A wordmark four letters long, sitting on a billion data points. PDF is not the file format - it is the company quietly grading every chip the modern world runs on.
Since 1991, PDF Solutions has turned the firehose of semiconductor manufacturing and test data into something a yield engineer can actually use: a decision.
Achip fab never sleeps, and neither does the data it produces. Each wafer that passes through hundreds of process steps leaves a confession behind: voltages, defect maps, test results, tool readings. Multiply that by thousands of wafers, dozens of tools and a supply chain that spans foundry, assembly and test houses on three continents, and you get a problem no spreadsheet was built for.
That 3 a.m. twitch - a yield curve dipping half a percent - is worth millions if you catch it tonight and a recall if you catch it next quarter. PDF Solutions exists for that exact moment. Its software watches the twitch, traces it back through the silos, and tells an engineer in Santa Clara, Hsinchu or Kariya which tool, which step, which die.
It is unglamorous work. There is no consumer app, no viral logo. And yet TSMC, Intel, Qualcomm, Analog Devices and DENSO all feed their most sensitive manufacturing data into the same platform - because the data layer is neutral, and the math does not take sides.
One platform, three audiences, a single promise: connect every wafer, die and test result, then make it analyzable in real time.
Data acquisition, normalization, semantic modeling, big-data cloud management, AI/ML and visualization - with tailored editions for IDMs, fabless designers and OSATs.
A ModelOps layer that pairs Exensio analytics with Intel's Tiber AI Studio to build, deploy and run machine-learning models directly on manufacturing operations.
Equipment-connectivity and factory-automation software (SECS/GEM, Industry 4.0) that wires fab tools into the data exchange network.
On-chip test structures plus e-beam inspection so defects in advanced process nodes effectively confess themselves early.
Test chips and electrical characterization that diagnose process variation and accelerate the yield ramp on new nodes.
Secure remote-access infrastructure connecting customers, fabs and partners across the global data network.
130+ semiconductor companies and 500+ clients across 36 countries - foundries, integrated device makers, assembly-and-test houses and fabless designers. The users are the engineers who own yield, test, quality and reliability.
Approximate annual revenue. 2025 up ~22% year over year. Market cap reached roughly $2.3B by December 2025. Source: company filings & market data.
John Kibarian and Kimon Michaels found PDF Solutions in San Jose to bridge aggressive IC designs and stubborn submicron process variability.
Co-founder John Kibarian, a Carnegie Mellon PhD, steps into the chief executive role he still holds.
The company goes public in July under the ticker PDFS.
Si Automation (2006), StreamMosaic (2019), Cimetrix (2020) and Lantern Machinery Analytics (2023) extend connectivity and analytics.
Exensio Studio AI launches with Intel; a landmark IDM contract is signed; DENSO goes on the record about Exensio's impact.
A Carnegie Mellon electrical & computer engineering PhD who developed yield-diagnosis algorithms at the SEMATECH Center for Rapid Yield Learning before co-founding the company. President since day one, CEO since 2000.
Co-founder alongside Kibarian in 1991, steering the products and solutions side of a business that has kept its founding mission intact for over thirty years.
"Software, hardware, IP and services to tackle the hardest development and manufacturing challenges in semiconductors."
Licensing partnership bringing Intel's Tiber AI Studio into Exensio Studio AI.
Integration of DFT diagnosis and layout context with Exensio's manufacturing and test datasets.
Collaboration on AI-driven, agent-assisted fab decision-making.
Interviews, product context and the official channels.