The Map Wars · Part 3
At CES 2023, HERE unveiled its next-generation map-making platform, UniMap: automated production, SD, HD and indoor maps unified into a single digital representation, changes publishable within 24 hours. Put its product page next to TomTom's Orbis and you can match them almost line for line.
After forty years of combat, the two companies' technology slides have become interchangeable.
That is not the end of the story; it is where the real story begins. When technology no longer decides the winner, what does? The answer hides in a choice each company made around 2022. But to understand that choice, you have to go back to the genes.
Two Sects
HERE belongs to the sect of field survey. The Navteq bloodline: survey fleets, field offices, a "ground truth" faith in having personally seen every road — the HERE True LiDAR fleet is that tradition's continuation. The trust it sells carmakers is built on a chain of physical verification: behind every data point, someone checked the real world.
TomTom belongs to the sect of probes. The Tele Atlas-plus-PND bloodline: as early as 2008 it built the world's first community-validated map database on twenty million devices — over 50,000 community-verified corrections in a single release, matching all other sources combined — and later grew the probe network to hundreds of millions of devices. Its trust is built on statistics: hundreds of millions of eyes watching the roads on its behalf.
The two philosophies have opposite cost curves. Field survey: high fixed cost, high precision, low freshness. Probes: low marginal cost, high freshness, precision by statistical convergence. Forty years on, each is poaching from the other — HERE chases fleet data feedback for freshness, TomTom builds LiDAR vehicles for precision — but genes decide default behavior: under budget pressure, HERE protects the fleet first; TomTom protects the probes.
A Misaligned Decade
2016 to 2018 was HERE's golden hour. HD Live Map introduced the "self-healing map"; the Open Location Platform aspired to be an exchange for location data; the sensor-data consortium of its three German shareholders handed it a real-time data source no one else had; it donated the SENSORIS standard to the industry. The route's crowning proof: Mercedes' Drive Pilot — among the first certified Level 3 systems anywhere — runs on HERE's HD map. For those few years, HERE was the industry's designated definer of the next-generation map.
TomTom spent the same period doing something unglamorous: demolishing and rebuilding its production line. The new map-making platform, internally codenamed Genesis, took heavy investment from around 2019; in 2020 Uber became a map-editing partner, feeding customer business data back into the line; in 2021 the cutover stumbled — releases slowed, automotive deliveries suffered, the stock took a beating; in 2022 the automation paid off and roughly 500 map-production jobs, a tenth of the workforce, were cut — an ING analyst called it the "holy grail of fully automated map-making" finally reached; that November the new Maps Platform launched, and in 2023 OpenStreetMap data formally entered the map.
The misalignment sits right there: HERE did "map as a service" first, on an old production line; TomTom rebuilt "the line as a factory" first, and its platform came late. HERE's platform vision was five years early and never became an exchange; TomTom's rebuild hurt for two years and genuinely changed the cost structure. A decade on, one had accumulated a narrative; the other had accumulated a factory.
Converging on the Same Destination
Then the opening scene: 2023, UniMap, and HERE too converged on "one map, many views, continuous publication." The two companies had arrived, by opposite roads, at the same technical destination — automation replacing manual editing, a single database deriving every product, updates measured in days rather than quarters.
Headcount curves cross-validate the story. TomTom cut 500 map editors in 2022 and another 300 staff in June 2025 in its AI pivot; HERE has run round after round of restructuring — in its employees' own words, "constant reorgs, management churn, random layoffs." The direction is identical: editing jobs vanish, AI jobs grow. The difference is the narrative — each TomTom cut maps to a production-line milestone; HERE's cuts map to financial pressure. The same act: for one, a byproduct of upgrading the factory; for the other, a tourniquet.
The technology has fully converged. That is the most important signal of all: whatever decides the winner next will not be found in the technology.
The 2022 Question
The fork came over raw materials.
In late 2022, TomTom co-founded Overture and swapped its base map for open data — choosing "open ingredients plus a proprietary factory": donate the base layer to the commons, and pour every euro saved into the dynamic layers — live traffic, lane-level data, charging-station status. That bet is about cost structure: base-map commoditization is irreversible; concede early, save early, migrate value upward.
HERE chose not to move. UniMap remains a fully proprietary vertical stack — its own ingredients, its own line, its own delivery. That bet is about liability structure: automotive customers want an end-to-end chain of responsibility, and carmakers won't touch open data's licensing contamination and liability ambiguity.
Both bets are coherent. But a preliminary verdict is already in: Volkswagen Group's software company CARIAD signed its ADAS-grade map contract with TomTom — the German OEM system personally certifying that open ingredients can clear automotive-grade delivery. The hardest plank in HERE's argument is failing.
The deeper difference lies beneath the moats. HERE's trump card — the shareholder fleet sensor consortium — has a structural flaw: the data moat is welded to the equity; as shareholders drift, the moat leaks. Mercedes piping Google's data into its own navigation is the sound of the leak. TomTom's probes come from customer products and commercial contracts — Uber's trips, Microsoft's platforms, its own apps — untouched by anyone's shareholding. One moat is held together by a marriage, the other by business. Which holds longer over time needs no argument.
One Question Left
Forty years bring the account to this: the collection philosophies have converged — the surveyors are hunting probes, the probe people are buying LiDAR; the production technology has converged — UniMap and Orbis state nearly identical goals. The only thing that has not converged is commercial philosophy: should the raw material be open, and at which layer do you charge?
That question will not be settled by technical review. It will be settled by market clearing — and the clearing is underway: contract flows, shareholder patience, where the talent goes, a scoreboard updated every quarter.
For readers making cockpit sourcing decisions for export programs, here is a reading tool worth more than any platform name. Ask a supplier three questions. Where does your freshness come from — a probe network, fleet feedback, or an open community? Where is your raw-material cost curve heading — proprietary base-map maintenance only rises, while open-base-plus-curation is falling? And where does your chain of liability close — end-to-end proprietary, or open base plus an automotive-grade validation layer? Ask all three, and the answer surfaces on its own.
In an industry whose technology has converged, choosing a supplier is choosing a commercial philosophy. And the correctness of a commercial philosophy is never decided by debate — it is decided by whichever company is still standing.
Sources: HERE UniMap launch materials (CES 2023); TomTom press archives and annual reports; Reuters and euronews coverage of TomTom's 2022 map-making automation restructuring; Tele Atlas's 2008 MultiNet release notes; Glassdoor employee reviews; Bundeskartellamt decision B7-25/22.
《地图战争》系列 · 第3篇
2023年的CES上，HERE发布了新一代制图平台UniMap：自动化制图，SD、HD、室内地图统一到一个数字表示，变更24小时内可发布。把它的产品页和TomTom的Orbis并排放，几乎可以逐条对应。
四十年缠斗，两家的技术PPT终于可以互换了。
这不是故事的结束，是真正的故事开始的地方：当技术不再分胜负，分胜负的是什么？答案藏在两家公司2022年前后各自做的一个选择里。但要看懂那个选择，得先回到基因。
两个教派
HERE是实测教派。Navteq的血统：测绘车队、外业办公室、亲眼看过每条路的"ground truth"信仰，后来的HERE True激光雷达车队是这个传统的延续。它卖给车厂的信任，建立在实地测量的责任链上——每个数据点背后都有一次物理世界的核对。
TomTom是探针教派。Tele Atlas加PND的血统：2008年就用两千万台设备做出全球第一个社区验证的地图数据库，单个版本里靠社区输入校验的变更超过五万处，检出量追平其他全部数据源的总和；后来探针网络扩到数亿设备。它的信任建立在统计学上——几亿只眼睛替它看路。
两种哲学的成本曲线正好相反。实测：高固定成本，高精度，低鲜度。探针：低边际成本，高鲜度，精度靠统计收敛。四十年后的今天，两家都在偷师对方——HERE拼命找车队回流数据补鲜度，TomTom建激光雷达车队补精度——但基因决定了各自的默认动作：遇到预算压力，HERE先保车队，TomTom先保探针。
错位的十年
2016到2018年是HERE的高光。HD Live Map提出"自愈地图"，Open Location Platform想做位置数据的交易所，德系三家股东的车队传感器数据联盟给了它独一无二的实时数据源，SENSORIS标准也是它送进行业的。这条路线的巅峰证明是奔驰Drive Pilot——全球最早获认证的L3之一，跑在HERE的高精地图上。那几年的HERE，是全行业眼里"下一代地图"的定义者。
同一时期的TomTom在做一件不性感的事：推倒重建产线。内部代号Genesis的新制图平台从2019年前后开始重投入；2020年把Uber变成地图编辑伙伴，客户业务数据反哺产线；2021年切换翻车，发版放缓、汽车交付受扰、股价挨打；2022年自动化兑现，裁掉Maps部门约500人、占全员一成，ING分析师的评语是"多年追求全自动制图的圣杯，这轮投资看来成功了"；同年11月发布新Maps Platform，2023年OSM数据正式入图。
错位就在这里：HERE先做了"地图当服务"，但产线还是老的；TomTom先重建了"产线当工厂"，但平台来得晚。HERE的平台愿景超前了五年，没做成交易所；TomTom的产线重建疼了两年，真改了成本结构。十年下来，一家攒了叙事，一家攒了工厂。
殊途同归
然后是开头那一幕：2023年，UniMap发布，HERE也收敛到了"一张图、多视图、连续发布"。至此两家在技术路线上殊途同归——自动化替代人工编辑，单一数据库派生全部产品，更新以天计而不是以季计。
人员曲线可以做交叉验证。TomTom在2022年裁500名制图员、2025年6月再裁300人转向AI；HERE近年多轮重组，员工评价的原话是"不断重组、管理层轮换、随机裁员"。方向完全一致：编辑岗位消亡，AI岗位扩张。区别在叙事——TomTom的每次裁员都对应一个产线里程碑，自动化兑现一次、收缩一次；HERE的裁员对应的是财务压力。同样的动作，一个是工厂升级的副产品，一个是止血带。
技术全收敛了。这正是最重要的信号：接下来分胜负的，必然不在技术里。
2022年的选择题
分岔发生在原料上。
2022年底，TomTom共创Overture、把底图换成开放数据，选了"开放原料加专有工厂"：底图交给公共品，省下的维护成本全部投向实时交通、车道级数据、充电桩状态这些动态图层。这一注赌的是成本结构——底图商品化不可逆，早认早省，价值往上层迁移。
HERE选择不动，UniMap依然是自家原料、自家产线、自家交付的全专有垂直栈。这一注赌的是责任结构——车规客户要的是端到端责任链，开放数据的许可证问题和责任界定，车厂不敢碰。
两个赌注都成立，都有逻辑。但初步判决已经出来了：大众集团的软件公司CARIAD把驾驶辅助级地图签给了TomTom——等于德系车厂亲手证明了"开放原料可以完成车规化交付"，HERE论点里最硬的那一条正在失效。
更深一层的差别在数据护城河的地基上。HERE的王牌——股东车队传感器联盟——有个结构性弱点：数据护城河和股权绑在一起，股东离心，护城河跟着漏水。奔驰把谷歌的数据接进自家导航，就是漏水的声音。TomTom的探针来自客户产品和商业合同——Uber的行程、微软的平台、自有应用——不随任何股权关系松动。一条护城河靠婚姻维系，一条靠生意维系，长期哪条更稳，不需要论证。
只剩一道题
四十年走到这里，可以做个总结账：采集哲学收敛了——实测派在找探针，探针派在买激光雷达；产线技术收敛了——UniMap和Orbis的目标几乎逐条相同。唯一没有收敛的，是商业哲学：原料该不该开放，价值该在哪一层收钱。
这道题不会由技术评审决定，会由市场清算决定。清算已经在进行：合同的流向、股东的耐心、人才的去处，每个季度都在给这道选择题打分。
给正在做出海座舱选型的读者一套判读工具，比记住任何平台名字都有用。问供应商三个问题：你的鲜度从哪里来——探针网络、车队回流，还是开放社区？你的原料成本曲线往哪走——专有底图的维护费只涨不跌，开放底图加策展的成本在跌？你的责任链在哪里收口——端到端专有，还是开放底图加车规验证层？三个问题问完，答案会自己浮出来。
技术收敛的行业里，选供应商本质上是在选一种商业哲学。而商业哲学的对错，从来不由辩论决定——由活下来的那家决定。
数据来源：HERE UniMap发布材料（CES 2023）、TomTom历年新闻稿与年报、Reuters与euronews对2022年TomTom制图自动化重组的报道、Tele Atlas 2008年MultiNet发布说明、Glassdoor员工评论、Bundeskartellamt B7-25/22决定。