A hill chief who had never sailed decided to grow a navy the way you grow a crop. In June 1664, alarmed English clerks in Surat wrote to Karwar that Shivaji was "fitting up and building" sixty frigates at once.1 He had no shipwright tradition, no deep-water fleet, no naval caste — just a coastline he refused to leave to the Portuguese, the Dutch, and the British. So he started from raw timber and farmhands and willed a fleet into existence. The building of the Maratha fleet is the story of a man manufacturing a capability his people did not have, in plain sight of every power that already had it, and frightening all of them anyway.
Picture the western Indian seaboard in the 1660s. The waters belong to outsiders. Portuguese carracks, Dutch East Indiamen, British convoys — these move freely, tax freely, fight freely. No local ruler contests them at sea. Shivaji looked at that arrangement and saw a hole in his own kingdom: a king who controls hills and forts but not the water that borders them controls only half a realm.2 From the late 1650s he had quietly begun acquiring frigates in small numbers.3 After the 1664 sack of Surat — which filled his treasury and put a scare into the Mughals — the trickle became a program. He scouted the coast himself, rebuilt sea forts, and turned shipbuilding into a standing project rather than an occasional purchase.4 The raw situation this page ingests is simple and brutal: a land power, late to the maritime race, with poor knowledge of shipbuilding technology and limited ammunition, deciding to compete anyway.5
The fleet's most revealing feature is that nobody watching it could agree on its size. The British in Surat reported sixty frigates being built.6 A month later the Portuguese captain at Chaul counted fifty ships fitted for combat, with seven already finished in upper Chaul and nearly ready for sea.7 The Dutch separately logged fifty frigates "on the stocks along the straits."8 Three observers, three counts — and that spread is not sloppy bookkeeping. It is the mechanism working. A fleet under construction in scattered coves, by a power that does not publish its order of battle, generates exactly this fog. Each European report is a guess filtered through fear, and fear inflates. The contradictory counts are themselves the weapon: an enemy who cannot pin your strength must plan against the worst case. The Portuguese council at Goa told its Chaul captain not to lock horns with Shivaji and not to block his ships — and still sent fifty men from Bassein in case.9 Uncertainty did the work of an actual battle fleet before a single Maratha hull fired a shot.
This page isolates the construction angle — the from-nothing manufacture of a capability — and hands the rest of the vault a clean case of bootstrapped state power. It feeds Maratha Maritime Power and Naval Strategy the raw industrial substrate beneath the strategy: before there is a doctrine of coastal raiding there must be hulls, and here is where the hulls come from. It hands Sindhudurg the Island Fortress its purpose — a fleet needs a fortified harbor to shelter in, and the forts and the frigates are one program in two materials, stone and timber. It supplies The Basrur and Karwar Naval Raid 1665 the instrument: the raid is what these frigates were built to do. And it sets the table for The Maratha Naval Commanders — once you have ships, you need men who can sail them, and the question of who manned them becomes urgent precisely because the ships came first.
Read the chronicler Sabhasad's list of Shivaji's vessels and you can almost walk the dockyard. The workhorse was the galbat — a small two-masted craft of forty to seventy tons carrying four to eight guns, nimble enough to dart and harass.10 Above it sat the tarande, the big sailing vessels, the heavy lifters of the line.11 The prize of the taxonomy was the gurab: a two-masted ship of roughly 150 tons, built broad in proportion to its length and narrowing toward the stern, deliberately shaped to draw very little water.12 That last detail is the whole strategic idea compressed into a hull. A shallow-draft warship can run up rivers and hug coasts where a deep-keeled European galleon would ground itself. The British feared exactly this — that Shivaji meant to "run up the river of Cambay" and land an army to march on Ahmadabad.13 The gurab is the answer to a question about geography.
The list keeps going, and it tells you the fleet was not purely military. The taru was a plain sailing vessel.14 The sibad was a large, square-sterned, flat-bottomed two-master with no deck, known chiefly for trade.15 The pagar was barely a warship at all — a well-smoothed canoe.16 A navy that includes trading craft and canoes alongside gun-carrying galbats is a navy doing two jobs: fighting and commerce. The Dutch caught this plainly, noting that Shivaji "intends to open up an intensive traffic in many quarters" and was "stimulating the merchants to traffic."17 In the very same year, his official Raoji Somnath Pandit got two small ships back from Mocha at such a profit that he readied eight or nine more for Mocha, Persia, Muscat, and beyond.18
The strategic intent reads straight off the vessel types. The fast galbats and shallow gurabs exist to waylay the Surat merchant traffic returning from Basra and Persia, to threaten Cambay, and to break the foreign monopoly on the sea-lanes.19 The fat-bellied sibads and trading taru exist to carry Maratha goods and turn the same coast into revenue. Build a mixed fleet and you have built both a blockade and a trading house. The taxonomy is the business plan.
One more thing the case study exposes: the human cost hidden under the timber. To row these vessels Shivaji conscripted farmers, pulling cultivators off the land to pull oars.20 It did not hold. Rowers deserted mid-voyage — men who knew soil, not sea, slipping away when the coast was close enough to walk home from.21 A fleet can be built faster than a crew can be made loyal, and the desertions are the seam where the rush shows.
You are a Dutch factor at Vengurla, late summer 1664, and a letter from upcountry says the hill-king has fifty hulls on the stocks. You have never seen them. You count what you can: seven ships finished at Chaul, frigates rumored along the straits, a number that climbs every time a different informant talks. Sit with that climbing number. Notice that you cannot verify it and that this is the point. The man building the fleet has never sailed — your own report says "he never before sailed on the sea" — and yet "the awe he has inspired" already lets him threaten losses he has not yet inflicted.22 Feel the trap close: if you wait for proof of the fleet, the fleet arrives; if you act on the rumor, you legitimize it. You advise your superiors to keep friendly relations, decline to lend ships against him, and quietly move men to Chaul anyway. You have just been governed by a navy that has not put to sea. The transferable move is to read the contour of a capability while it is still being poured — to treat a competitor's announced build as a present fact, because by the time it is verified it is also too late. The skill is not naval. It is the discipline of pricing in what is coming instead of waiting for what has come.
A fleet-building program rots in recognizable ways, and the source shows two of them mid-collapse. The first is the hollow hull — ships counted but never crewed. When you have fifty frigates on the stocks and farmers fleeing the oars, you have tonnage without sea-power; the dockyard is full and the water is empty. Watch for the gap between hulls and hands: a capability measured in inventory rather than in operations is a capability that exists on paper and dies on launch day. The second sign is the unowned skill — a navy whose technology its builders barely understand, dependent on conscripts who would rather be farming. The remedy Shivaji eventually reached for, recruiting coastal communities who actually belonged to the sea, points at the failure by contrast: when the men manning the capability do not love or know it, the capability leaks at the seams. A third diagnostic is subtler — believing your own enemy's count. If Shivaji had taken the inflated European tallies as truth, he might have slackened, trusting in ships that were rumor. The build works only while the builder knows the real number and lets everyone else guess.
This page rests on Vaibhav Purandare's Shivaji: India's Great Warrior King [POPULAR SOURCE], a narrative biography that here leans heavily on contemporary European archives — English East India Company letters, Portuguese Chaul–Goa correspondence, and Dutch factory reports — alongside the Marathi chronicle Sabhasad.23 The European records are strong corroboration for one another on the fact of a major build, but they openly disagree on scale: British sixty, Portuguese ~fifty combat-ready, Dutch fifty on the stocks.24 That spread is preserved here, not resolved — it is the genuine threat-uncertainty of the moment, and flattening it to a single figure would erase the very fog that gave the fleet its early power. The ship taxonomy is [SINGLE SOURCE] on Sabhasad for the Marathi vessel names and tonnages, though the gurab and galbat are well attested elsewhere in Indian Ocean maritime history. Open: the actual launched-and-operational count is unknown; "on the stocks" and "ready for sea" are different states the records blur. Open: how many conscripted rowers deserted, and whether desertion materially delayed the early voyages, is not quantified.
Only one modern author writes here — Purandare — but he is a referee standing over several quarreling witnesses, and the quarrel is the content. The English, Portuguese, and Dutch factors converge on the headline: Shivaji is building a serious fleet and it frightens everyone. They diverge the instant they count. The British put it at sixty; the Portuguese captain at Chaul says fifty fitted for combat with seven ready; the Dutch log fifty on the stocks.25 Set against these European tallies sits Sabhasad, the Maratha chronicler, who offers no grand number at all but instead a granular vessel list — galbat, tarande, gurab, taru, sibad, pagar.26 The split between the European witnesses and Sabhasad reveals something neither sees alone. The Europeans, anxious and outside the program, measure the threat as a quantity — how many hulls, how soon. Sabhasad, inside the tradition, measures it as a kind — what sorts of craft, for what work. Put the two together and you recover what each misses: the Europeans tell you the fleet was big enough to terrify; Sabhasad tells you it was mixed enough to do two jobs, war and trade, at once. The threat was real (European convergence), its size was genuinely uncertain (European tension), and its design was deliberate (Sabhasad's taxonomy). Purandare's restraint in leaving the numbers unreconciled is the right call — the disagreement is evidence, not error.
A hill-king building ships he doesn't know how to build connects to two questions that have nothing obviously naval about them: how do you create a capability from zero, and how does a ruler legitimize ownership of something — here, the sea — that no one thought he could own? Both reach out of history into other domains.
History (statecraft): Arthashastra — Kingship and the Rajarshi Ideal — Kautilya's manual treats the building of new state capacity (forts, treasuries, armed forces) as a deliberate, sequenced act of the ideal king rather than an inheritance. Shivaji's fleet is that doctrine run live: a from-zero capability willed into existence as a matter of policy. What differs is the medium — the Arthashastra assumes a land kingdom, and there is no chapter on growing a navy you have no tradition for. The parallel produces an insight neither text gives alone: the rajarshi ideal of the self-making king implies that a true ruler must be able to manufacture entirely new domains of power, not just administer existing ones — and the sea-fleet is the test case of a king extending sovereignty into an element his caste and his geography never claimed. Capability-building is itself a kingly virtue, and the dockyard is where you watch the virtue get exercised on raw timber.
Behavioral-mechanics (deterrence by ambiguity): Sun Tzu — Field Intelligence and Signal Reading — Sun Tzu's core teaching is that the read precedes the fight: the commander who controls what the enemy can know controls the engagement before contact. Shivaji's fleet weaponizes exactly the gap Sun Tzu wants to manage — the European powers cannot get a clean signal on his strength, so their counts scatter from fifty to sixty, and they end up planning against a worst case that may not exist. What differs is who holds the advantage: Sun Tzu writes for the reader of signals, while here Shivaji is the emitter of an unreadable signal, governing his enemies through the noise his scattered dockyards produce. The insight neither domain alone generates: a capability under construction can deter more effectively than a finished one, because the unfinished version is uncountable and therefore unboundedly frightening — deterrence runs on the variance of the estimate, not its mean. The Dutch said it themselves, that "the awe he has inspired" let him threaten losses he had not yet inflicted.27
These two handshakes split the fleet into its two faces. The Arthashastra connection reads it as construction — a king making new power. The Sun Tzu connection reads it as signal — a king making new fear. The fleet is both at once, and that doubleness is exactly why a land power could rattle four maritime empires before its ships were even crewed.
The Sharpest Implication If a man who had never sailed could frighten the Portuguese, Dutch, and British by building ships he barely understood, then the threat was never the ships. It was the decision to build them, made visibly, by someone who refused to accept that the sea belonged to other people. That should unsettle any assumption that you compete by first becoming competent. Shivaji competed by announcing intent and pouring hulls, and the competence chased the commitment rather than preceding it. The uncomfortable reading: the moat around an incumbent's advantage is often just everyone else's belief that you need their tradition to enter — and that belief dissolves the moment one outsider starts building in the open and lets the worst-case estimate do the rest.
Generative Questions