The $3.8 Billion Evaporation
On October 19, 2023, Dan Lewis sent a memo to approximately 500 remaining employees — down from a peak of 1,500 — informing them that "Today is your last day at the company." Convoy, the Seattle-based digital freight brokerage that had raised over $1 billion from investors including
Jeff Bezos,
Bill Gates, Reid Hoffman,
Marc Benioff, CapitalG, Generation Investment Management, Baillie Gifford, and T. Rowe Price, that had been valued at $3.8 billion just eighteen months earlier, that had been named to the CNBC Disruptor 50 list five separate times, was shutting down. Not pivoting. Not acqui-hiring. Not restructuring through bankruptcy with a plan to emerge leaner. Shutting down. The company had spent four months exhausting every strategic option — the most logical acquirer, Lewis wrote, had itself been crushed by the same forces destroying Convoy — and was weeks from running out of cash entirely.
The number that matters here is not the $3.8 billion peak valuation, though it is staggering in retrospect. It is not even the $1 billion-plus in total capital raised, though that figure represents one of the more spectacular capital destructions in the 2020s venture landscape. The number that matters is the spread between the cost of a truckload in early 2022 and the cost of that same truckload in late 2023 — a freight rate collapse so severe that Lewis would describe it as an "unprecedented freight market collapse" coinciding with "dramatic monetary tightening." This was the vise. A business model designed to thrive on transaction volume in a liquid, digitized freight marketplace was instead squeezed between plummeting rates and vanishing capital, and neither jaw relented.
What makes the Convoy story worth anatomizing in detail is not that it failed — startups fail constantly, and freight brokerages fail with particular regularity during cyclical downturns. What makes it consequential is the specificity of its ambition, the pedigree of its backers, the quality of the thesis, and the instructive brutality of a market cycle that did not care about any of those things. Convoy was not a half-baked idea that attracted dumb money. It was a carefully considered attempt, led by experienced operators from Amazon and backed by some of the most sophisticated capital allocators on the planet, to apply the Uber marketplace playbook to a $800 billion U.S. trucking industry that moved freight the same way it had for decades — through phone calls, faxes, and relationships brokered by middlemen. The thesis was sound. The execution was real. The technology worked. And it died anyway.
By the Numbers
Convoy at the Peak
$3.8BPeak valuation (April 2022)
$1B+Total capital raised
1,500Peak headcount
~500Employees at shutdown (October 2023)
5xCNBC Disruptor 50 appearances
$260MSeries E raise (2022)
~$800BU.S. trucking market (annual)
~35%Truck miles run empty (industry average)
The Amazon Diaspora and the Supply Chain Epiphany
Dan Lewis is the kind of founder Silicon Valley's pattern-matching machinery is built to fund. Before starting Convoy, he led new shopping experiences at Amazon — a company whose logistics obsession had, by the mid-2010s, become the defining competitive weapon in American retail. Before Amazon, he worked at Microsoft. He was not a trucking guy. He was a technology operator who had spent years watching, from inside the world's most sophisticated supply chain organization, how the unglamorous movement of physical goods was becoming the primary axis of consumer decision-making. "It wasn't the location of the store, the ambiance, the quality of the sales staff," Lewis would later tell CNBC. "It was the fact that people could get things fast."
His co-founder, Grant Goodale, shared the Amazon pedigree. Together they represented a particular archetype of the 2015-era startup founder: deeply technical, operationally credentialed, attacking a massive legacy industry from the vantage point of having seen what a well-functioning technology platform could do to supply chains at scale. They founded Convoy in 2015 in Seattle — not San Francisco, which was already becoming the center of gravity for logistics startups like Flexport, but Seattle, where Amazon's cultural DNA was in the water supply.
The founding insight was deceptively simple, and Lewis articulated it with the clarity of someone who had spent years watching brokers operate from the Amazon side of the transaction. The U.S. trucking industry is staggeringly fragmented. Approximately 80% of freight transportation dollars flow to trucks — more than air, ocean, rail, and pipeline combined. But the industry is dominated by small carriers: owner-operators running one or two trucks, small fleets of five or ten, family businesses spanning a few dozen rigs. These carriers find loads through freight brokers, intermediaries who match available trucks with shippers who need goods moved. The brokerage process was, in 2015, still overwhelmingly analog. Phone calls. Faxes. Relationships built over decades between individual brokers and individual carriers. A broker working a particular set of loads wouldn't know about all available trucks, and a carrier working with a handful of brokers wouldn't see all available freight. The result was endemic inefficiency: roughly one-third of the time, trucks ran empty. Not partially loaded. Empty. Burning diesel, wearing rubber, occupying highway lanes, generating carbon — and earning nothing.
About one-third of the time, trucks run empty. It's kind of a tragedy of the commons, where you're trying to get the right appointment for pickup, trying to get the right point for drop off, but you might not get the right appointments.
The waste was not just financial — though Convoy would later estimate that empty miles cost the industry billions annually. It was structural. In a fragmented market mediated by human brokers, information asymmetry was the default state. No single broker could see the entire supply-demand picture, and no carrier could access all available loads. The market-clearing mechanism was a phone tree. Lewis and Goodale saw a platform opportunity: build the digital layer that connects every shipper with every carrier, use data and algorithms to optimize matching, reduce empty miles, and capture a slice of the spread that had historically flowed to brokers whose primary asset was a Rolodex and a phone.
This was, in essence, the Uber pitch applied to freight. And in 2015, that was the most fundable sentence in the English language.
The Architecture of the Digital Freight Marketplace
What Convoy actually built — the product, the technology, the operational system — deserves more precise examination than the shorthand "Uber for trucking" allows. The analogy was always imperfect, and the ways in which it was imperfect illuminate why the business was both more promising and more fragile than its elevator pitch suggested.
In the ridesharing model, Uber connects a relatively homogeneous supply (drivers with cars) with a relatively homogeneous demand (passengers who need rides). The matching problem is complex at scale but simple in structure: find the nearest available driver for this passenger. In freight, the matching problem is multidimensional. A load has a pickup location and a delivery location, but also a pickup time window and a delivery time window, cargo type, weight, equipment requirements (flatbed? refrigerated? dry van?), and hazmat classification. A carrier has a current location but also a preferred lane, a home base to return to, maintenance schedules, hours-of-service regulations, and a set of equipment capabilities. Matching a load to a carrier requires solving, at minimum, a constrained optimization problem across geography, time, equipment, and economics — and doing it thousands of times per day across a network that spans the continental United States.
Convoy's technology stack was built to solve this problem at scale. The company developed a mobile app for carriers — "tens of thousands of small trucking companies and owner-operators," as Lewis described it — that allowed them to see available loads, accept bookings, and manage their schedules digitally. On the shipper side, Convoy offered a platform for tendering freight, tracking shipments, and accessing capacity on demand. The matching algorithm sat in between, attempting to find the optimal carrier for each load based on proximity, lane preference, price, and historical performance.
The company also expanded beyond pure spot-market matching. Convoy Go, a trailer-pool product, allowed shippers to preload trailers before a driver arrived, decoupling the driver's schedule from the shipper's loading dock operations. This was operationally significant: in traditional trucking, a driver might wait two to four hours at a warehouse for a load to be ready — time known as "detention" that is enormously costly in a business where drivers are paid by the mile, not the hour. Convoy Go attacked detention directly. The company also moved into contract freight, offering carriers "dedicated work on consistent, dedicated lanes" through the app — a shift from pure spot-market volatility toward the more stable, if less lucrative, contract business.
By 2021, when supply chain chaos dominated headlines and demand for trucking capacity reached historic levels, Convoy had assembled what looked like the critical elements of a genuine platform business: a two-sided network of shippers and carriers, proprietary technology for matching and pricing, data exhaust from every transaction that could improve the algorithm over time, and product extensions that increased engagement on both sides. Lewis could truthfully claim to have built "the first company to create a completely digital trucking network."
The Investor's Dream and the Operator's Nightmare
The capital that flowed into Convoy tells its own story — one that rhymes uncomfortably with the broader venture dynamics of the 2019–2022 era.
Convoy's capital trajectory from seed to shutdown
2015Founded by Dan Lewis and Grant Goodale. Early seed funding from Allen & Co. and angel investors including Jeff Bezos, Reid Hoffman, and Marc Benioff. Valued at less than $60 million.
2017–2019Successive venture rounds from CapitalG (Alphabet's growth fund), Generation Investment Management (Al Gore), and others. Rapid headcount growth.
2019Mark Okerstrom, former CEO of Expedia, joins as President and COO. Engineering team recruited from Amazon, Google, Facebook, and Expedia.
2022Series E: $260 million led by Baillie Gifford and Hercules Capital. JPMorgan extends $150 million credit line. Valuation: $3.8 billion. Allen & Co. sells its entire position — a grand slam on its early check.
2023Multiple layoff rounds. Atlanta office closed. Four-month search for acquirer fails. Shutdown on October 19.
The trajectory from sub-$60 million to $3.8 billion in seven years was not unusual for the era. What is notable is the composition of the cap table. This was not a group of trend-chasing tourists. Bezos understood logistics at a molecular level. CapitalG, Alphabet's growth fund, brought the data infrastructure perspective of the world's most sophisticated information company. Generation Investment Management, co-founded by Al Gore, focused on long-duration sustainability plays — and reducing empty truck miles had a genuine carbon-reduction thesis. Baillie Gifford, the Edinburgh-based investment trust that has backed Amazon, Tesla, and Moderna, specialized in long-horizon conviction bets on technology-driven transformation. These were, by any measure, smart, patient, thesis-driven investors.
And yet. Allen & Co., which had written one of the earliest and smallest checks, sold its entire position in the 2022 round at the $3.8 billion valuation — the only investor, according to PitchBook data, to completely exit before the collapse. The timing was either prescient or lucky; either way, it was a grand slam on the original investment. Every other investor who participated in the Series E — Baillie Gifford, T. Rowe Price, Hercules Capital — would watch their stakes go to zero eighteen months later. The lesson here is not that these investors were foolish. It is that even the most sophisticated capital allocators, deploying the most rigorous analytical frameworks, can be undone by a cyclical collapse in the underlying commodity market that a business is built upon.
We are in the middle of a massive freight recession and a contraction in the capital markets. This combination ultimately crushed our progress at the same time that it was crushing our logical strategic acquirer — it was the perfect storm.
The phrase "perfect storm" is overused in corporate apologetics, but Lewis's usage was, for once, precise. Convoy faced not one existential pressure but two, and they were correlated. The freight recession cratered revenue and gross margin simultaneously, while the capital markets contraction — the post-2022 repricing of venture-stage technology companies — eliminated the possibility of raising the bridge financing that might have carried the company through the trough. In a normal freight downturn, a well-capitalized brokerage can survive by cutting costs and waiting for the cycle to turn. In a normal capital crunch, a business with strong revenue trends can find a buyer or a down-round investor. Convoy got both at once.
The Cyclical Trap: Why Freight Ate the Thesis
To understand why Convoy died, you have to understand how the freight cycle works — and why a venture-backed technology company is structurally ill-suited to survive one.
The U.S. freight market is violently cyclical. During periods of high demand and constrained supply — as in 2020–2021, when the pandemic simultaneously spiked e-commerce volumes and disrupted labor availability — spot-market trucking rates can increase 50% to 80% year-over-year. Carriers who had been struggling to fill trucks are suddenly fielding calls from desperate shippers willing to pay almost any rate to move goods. Brokerages that earn a percentage spread on each transaction see revenues balloon. For a digital freight marketplace like Convoy, this environment is euphoric: more transactions at higher rates, each one generating margin and data.
Then the cycle turns. As it did, savagely, in 2022 and 2023. Demand softened as the post-pandemic consumption binge faded and inflation began eroding purchasing power. New truck capacity that had been ordered during the boom began arriving, flooding the supply side. Rates collapsed. According to data that Lewis referenced in his shutdown memo, the freight market experienced an "unprecedented" decline. Spot rates in many lanes fell below the cost of operation for carriers, driving thousands of small trucking companies out of business — the very carriers who comprised Convoy's supply-side network.
For a traditional freight brokerage — the kind staffed by brokers making phone calls — a downturn is painful but survivable. The business model is inherently flexible: brokers are often compensated on commission, so labor costs adjust downward with revenue. Office leases can be renegotiated. The fixed-cost base is modest. For a venture-backed technology platform with hundreds of software engineers recruited from Amazon, Google, and Facebook, the cost structure is dramatically different. Engineers are salaried. The technology platform requires ongoing investment regardless of transaction volume. The company had been hiring against the up-cycle and building infrastructure for a market that was growing — and then the market reversed.
The cruelty of this dynamic cannot be overstated. Convoy's technology genuinely worked better in a downturn in certain respects: when capacity is abundant, the algorithmic matching engine has more options, and the efficiency gains from digital matching are arguably larger. But the revenue to fund the technology evaporated precisely when the technology's theoretical value was highest. It is the classic venture dilemma applied to a cyclical industry: the growth capital model assumes a relatively monotonic growth trajectory, or at least the option to raise through temporary setbacks on the strength of long-term TAM. When the industry-level revenue drops 30–40% and the capital markets close simultaneously, even the most compelling long-term thesis cannot survive the short-term liquidity crisis.
The Analog Fortress: Why Brokers Survived and Platforms Didn't
The most instructive contrast is not between Convoy and Uber Freight — the other high-profile entrant into digital freight — but between Convoy and C.H. Robinson, XPO Logistics, and Landstar Systems, the incumbent brokerages whose analog business model Convoy was explicitly designed to replace.
According to Morgan Stanley data referenced in Convoy's own narrative, thousands of private companies controlled approximately 60% of the U.S. freight brokerage market, with a handful of public companies like C.H. Robinson, XPO, and Landstar capturing the remainder. These incumbents were not ignorant of technology — C.H. Robinson had been investing in its Navisphere platform for years, and XPO had built significant digital capabilities. But their core business model retained the flexibility of human-mediated brokerage: variable-cost brokers who could be hired and fired with the cycle, deep shipper relationships that provided contract freight as a buffer against spot-market volatility, and decades of accumulated data on lane pricing, carrier reliability, and seasonal patterns.
When the freight recession hit, the incumbents contracted. C.H. Robinson's margins compressed. XPO spun off its brokerage operations. Landstar's load volumes fell. But they survived. Their cost structures bent rather than broke, because the human broker, for all the inefficiency that Lewis rightly identified, is inherently a variable-cost resource in a way that a software engineering team is not.
Lewis anticipated the competitive dynamic clearly enough. "There will always be analog components because people need to pick up the phone and connect with each other," he told CNBC in 2022. "But I think that over the next decade, the industry is going to consolidate and it will consolidate around companies that are digitally led and data led, that are able to use data" — and here the sentence trails off in the source, but the implication is clear. Lewis believed that the structural advantages of digital platforms would, over a full cycle, overwhelm the operational flexibility of analog brokerages. He may have been right about the decade-long trend. He was fatally wrong about the eighteen-month timeline.
The Uber Freight Shadow
Convoy was not alone in pursuing the digital freight thesis. Uber, whose brand was synonymous with marketplace-driven disruption of legacy transportation, launched Uber Freight in 2017 as a natural extension of its platform into the freight brokerage space. The two companies became the most prominent digital freight startups of the era, competing for the same carriers, the same shippers, and the same narrative: technology would disintermediate the analog freight brokerage.
The competitive dynamic between Convoy and Uber Freight illuminates a structural truth about platform businesses in non-winner-take-all markets. In ridesharing, Uber and Lyft achieved effective duopoly because riders and drivers in a given city tend to converge on one or two platforms for liquidity. In freight, the marketplace dynamics were different. A carrier running a route from Dallas to Atlanta does not need continuous, real-time matching the way a rideshare driver does — the carrier needs one load per day, or one load per trip, and has tolerance for a multi-hour search window. This means the liquidity advantage of a dominant platform is weaker in freight than in ridesharing. A carrier can — and typically did — use multiple platforms and multiple brokers simultaneously, splitting their attention without meaningful cost.
This fragmented attention on the supply side meant that neither Convoy nor Uber Freight could build the kind of compounding network effects that had powered Uber's ridesharing business. More shippers on the platform attracted more carriers, but not exclusively more carriers — the same carriers were also on Uber Freight, on C.H. Robinson's platform, and answering phone calls from independent brokers. The marketplace was leaky. And in a leaky marketplace, technology advantage alone is insufficient to build a durable moat.
Uber Freight, critically, had one advantage Convoy lacked: a parent company with a balance sheet measured in billions and the strategic commitment to subsidize a freight operation through a full cycle. When the freight recession hit, Uber Freight contracted — it was not immune to the cycle — but it had access to Uber's corporate treasury and its public-market capital-raising apparatus. Convoy, as a standalone venture-backed company, had no such backstop.
The Talent Machine That Couldn't Survive Its Own Overhead
One of Convoy's genuine competitive advantages, and ultimately one of its liabilities, was the caliber of its engineering and leadership team. Lewis and Goodale had recruited aggressively from the upper echelons of Seattle's technology ecosystem. Mark Okerstrom, who joined as President and COO, had been CEO of Expedia — a company that had pioneered the application of data science to matching supply (hotel rooms, flights) with demand (travelers) across a global marketplace. The parallels between Expedia's marketplace and Convoy's were obvious.
We have assembled a team of really world class engineers from places like Amazon, Google, Facebook, Expedia, etcetera, which are working on solving freight customers' hardest problems. That is our primary investment area.
The engineering investment was not vanity. Convoy's matching algorithms, pricing models, and carrier-facing mobile applications represented genuine technical achievement. The Convoy Go trailer-pool product required sophisticated scheduling systems that coordinated shipper loading windows, trailer locations, and driver availability across complex logistics networks. The contract freight product built into the app — allowing carriers to find dedicated work on consistent lanes — required a recommendation engine that understood carrier preferences, historical reliability, and economic optimization simultaneously.
But world-class engineers from Amazon, Google, and Facebook command world-class compensation. Seattle software engineer salaries in the 2019–2022 period were among the highest in the industry, and Convoy was competing for talent against its own former employers. The fixed cost of maintaining a several-hundred-person engineering organization was enormous relative to the gross margins of a freight brokerage, even a digital one. When revenue collapsed in the downturn, this cost structure became an anchor. The layoffs that preceded the shutdown — from 1,500 to roughly 500 employees — represented an attempt to cut fast enough to extend the runway, but in a business where the core product was the technology, every engineering layoff degraded the product that was supposed to be the competitive advantage.
The Four-Month Death March
The final chapter of Convoy's operational life lasted from approximately June to October 2023. Lewis and his team spent four months — "an exhaustive process," he wrote — exploring "all viable strategic options for the business." The details of this process remain largely private, but the contours are visible in Lewis's shutdown memo and subsequent reporting.
The "most likely acquirer," whose identity Lewis did not disclose publicly, was itself a victim of the same forces destroying Convoy. This is telling. In a freight recession, the natural acquirers of freight technology companies are other freight companies — and those companies are themselves struggling. A strategic acquirer in the logistics space in mid-2023 would have been looking at its own compressed margins, uncertain demand outlook, and a capital environment that made financing acquisitions difficult. The technology was valuable in abstract — Lewis wrote that Convoy had built "the conditions for a truly scalable technology platform and business model that would have yielded real financial gains when market conditions improve" — but value in abstract does not generate the cash required to acquire a company burning through its runway.
The technology, stripped of its operational context, did eventually find a buyer. In a post-shutdown transaction, freight data and analytics company DAT acquired Convoy's technology assets — the algorithms, the data, the intellectual property — in a deal that represented a fraction of the company's peak valuation. For Convoy's investors, the recovery was negligible. For DAT, it was a bargain acquisition of technology that had cost over $1 billion to develop.
Lewis's memo to employees was, by the standards of corporate death notices, notably honest. "We moved all business levers possible. But we were running up the down escalator… and it kept speeding up." The metaphor is apt. A down escalator doesn't care how fast you run.
The Carbon Thesis That Died With the Company
One of the underappreciated dimensions of the Convoy story is the environmental thesis. The company's pitch to shippers included not just cost savings and efficiency, but a genuine carbon-reduction argument. If roughly one-third of truck miles in the United States are empty miles — trucks running without cargo between loads — then any technology that reduces empty miles directly reduces carbon emissions from trucking. Trucking is one of the largest sources of transportation-related carbon emissions in the United States, and the scale of the inefficiency meant that even modest improvements in utilization could yield meaningful environmental gains.
This was not greenwashing. The empty-miles problem is real, measurable, and directly addressable through better matching algorithms. Generation Investment Management, Al Gore's sustainability-focused fund, invested in Convoy in part on this thesis. Grant Goodale, Convoy's co-founder, spoke publicly about the company's work on electric truck deployment and the logistics infrastructure required to support electrified freight. The company was thinking seriously about not just digitizing the existing system but using the digital layer to enable the transition to lower-carbon trucking.
The death of Convoy did not eliminate the empty-miles problem. The one-third empty-mile statistic persists. The carbon emissions persist. The technology that Convoy built to address them was acquired by DAT — a company whose primary business is providing freight rate data and analytics, not operating a carrier-matching marketplace. Whether that technology will be deployed in a way that captures the environmental value Convoy envisioned remains an open question. What is certain is that the urgency of the problem outlived the company that was best positioned to solve it.
Lessons from the Wreckage
The Convoy story contains multiple cautionary narratives layered atop each other, each legible from a different angle.
For founders: the danger of applying a venture-growth-capital structure to a cyclical commodity business. The venture model assumes that time is on the company's side — that each additional dollar of capital buys time to build technology, acquire users, and reach the scale at which the business becomes self-sustaining. In a cyclical industry, time is not reliably on your side. The cycle can turn at exactly the wrong moment, and the capital markets can close at exactly the wrong moment, and these two events can be correlated because they share a common cause (macroeconomic tightening).
For investors: the limits of thesis-driven conviction in the face of cyclical risk. Every investor on Convoy's cap table could articulate a compelling long-term thesis for digital freight brokerage. The thesis may still be correct. But a thesis about a ten-year industry transformation does not immunize a portfolio company against an eighteen-month liquidity crisis. The question "Is this the future of freight?" and the question "Can this company survive the next two years?" are different questions, and the answer to the first can be yes while the answer to the second is no.
For operators watching the freight industry: the analog brokerage, for all its inefficiency, possesses a form of anti-fragility that digital platforms do not. Variable costs, human relationships, and operational flexibility allow traditional brokerages to compress during downturns and re-expand during recoveries. The fixed-cost technology infrastructure that gives a digital platform its efficiency advantage in good times becomes a liability in bad times. The industry may indeed consolidate around digitally-led companies, as Lewis predicted. But the survivors of that consolidation may be incumbents that added technology to an analog core, rather than technology companies that attempted to rebuild the industry from scratch.
The Phantom Convoy
There is a resonance — unintended but structurally perfect — between the trucking startup's name and the word's deeper history. The original convoys were formations of merchant ships sailing in groups during wartime, protected by naval escorts against submarine attack. The logic was collective: individual ships were vulnerable; a coordinated group, sharing intelligence and protection, could survive threats that would destroy any single vessel. The most famous convoy battles of World War II — the brutal Atlantic crossings documented in Martin Middlebrook's
Convoy SC122 and HX229 — were studies in the tension between coordination and chaos, between the theoretical advantage of system-level optimization and the brutal reality of individual ships being torpedoed despite every precaution.
Convoy the company attempted something analogous: to coordinate the fragmented, individual-operator chaos of American trucking into a system-level network that would be more efficient, more resilient, and more valuable than the sum of its parts. The theoretical advantage was real. The optimization algorithms worked. The network grew. And then the economic equivalent of a U-boat wolfpack — the simultaneous freight recession and capital contraction — found the convoy in open water, and the escort was insufficient.
On the day Convoy shut down, the trucking industry's empty-mile rate remained approximately where it had been when Lewis and Goodale founded the company eight years earlier. One-third of the time, trucks still ran empty. The phones still rang in brokerage offices from Memphis to Chicago. The fax machines, improbably, still hummed. And somewhere on I-80, a truck drove 150 miles with nothing in its trailer to pick up a load that a phone call had arranged, the diesel burning, the tires wearing, the carbon climbing — exactly the problem Convoy had been built to solve, persisting now without it.
