail-specific data, developing computer vision for product matching across competitor catalogs, and building the kind of real-time data pipelines that could make sub-hour decisions meaningful.

By 2021, that bet had attracted $35M in Series B funding led by B Capital Group, with Intel Capital, Firebolt Ventures, and investor Gokul Rajaram joining the round. The valuation hit $200M. Total funding: $50.1M.

The product is called ProfitGPT, and the name is deliberately unambiguous. It's not a "decision support tool." It's not a "business intelligence dashboard." It's a generative AI platform built to maximize retail profit - and it does so through six interconnected AI modules that can be deployed individually or as a unified system.

Jarvix deserves its own paragraph. It's the part of the platform that retail teams actually interact with day-to-day - a conversational AI that gives non-technical users instant access to pricing, inventory, and competitive intelligence without requiring them to know SQL or navigate a dozen dashboards. Think of it as giving every category manager their own data analyst who never sleeps and has already read every competitor's website today.

In April 2025, Hypersonix added Agentic Tariff Workflows - a direct response to the tariff volatility that has been reshaping global retail supply chains. The feature gives retailers automated visibility into tariff impacts at the SKU level, with real-time competitor pricing alerts and automated repricing rules that fire when trade costs shift overnight.

A specialty home improvement retailer saw a 16.52% profit lift. A books, gifts, and baskets eCommerce retailer achieved a 14.4% profit lift and a projected $1.4-$2.3M annual profit impact. One retailer cut stockout rates, another halved the hours their pricing team spent on analysis that was already outdated by the time it was finished.

There's a certain credibility that comes from having both a fast food chain and a luxury fashion conglomerate in your customer list. KFC cares about food costs and throughput. LVMH cares about brand integrity and premium positioning. The fact that Hypersonix serves both says something about the platform's configurability - and about how universal the pricing problem actually is.

The platform covers 12 industry verticals: grocery, apparel, beauty, furniture, electronics, sporting goods, industrial supplies, optical, health and wellness, food and beverage, automotive, and manufacturing. The common thread is not the category; it's the problem. Every one of these sectors has too many products, too many competitors, and too few analyst hours to price correctly without help.

Intel Capital's participation in the Series B is worth noting. Intel doesn't invest in every enterprise SaaS company with a good pitch deck. Their presence signals something specific: confidence in the underlying computational architecture. Hypersonix's ML stack - Python, PySpark, Databricks, Delta Lake, Snowflake - is built for the kind of throughput that makes 100,000 daily decisions feel lightweight.

Hypersonix's stated mission is to give every retailer the autonomous AI intelligence needed to maximize profitability. The word "autonomous" is doing a lot of work in that sentence, and intentionally so. The company is not building a better analytics dashboard. They're building a system that can act - that can decide and adjust and reprice and reorder without waiting for a human to review a report.

In April 2025, when tariff volatility started hitting retail supply chains in unpredictable ways, Hypersonix didn't wait for customers to ask for a solution. They shipped Agentic Tariff Workflows within weeks - a feature that automatically monitors tariff impacts at the SKU level and fires repricing rules when trade costs shift. That's what it looks like when a company actually believes its own thesis about autonomous AI.

The broader bet is on a retail industry that continues to compress margins, complicate supply chains, and demand faster decisions than human teams can make. That trend line has been moving in one direction for 20 years. Hypersonix is building in the direction of where retail is going, not where it has been.

Back to that grocery chain's pricing department. The category manager is on row 47. But now there's a Hypersonix terminal on her second monitor. She asks Jarvix - in plain English, like a text message - "Which of our dairy SKUs are we losing on compared to Competitor X this week?" The answer comes back in seconds: 23 SKUs, specific price gaps, recommended adjustments, and a projected margin impact if she acts now versus Thursday.

She doesn't reprice 42,000 SKUs by hand anymore. ProfitGPT handles that. What she does now is make judgment calls on edge cases, approve strategy shifts, and spend her expertise on the decisions that genuinely require human context. The rest is autonomous.

That's the practical version of Hypersonix's mission statement. Not a technology pitch - a different kind of working day for the people who keep retail margins from evaporating. The spreadsheet is still open. The decisions are just better now, and they happen faster than any analyst could manage alone.

Hypersonix didn't change what retail does. They changed how fast it can think.