Andrew
Verhalen.

Andrew Verhalen does not show up in many founder podcasts. He shows up in the cap tables. The earliest ones, before there is much to talk about, when the deck still has a typo on slide three and the demo only works on Tuesdays.

Today he is a General Partner at Matrix Partners, the firm that has been writing first checks since the seventies and counting Apple, FedEx, Oculus, and Canva among the names it has helped early. Verhalen joined Matrix in 1992. That sentence is easy to skim past, so read it again. Nineteen ninety-two. The internet was a curiosity. Mosaic had not shipped. The S&P 500 sat just below 450. He has been underwriting infrastructure bets longer than most of the founders he funds have been alive.

What does an infrastructure bet look like in his hands? Quiet, mostly. He invests from seed through Series A, around a million-dollar sweet spot, into the unglamorous middle layer of computing: networking gear, semiconductors, the protocols and silicon that make the glamorous stuff possible. The category does not trend. It just compounds.

Before The Term Sheet, A Chip.

Verhalen is an electrical engineer first. Cornell, class of 1978, B.S. in EE. He stayed for an M.Eng., then squeezed an MBA out of Cornell Johnson at the same time. The combination is rare even now. In the early eighties it was practically experimental.

He took it to Intel and became a product manager for the 8086 microprocessor. The 8086 is the chip that anchored the x86 instruction set - the architectural lineage that runs straight through the PC era and into half the data centers still humming today. Verhalen's job was to convince the world to use it. The world obliged.

From Intel he moved to 3Com, where he built a new division focused on network adapters and hubs as a Divisional Vice President. If the 8086 was the brain, 3Com made the nerves. He spent the back half of the eighties watching local area networks turn from a corporate curiosity into the default plumbing of the office.

The Grand Junction Footnote.

In the early 1990s Verhalen helped start Grand Junction Networks, an Ethernet switching company. Switching, at the time, was the right answer to a problem the industry had not fully admitted it had. Cisco bought Grand Junction. The technology went on to anchor a major division at Cisco for years. If you have ever plugged an Ethernet cable into anything, there is a non-trivial chance you have used a descendant of work he helped seed.

The pattern repeated. Unwired Planet, the company that became Openwave, went public in 1999 carrying his board imprint. SiTera, a networking silicon outfit, sold to Vitesse. Alteon WebSystems listed on Nasdaq and was acquired by Nortel for billions. None of these names trend on social. All of them quietly mattered.

A Portfolio Made Of Pipes.

Look at where Verhalen sits today and the through-line holds. He is the lead independent director at Ambarella, the publicly traded computer-vision semiconductor company best known for putting smart chips inside cameras, drones, and the eyes of autonomous machines. Among his more recent infrastructure bets is Sila, the next-generation battery materials company building silicon-dominant anodes - a hardware story, not a software one, with a multi-year horizon and capex to match.

Hardware is back in vogue in 2026 because AI made it impossible to ignore. Verhalen never stopped paying attention. He has been buying picks and shovels through three platform shifts: client-server, mobile, cloud, and now AI infrastructure. The same questions apply each time. Who controls the bottleneck. Who owns the standard. Who is one Moore's-law turn from owning the category.

The Contrarian Posture, Without The Costume.

Matrix describes itself as a metrics-first SaaS investor and a partner for company builders from seed through Series A. Verhalen sits inside that culture as the operator with the longest receipts. He is not a thesis-of-the-month investor. He is the partner you call when the data sheet matters more than the narrative.

There is a kind of contrarianism that is performative and a kind that is structural. His is structural. He has worked inside the parts of the computing stack that get described in white papers, not keynotes. That history means he can underwrite hardware risk, manufacturing risk, standards risk - the categories most generalist VCs outsource to their associates. The reward is dealflow. Founders building in chips, networking, energy storage, sensors, and the awkward middle of deep tech tend to find their way to partners like Verhalen because the partner can actually read the schematic.

What He Funds, In One Line.

Hardware-shaped companies. Network-shaped companies. Anything where the moat is a deeply technical thing that takes years to build and then takes years to displace. Geography is wide - Matrix invests across the U.S., Israel, India, Asia-Pacific, and Latin America. Stage is narrow - seed to Series A, written early enough to actually shape the company. Check size starts around a hundred thousand and runs to ten million, with most checks landing near a million. The thesis, in 1992 and in 2026, is the same. Back the contrarian who has earned the right to be contrarian.

The Cornell Through-Line.

Three Cornell degrees, all stacked between 1974 and 1981. B.S. and M.Eng. in electrical engineering, MBA from Johnson in marketing and operations. The combination is itself a thesis. Engineers who can sell are rare. Engineers who can sell and then learn to allocate capital across three decades are rarer. He has the credential trio that explains the career and never needs to be name-dropped.

Why The Headline Matters.

Andrew Verhalen does not pose for the cover of the magazine. He is one of those partners other partners quietly cite when they need to make a hardware bet defensible to their committee. His name appears on board lists more often than on stages. The companies that win in the infrastructure layer rarely throw launch parties, and when they do, the founders thank the engineers, not the investors.

That suits him. The check writes the same either way.