SDVerse’s ‘AI Sourcing Engine’ Confirmed at CES 2026

SDVerse argues that the real bottleneck of the software era is discovery and sourcing. If this process is slow, SDV inevitably slows down as well - and SDVerse says it can compress that timeline from months to minutes. At CES 2026, that promise moved closer to reality through an AI-powered sourcing engine and a private architecture designed to address the cloud concerns that OEMs continue to hesitate over.

By | Sang Min Han _han@autoelectronics.co.kr
한글로보기






CES 2026 was a place where robotics poured out onto the show floor, and AI seemed to cover the surface of every industry. Automotive was no exception - each year brings new buzzwords: SDV, SD, AI-defined vehicle. The direction of technology stays similar, but the terminology keeps changing. What matters more now is whether those words are actually changing the field.
That was exactly why I sat down again - inside SDVerse’s private room at CES - with Prashant Gulati, CEO of SDVerse, and Jeff Walker, CCO, who is leading much of the platform’s expansion in practice. I wanted to confirm whether SDV is being consumed as just another CES trend, or whether the industry is truly moving into a software-first era - and, in the middle of that shift, why SDVerse becomes necessary, and how the platform is evolving into a “real tool” through AI.
The conclusion was clear. What SDVerse emphasized wasn’t “what AI can do,” but something far more structural: AI is compressing the time spent on discovery and sourcing - and that change aligns precisely with what OEMs actually want on the ground.
“Our goal is to take what used to take months and make it possible in minutes. We talk about ‘Months to Minutes,’” Gulati said, summarizing SDVerse in one line.



Whatever You Call It, Cars Become Software

One of the most striking moments in the meeting was how quickly the conversation ended the buzzword debate. Whether “SDV” is the right term, or whether “SD” is the new trend - SDVerse considers that discussion increasingly irrelevant. Gulati redirected the point back to one thing: the growth of code.
“Call it whatever you want. In the end, cars become software.”
The numbers he referenced were symbolic. A single vehicle today may already carry roughly 250 million lines of software, and projections suggest it could climb toward 800 million lines in the not-so-distant future. The exact number matters less than the direction: software is increasingly defining the vehicle’s functions, interface, and the way customers use it.
And this is where SDVerse targets a different problem. It is not trying to “build features.” It is addressing the question the industry must solve before features even begin.
“The real issue is where we find the software, how we find it, and how fast we can find it.”
That line captures SDVerse’s focus with precision.



What SDVerse Is Changing Isn’t Development
- It’s the Structure of Discovery

Software is exploding as SDV accelerates, but the industry’s sourcing behavior is still shaped by hardware-era inertia. In practice, many OEMs and Tier suppliers still discover software in familiar ways: they search Google, attend conferences, call existing vendors - and the process stretches into months.
Once RFP and RFI cycles begin, people and schedules get locked in. Requirements documents go out, answers come back, meetings get scheduled, clarifications bounce back and forth. Travel and coordination costs pile up, and discovery itself becomes the bottleneck.
Gulati described the reason for this slowness in extremely practical terms.
“Because it becomes face-to-face meetings, travel happens, scheduling becomes necessary, and just getting through the questions in documents takes time.”
SDVerse keeps returning to one point: if sourcing is slow, SDV becomes slow.
So instead of trying to “speed up development,” SDVerse attempts to change what happens before development - by rebuilding discovery as a structure.
The idea is simple but aggressive: find what already exists first, reduce unnecessary internal development, and improve capital efficiency. The moment companies attempt to build everything in-house, time, talent, and cost start collapsing together.



Once AI Entered, It Became More Than “Search”
- It Became a Sourcing Engine

The biggest update in the CES meeting was AI. But the real point wasn’t that SDVerse “uses AI.” The point was that the tool itself shifted - from keyword search to a requirements-based sourcing engine.
Traditional search is simple: type in “parking assist” or “cybersecurity,” then browse a list. What SDVerse demonstrated at CES was different. You can input the exact sentences engineers use in reality - what procurement teams write directly into documents - and then receive results.
“We need an AUTOSAR-compatible ADAS stack validated for production.”
“We are looking for a VCU that meets ISO 21434 requirements.”
“We need functional safety consulting compliant with ISO 26262.”
Results appear - but what matters is that they don’t stop at “listing options.” The system explains why something matched, and also why the score was reduced. A solution might match highly, but get penalized because it isn’t truly production-grade - only a pilot.
In other words, AI isn’t acting like a faster search bar. It becomes part of sourcing judgment.
“Before, it was just keyword search. Now it’s shifting to contextual search, like ChatGPT understands meaning,” Walker said.
Another major shift is the handling of the requirements document itself. Upload an RFP, and the system parses it, identifying suppliers that meet the conditions. The initial discovery phase - often two to three months - compresses down to minutes.
This is where “Months to Minutes” becomes more than a slogan. It becomes functionality.



Why OEMs Hesitate on the Cloud: Because Search Is Strategy

As AI sourcing becomes faster, OEM sensitivity increases around one point: the cloud. OEM hesitation isn’t driven by vague fear like “the cloud is dangerous.” The reason is far more concrete.
For an OEM, search history and requirements documents are strategy. What they’re searching for, what specifications they demand, and where they see risk or weakness - these reveal their development direction and business decisions.
Gulati framed it directly.
“Large members don’t want this search to go to the cloud and let others see what we’re looking for - because that’s strategy.”
That is why SDVerse emphasized its private architecture - an approach where AI can be referenced, but OEM data does not leak outward or become shared. In practice, OEMs don’t want “AI” as the goal. They want a structure where AI can exist without compromising security and control.



OEMs Are Moving Through a Learning Journey: “Build or Buy”

SDVerse also spoke about ongoing OEM discussions during CES. The interesting part is that the questions OEMs ask are fundamentally the same.
“Can we do this in-house? Do we need an external platform? Can we actually get our engineers to use it? It’s a learning journey.”
At first, internal development looks attractive. But as projects grow and schedules slip, everything returns to sourcing efficiency. SDVerse is trying to prove value at exactly that moment - especially in a way engineers can accept. That’s why AI-based discovery becomes so important: it doesn’t convince procurement alone - it makes engineers feel, “This is usable.”



SDVerse executives meet again at CES 2026 - (from right) Jeff Walker, Prashant Gulati, and Brian Carlson of ClearCatalyst Consulting. 



A Opportunity for Small Companies, Scale for Big Companies

SDVerse becomes especially interesting because of a scale paradox. Large OEMs no longer want only large suppliers. They increasingly need specialized solutions from small companies, the speed of startups, and niche capabilities. The problem has always been connection.
In the traditional structure, it can take years for a small company to even get in front of an OEM. A marketplace compresses that timeline. Small companies can surface through structured discovery - not through sales networks. OEMs can compare verified attributes and still find suppliers beyond their existing vendor lists.
Gulati described the mechanism clearly.
“For example, GM doesn’t only want to work with big companies like Bosch or Forvia. They also want to work with smaller companies like PopcornSAR or Methodica. We’re trying to make that matching possible.”
That line sits close to the essence of SDVerse’s ecosystem: it is not just “more participants.” It is a different way of connecting them.



The Next Step Isn’t “Standards” - It’s Commercialization

SDVerse does not position itself as a standards organization. While AUTOSAR, COVESA, SOAFEE, and others focus on standards and interoperability, SDVerse focuses on bridging the outcome into commercial sourcing and execution. Whether the software is open source or proprietary, whether it follows one standard or another - SDVerse’s stance is that the market chooses.
“We’re not an organization that defines standards. We focus on commercialization. Open source, proprietary software, AUTOSAR - anything can come in, and the market decides,” Walker said, stating SDVerse’s position plainly.
The more technology expands, the more urgently the industry needs “something usable.” SDVerse is moving the platform toward that reality - and AI acts as an accelerator.

CES is always a place where the future is spoken about. But what SDVerse presented felt less like “future talk,” and more like an answer to what the industry needs right now: speed in sourcing.
The moment SDV moves beyond buzzwords and into execution begins where tools like this actually attach to the field.

AEM(오토모티브일렉트로닉스매거진)

---