Via Transportation (VIA): Everyone Argued About the Wrong Question

Disclosure: I sold my position and no longer own shares. This is my personal research, not investment advice.

In December 2025, Bleecker Street Research published a short report on Via Transportation called "Road to Nowhere." The argument: Via is a labor-intensive transit contractor masquerading as a software platform. The stock fell from a $46 IPO price to under $16.

The bulls spent the next six months arguing Via really is a software company. The bears spent it arguing it isn't. I think both camps were asking the wrong question. Via isn't a failed software company. It's a successful services company that nobody had a clean comp for — and that's where the mispricing lived.

The bears had the facts right and the conclusion wrong

Almost everything in the short report checks out against the 10-K. 85% of cost of revenue is tech-enabled services: drivers, fleet, support staff. R&D spending was essentially flat for three straight years — $96M, $89M, $92M — while revenue grew 75%. Organic customer growth was 9%; the headline 23% came mostly from an acquisition. Auto liability insurance sits in G&A rather than COGS, which flatters gross margin by a point or so.

None of that makes it a bad business. It makes it a different business than the one being priced.

What the short report underweighted: adjusted EBITDA margin went -37% in 2023, to -16% in 2024, to -8% in 2025. That is a services business with real operating leverage, and leverage is the thing services contractors almost never have.

The actual mechanism

Gross margin was never going to be the story. It sits at 40-41% and it will stay there, because the cost of revenue is people and vehicles.

The story is operating expense. Quarterly revenue went from $53M in Q1 2023 to $136M in Q2 2026 — up 156%. Adjusted operating expenses over the same period went from $47M to $60M, up 28%. Services-shaped gross margin, software-shaped cost structure. That combination is rare and it's what the market wasn't pricing.

Why it was cheap

The comp set is the whole thesis. Not Trade Desk or Tyler Technologies at 9-10x sales. The right comparison is government services: ICF at roughly 0.7x revenue, Maximus near 1.0x, Booz Allen near 1.5x. Those businesses grow 5-10% a year and run 9-11% operating margins. Via was growing 27% — four to five times the comp set — and in May it traded at 1.7x forward sales. Below the comp range, as the fastest grower in the category. I bought at $17.78.

What happened

Q1 and Q2 both beat. Q2 revenue came in at $135.7M against a $132.5-134M guide. Adjusted gross margin 41%, adjusted EBITDA margin -2.5% — best on record — adjusted net loss of $838K, essentially breakeven. Full-year guidance raised twice. Management committed to positive adjusted EBITDA in Q4 2026.

The thesis worked. Faster than I expected.

Why that's the problem

At ~$31, enterprise value is roughly $2.2B against about $551M of 2026 revenue. That's 4x sales, against a comp set topping out at 1.5x.

I ran it backwards: to justify the current price, Via needs roughly 2.6x the 2032 free cash flow my model produces — an 18% FCF margin on a business whose Q2 FCF margin was negative 10%. I stress-tested every input. Lower discount rate, extreme cost discipline, gross margin expanding to 45%, stock comp cut in half. No single change closes the gap. You need all of them at once.

What could make me wrong

• Multiple expansion is exactly the variable I modeled worst. In May I projected a re-rate to 2.5x over 12-18 months. It hit 3.6x in three and a half. If the market pays software multiples for this indefinitely, selling was early.
• The pipeline is real: over $700M in Growth Annual Contract Value, doubled year-over-year two quarters running. $1M+ customers grew 36% to 114. Revenue per customer hit a record $641K.
• Q4 2026 is a genuine step function — the first positive adjusted EBITDA quarter. If it lands clean, the multiple may be defensible on actuals rather than promises.
• International outside Germany grew 41% in the first half. Germany is an isolated drag, not a global one.

What I'm less comfortable with

• Free cash flow went the wrong way. H1 2025 was -$25.0M; H1 2026 was -$36.5M, while adjusted EBITDA improved by $8M. Receivables grew from $82M to $105M and collection time stretched from about 63 to 70 days.
• Stock comp is 11.8% of revenue, up from 4.4% a year ago. "Positive adjusted EBITDA" is not GAAP profitability and won't be for years.
• Growth deceleration is a straight line: 42% at peak, 27% now, roughly 1.5 points per quarter for six quarters.
• Federal funding partially broke. The Senate voted 90-6 in August to extend surface transportation authorities through December 11, but deliberately excluded the IIJA's advance appropriations — $36.8B per year, which lapsed October 1. Via is 90%+ government revenue.
• The offsetting piece: the House's replacement bill authorizes roughly $580B with 90% delivered through formula funding, a significant shift away from IIJA's reliance on competitive grants. Formula programs are Via's revenue base. The House bill hasn't reached the floor and the Senate hasn't started.

Bottom line

I was right about the business and the market agreed with me faster than I expected. That's a good outcome and also the reason the position stopped making sense.

The thing I'd want someone to take from this: a closed mispricing is an exit signal even when the company is still executing beautifully. Those are separate facts. Via got better through my holding period — and that's precisely when it became a worse bet, because the price moved more than the business did.

Over the whole period my revenue estimate moved 1.6%. The multiple moved 111%. Essentially all of the return came from the market changing its mind about what to pay. Forecasting revenue and margin is the tractable part. The multiple carries most of the error, and it deserves an explicit range rather than a point estimate.